Sylloge of Defixiones from the Roman West, Volumes I and II: A comprehensive collection of curse tablets from the fourth century BCE to the fifth century CE 9781407359311, 9781407359328, 9781407315324, 9781407353821

Citation preview

B A R I N T E R NAT I O NA L S E R I E S 3 0 7 7 ( I )

2022

Sylloge of Defixiones from the Roman West A comprehensive collection of curse tablets from the fourth century BCE to the fifth century CE

Volume I C E L I A S Á N C H E Z N ATA L Í A S

B A R I N T E R NAT I O NA L S E R I E S 3 0 7 7 ( I )

2022

Sylloge of Defixiones from the Roman West A comprehensive collection of curse tablets from the fourth century BCE to the fifth century CE

Volume I C E L I A S Á N C H E Z N ATA L Í A S

Published in 2022 by BAR Publishing, Oxford BAR International Series 3077 (I) Sylloge of Defixiones from the Roman West, Volume I ISBN  978 ISBN  978 ISBN  978 ISBN  978

1 1 1 1

4073 4073 4073 4073

5931 1 Volume I (paperback) 5932 8 Volume II (paperback) 1532 4 (Set of both volumes) paperback 5382 1 (Set of both volumes) e-format

doi  https://doi.org/10.30861/9781407315324 A catalogue record for this book is available from the British Library. © Celia Sánchez Natalías 2022 Cover image  Numen from SD 118 (drawing: Celia Sánchez Natalías). The Author’s moral rights under the 1988 UK Copyright, Designs and Patents Act are hereby expressly asserted. All rights reserved. No part of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any form of digital format or transmitted in any form digitally, without the written permission of the Publisher. Links to third party websites are provided by BAR Publishing in good faith and for information only. BAR Publishing disclaims any responsibility for the materials contained in any third-party website referenced in this work.

BAR titles are available from: Email Phone Fax

BAR Publishing 122 Banbury Rd, Oxford, ox2 7bp, uk [email protected] +44 (0)1865 310431 +44 (0)1865 316916 www.barpublishing.com

Of Related Interest

Nuove epigrafi greche da Halaesa Archonidea Dati inediti sulle Tabulae Halaesinae e su una città della Sicilia tardo-ellenistica Emiliano Arena BAR International Series 3017

Oxford, BAR Publishing, 2020

Blame it on the Gender Identities and transgressions in Antiquity Edited by Maria Cristina de la Escosura Balbás, Elena Duce Pastor, Patricia González Gutiérrez, María del Mar Rodríguez Alcocer and David Serrano Lozano

BAR International Series 3005

Oxford, BAR Publishing, 2020 Paleoetnología de la Hispania Céltica, Tomo I y Tomo II Etnoarqueología, etnohistoria y folklore Pedro R. Moya-Maleno

BAR International Series 2996

Oxford, BAR Publishing, 2020 Le c.d. gammadiae nelle catacombe cristiane di Roma Censimento, confronti ed ipotesi interpretative Cristina Cumbo

BAR International Series 2947

Oxford, BAR Publishing, 2019 Roles for Men and Women in Roman Epigraphic Culture and Beyond

Gender, social identity and cultural practice in private Latin inscriptions and the literary record Peter Keegan

BAR International Series 2626

Oxford, BAR Publishing, 2014 The Archaeology of Gender, Love and Sexuality in Pompeii Lourdes Conde Feitosa

BAR International Series 2533

Oxford, BAR Publishing, 2013 Lifting the Veil: a New Study of the Sheela-Na-Gigs of Britain and Ireland Theresa C. Oakley

Oxford, BAR Publishing, 2009

BAR British Series 495

For more information, or to purchase these titles, please visit www.barpublishing.com iii

A Bonfilio, mi padre, a quien recuerdo cada día. To Bonfilio, my father, whom I miss every day. 1949–2020 Va, pensiero, sull’ali dorate; va, ti posa sui clivi, sui colli, ove olezzano tepide e molli l’aure dolci del suolo natal! (Nabucco, G. Verdi.)

v

I would like to thank the Ministerio Español de Ciencia, Innovación y Universidades for funding the research project (with reference number RTI2018-098339-J-I00) that allowed me to finish this volume.

vii

Acknowledgements I have been lucky to count on the contribution of many people, who have greatly helped me finish these volumes. The first incarnation came to light in June 2013 as a PhD dissertation, after six years filled with blood, sweat and tears. During this time, my research was carried out under the umbrella of the Grupo Hiberus, and more specifically, as part of three research projects on Ancient Magic led by Francisco Marco Simón (University of Zaragoza). I am extremely grateful to him and to Attilio Mastrocinque (University of Verona) for supervising my research as well as their outstanding dedication and patience. In addition, I am also thankful to many other colleagues and friends who have helped me improve the following pages over these years. Among them, special thanks are owed to Francisco Beltrán Lloris, Gabriella Bevilacqua, Alfredo Buonopane, Richard Gordon, Chris Faraone, Miguel Ángel Rodríguez Horrillo and Roger Tomlin.

In addition, not only I, but also the reader, should be extremely grateful to my husband Ben A. Jerue (Universidad San Jorge), who translated this volume from Spanish to English, posed insightful questions and left pesky comments, which definitely improved this book, all of which he did without asking me for a divorce. Together with Ben, the love of my life, I owe gratitude to Toil & Trouble for giving me the best deadline ever to finish this project. Zaragoza, 18 November 2020*

I also owe my deep gratitude to all those who have been cheering and helping me along at the different institutions where I have been lucky enough to work or research (the Universities of Zaragoza, Verona, Oxford, Padua, Basque Country, Complutense of Madrid, Heidelberg, Yale and Chicago, together with the Escuela Española de Historia y Arqueología at Rome, the Hardt Fondation and The Center for Epigraphical and Palaeographical Studies at The Ohio State University). Sincere thanks also go out to all the curators and personnel of the museums and laboratories that I have visited to carry out the (notso-morbid) ‘autopsies’ of many curse tablets (located in Madrid, Ampurias, Barcelona, Tarragona, Sagunto, Cuenca, Cordoba, Seville, Cadiz, Nantes, Verona, Este, Cremona, Florence, Bologna, Reggio Emilia, Perugia, Rome, Naples, Pompeii, Budapest, Bad Kreuznach, Worms, Mainz, London and Bath). I am very thankful to my friends who helped with the completion of this volume through their precious time and company. Last, but not least, I am extremely grateful to my family for their unconditional love and support and to my ancestors for providing me with possibilities that they could never have dreamt of. As I have talked about with my deeply missed father, I feel very lucky for having had the opportunity of benefiting from lifelong learning, especially when one of my great grandparents, Amado, was illiterate his whole life, and one of my grandfathers, Bonfilio, only learned to write and read in prison (1939 was not the best year for democrats and leftists here in Hispania). It is thanks to their effort and hard work that I have had the chance to continue with my studies, without ‘liquidizing’ my brain, as Anuncia, one of my grandmothers, had feared.

This book was last proofread in April 2022, after many virtual trips back and forth between Zaragoza and Oxford. Any mistake remains my responsibility (and should be blamed on my tired eyes). I thank Ruth Fisher and Lisa Eaton from BAR Publisher and eVC-Tech team for all their efforts and patience to make this happen.

* 

ix

Contents VOLUME I List of Figures................................................................................................................................................................... xiii List of Tables...................................................................................................................................................................... xv I. Prolegomena.................................................................................................................................................................... 1 1. A Brief History of Previous Scholarship...................................................................................................................... 3 2. A Definition..................................................................................................................................................................... 5 3. Media............................................................................................................................................................................... 7 3.1 Introduction................................................................................................................................................................ 7 3.2 Specific Media............................................................................................................................................................ 7 3.3 Non-specific media.................................................................................................................................................. 12 3.3.1 Perishable non-specific media......................................................................................................................... 12 3.3.2 Non-perishable non-specific media................................................................................................................. 12 4. Inscribing the Defixiones............................................................................................................................................. 15 4.1 Introduction: the written and spoken word.............................................................................................................. 15 4.2 The act of inscription............................................................................................................................................... 16 4.2.1 The text: layout, language and writing............................................................................................................ 16 4.2.1.1 The layout............................................................................................................................................ 16 4.2.1.2 Language and writing.......................................................................................................................... 19 4.2.2 Other components linked to the text............................................................................................................... 22 4.2.2.1 Imagines magicae................................................................................................................................ 22 4.2.2.2 Magical symbols and charaktêres....................................................................................................... 25 4.2.2.3 Other voces magicae............................................................................................................................ 26 4.3 Formulae defigendi.................................................................................................................................................. 28 5. The Manipulation of Defixiones.................................................................................................................................. 31 5.1 Introduction: technical considerations..................................................................................................................... 31 5.2 Folded tablets........................................................................................................................................................... 31 5.3 Rolled tablets............................................................................................................................................................ 32 5.4 Pierced tablets.......................................................................................................................................................... 33 5.5 Other modi operandi................................................................................................................................................ 35 6. Deposition Contexts: Where the Curses have been Discovered............................................................................... 37 6.1 Introduction: putting ‘context’ into context............................................................................................................. 37 6.2 Funerary contexts..................................................................................................................................................... 37 6.2.1 Geographical and temporal data about funerary deposits............................................................................... 38 6.2.2 The various modi operandi............................................................................................................................. 39 6.3 Aquatic contexts....................................................................................................................................................... 41 6.3.1 Geographical and temporal data about aquatic deposits................................................................................. 42 6.3.2 Aquatic contexts as sacred spaces................................................................................................................... 42 6.4 Sacred contexts........................................................................................................................................................ 44 6.4.1 Geographical and temporal data about deposits in sacred contexts................................................................ 44 6.4.2 Spaces for communicating with the divine..................................................................................................... 45 6.5 Spaces associated with the victim............................................................................................................................ 48 6.5.1 Geographical and temporal data about deposits closely associated with the victim....................................... 49 6.5.2 Domestic spaces.............................................................................................................................................. 50 6.5.3 The agonistic sphere........................................................................................................................................ 50

xi

Sylloge of  Defixiones from the Roman West 7. Categorization of the Defixiones................................................................................................................................. 53 7.1 Introduction.............................................................................................................................................................. 53 7.2 A tacit agreement...................................................................................................................................................... 53 7.3 Erotic defixiones....................................................................................................................................................... 55 7.4 Agonistic defixiones................................................................................................................................................. 56 7.5 Juridical defixiones................................................................................................................................................... 58 7.6 Defixiones against thieves........................................................................................................................................ 59 7.6.1 A different approach: prayers for justice?....................................................................................................... 59 7.6.2 Defixiones against thieves (or in fures).......................................................................................................... 61 8. The Pantheon of Deities Invoked................................................................................................................................ 63 8.1 Introduction.............................................................................................................................................................. 63 8.2 Graeco-Roman gods and supernatural entities......................................................................................................... 63 8.3 Indigenous divinities................................................................................................................................................ 65 8.4 Oriental powers........................................................................................................................................................ 66 9. Distribution of Defixiones in the Roman West.......................................................................................................... 69 9.1 Introduction.............................................................................................................................................................. 69 9.2 Distribution of defixiones during the Republic........................................................................................................ 70 9.3 Defixiones from the High Empire............................................................................................................................ 72 9.4 Defixiones during Late Antiquity............................................................................................................................. 75 9.5 Overall distribution of defixiones in the Roman West.............................................................................................. 76

VOLUME II II. Sylloge........................................................................................................................................................................... 81 Note to the reader........................................................................................................................................................... 83 Diacritical signs............................................................................................................................................................. 84 Location......................................................................................................................................................................... 85 Sylloge of defixiones from the Roman West (SD)......................................................................................................... 86 1. Italia........................................................................................................................................................................... 89 2. Hispania................................................................................................................................................................... 199 3. Galliae...................................................................................................................................................................... 229 4. Britannia.................................................................................................................................................................. 265 5. Germania.................................................................................................................................................................. 385 6. Raetia....................................................................................................................................................................... 437 7. Noricum................................................................................................................................................................... 445 8. Pannonia.................................................................................................................................................................. 449 III. Bibliography............................................................................................................................................................. 463 IV. Appendices................................................................................................................................................................. 495 1. Table of correspondences......................................................................................................................................... 497 2. Tables Prolegomenon............................................................................................................................................... 509 3. Plates........................................................................................................................................................................ 517 V. Indices.......................................................................................................................................................................... 567 Index of Deities and Supernatural Beings.................................................................................................................... 569 Select Index of Passages.............................................................................................................................................. 571 Index of Places............................................................................................................................................................. 573

xii

List of Figures Figure 2.1. Two daemones shackling a charioteer................................................................................................................ 5 Figure 3.1. Curse tablet made of lead and papyrus............................................................................................................... 8 Figure 3.2. Curse tablet on stone.......................................................................................................................................... 8 Figure 3.3. Different shapes of lead tablets: irregular......................................................................................................... 10 Figure 3.4. Different shapes of lead tablets: tabula cerata................................................................................................. 10 Figure 3.5. Different shapes of lead tablets: disc................................................................................................................ 10 Figure 3.6. Different shapes of lead tablets: tabula cum capitulo...................................................................................... 10 Figure 3.7. Different shapes of lead tablets: tabula ansata................................................................................................. 11 Figure 3.8. Non-specific media: tegula from Thysdrus/El Jem.......................................................................................... 13 Figure 3.9. Non-specific media: curse on a pewter plate from Aquae Sulis/Bath.............................................................. 14 Figure 4.1. Defixio from Lincolnshire................................................................................................................................. 15 Figure 4.2. Different types of right-to-left layouts: Fectio/Vechten.................................................................................... 18 Figure 4.3. Different types of right-to-left layouts: Colonia/Cologne................................................................................ 18 Figure 4.4. Different types of right-to-left layouts: Celti/Peñaflor..................................................................................... 18 Figure 4.5. DT 233: ‘Fence’-style layout............................................................................................................................ 18 Figure 4.6. Detail from the Aratispi/Villanueva del Cauche defixio................................................................................... 22 Figure 4.7. Imagines magicae: daemon Baitmo Arbitto..................................................................................................... 23 Figure 4.8. Imagines magicae: Saturnius............................................................................................................................ 24 Figure 4.9. Imagines magicae: tied charioteer.................................................................................................................... 24 Figure 4.10. Imagines magicae: victim encircled by snakes.............................................................................................. 24 Figure 4.11. Charaktêres attested more than 15 times........................................................................................................ 26 Figure 5.1. Folded defixio................................................................................................................................................... 32 Figure 5.2. Rolled defixio.................................................................................................................................................... 33 Figure 5.3. Defixio rolled around a bird bone..................................................................................................................... 33 Figure 5.4. Pierced defixio.................................................................................................................................................. 34 Figure 5.5. Pierced defixio from Pompei/Pompeii.............................................................................................................. 34 Figure 5.6. Cut defixio from Ateste/Este............................................................................................................................. 35 Figure 5.7. Melted defixio................................................................................................................................................... 35 xiii

Sylloge of  Defixiones from the Roman West Figure 7.1. Erotic defixio of attraction, in which charaktêres and various individuals can be seen (two of which appear to be kissing).................................................................................................................................... 55 Figure 7.2. DT 246: a servant characterized as Mercury Psychopomp checks the death of another gladiator in the arena......................................................................................................................................... 57 Figure 9.1. Defixiones dating to the Republican period...................................................................................................... 70 Figure 9.2. Distribution of defixiones during the High Empire.......................................................................................... 73 Figure 9.3. Distribution of defixiones during Late Antiquity.............................................................................................. 75 Figure 9.4. Distribution of defixiones in the Roman West.................................................................................................. 77

xiv

List of Tables Table 3.1. Materials used to make specific media................................................................................................................ 8 Table 3.2. Non-specific media used for writing defixiones................................................................................................. 12 Table 4.1. Defixiones grouped by provenance and their characteristics............................................................................. 17 Table 5.1. Defixiones classified according to provenance and type of manipulation......................................................... 32 Table 6.1. Defixiones discovered in funerary contexts....................................................................................................... 38 Table 6.2. Defixiones discovered in aquatic contexts......................................................................................................... 43 Table 6.3. Defixiones discovered in sacred spaces.............................................................................................................. 45 Table 6.4. Defixiones discovered in contexts closely associated with the victim............................................................... 49 Table 7.1. Defixiones classified according to provenance and typology............................................................................. 54 Table 8.1. Defixiones invoking a deity according to provenance....................................................................................... 64 Table 8.2. Table containing the supernatural beings that are invoked on more than one occasion in the curses from the Roman West.................................................................................................................................... 64

xv

I. Prolegomena

1

1 A Brief History of Previous Scholarship* Despite the fact that traditional historiography long spurned the study of ‘aggressive magic’ and branded it as an aspect of Graeco-Roman culture that was unworthy of serious study, defixiones have nevertheless attracted a good deal of scholarly attention when compared to other facets of ancient magic. While some scholars have identified the origin of the modern study of defixiones with the publication of isolated Greek tablets that appeared in 1796 and 1813,1 we can look even further back to 1737, when the priest and scholar A.F. Gori published the second volume of his Museum Etruscum exhibens insignia veterum Etruscorum Monumenta, in which he included an Etruscan defixio from Volterra.2 During the nineteenth century and especially from 1840 onwards, the discovery of new tablets precipitated the publication of a large number of isolated studies.3 It was not until later, and in conjunction with a series of groundbreaking archaeological discoveries in Cnidus and Cyprus, that these artefacts increasingly came to capture the scholarly imagination. Nevertheless, the publications of these corpora by C.T. Newton and L. Macdonald respectively,4 only gave a taste of what was to come.

Just one year later, Wünsch published a collection of the 48 so-called ‘Sethian’ defixiones from Rome, discovered in a columbarium near Porta San Sebastiano, which are largely written in Greek (cf. Wünsch 1898). Six years later, in 1904, Audollent published the third magisterial collection of curses, his Defixionum tabellae quotquot innotuerunt tam in graecis Orientis quam in totius Occidentis partibus praeter atticas in ‘Corpore Inscriptionum Atticarum’ editas (DT for short), in which he collected and edited 305 defixiones, the majority of which were written in Greek and Latin, though he included curses in Oscan, Etruscan, Iberian and Phoenician. Thus, with the appearance of Wünsch’s and especially Audollent’s volumes, the study of defixiones was forever transformed from being an occasional curiosity to an established and coherent corpus of inscriptions. This work proved to be the foundation for twentieth-century scholarship, which included the publication of numerous new finds as well as various studies that dealt with particular aspects of the defixiones. Among the latter are the linguistically focused work of M. Jeanneret6 and that of M. Besnier,7 who sought to improve the published readings of the defixiones that had been edited between 1904 and 1920.

The systematic investigation of ancient cursing practices, however, did not reach maturity until the very end of the nineteenth century and beginning of the twentieth, when three monumental corpora of defixiones edited by R. Wünsch and A. Audollent appeared. In the first, Wünsch studied and edited 220 curse tablets from Attica, which were published in 1897 as part of Inscriptiones Graecae [III, 3], where the author also provided the briefest of treatments of the other Greek and Latin curses that were known at that time.5

In the 1960s, two compilations were published that deserve special attention: the first, published by E. García Ruiz in 1967, is largely focused on linguistic questions and examines 100 curses;8 in the second, H. Solin published an edition of a new defixio from Ostia (cf. 53) with an appendix containing a list of 38 tablets published between 1920 and 1968.9 The number of publications dealing with individual defixiones continued to grow, thus leading to the need for a more systematic approach to the study of curse tablets. In an attempt to address the growing problem, Solin and D.R. Jordan announced in print their intention to compile a new authoritative corpus of defixiones. Nevertheless, this work never came to fruition,10 perhaps because of the

*  In the following pages, bold numbers refer to curse tablet entries. For ancient sources, abbreviations follow A. Spawforth and S. Hornblower (eds), Oxford Classical Dictionary4 (Oxford 2012). 1  Preisendanz (1930: 119–20) has argued that the first study of defixiones should be placed in 1796 when Ignarra published a discussion of a tablet found in Italy. Jordan (1990: 440), however, has rejected the identification of this item as a curse and hence dated the beginning of modern research on the topic to 1813, when Ǻkerblad published a Greek defixio discovered in a grave near Athens. Either way, and as the following discussion shows, the systematic study of the phenomenon did not begin until later. 2  Gori 1737: 404. Cf. 87. 3  The original publication of the following tablets belongs to this period: 1–4, 56, 61, 69–70, 85, 92–93, 150–55, 461, 520, etc. 4  The 14 tablets from Cnidus, all written in Greek, were still folded or rolled when found in the temenos of the sanctuary of Demeter. Several of them were written by women who had been robbed. For a discussion, see Newton 1863: 382f.; 719–45; Gager 1992: no. 89; Faraone 2011. The Cyprian defixiones were found in what appears to have been a common grave. The published texts from Cyprus (only 22 of 260!) mostly belong to the group of juridical defixiones. In these texts, Greek and Oriental deities are invoked in order to silence those who planned to testify against the defigens in court; for a discussion, see Macdonald 1891; Gager 1992: nos 45–46; Wilburn 2013: 169–218. 5  Cf. Wünsch 1897. Currently, J. Curbera is re-editing this corpus; for a preliminary notice, see Curbera 2012.

See Jeanneret 1916 and 1917. In both articles, he carried out a comparative study of 125 curse tablets (specifically, the 103 Latin defixiones collected by Audollent in his corpus as well as another 22 items that were mostly published between 1904 and 1917). 7  See Besnier 1920, where the author compiled the Latin defixiones that Audollent had not included in DT (specifically those published by Olivieri in 1899) as well as all curses published between 1904 and 1920. This tallied up to a total of 61 additional texts. Some of these tablets were identified as defixiones, though subsequent work has shown that in fact they were not (e.g., the lead labels CIL XI, 6722, no. 1, 3–12, 14 and 17, which I examined in the year 2010 at the Museo Archeologico Nazionale dell’Umbria, concluding that they were not curse tablets). 8  Cf. García Ruiz 1967. 9  See Solin 1968: 23–31, where 44 Latin defixiones, three Greek curse tablets (nos. 40 and 44–45) and two phylaktéria (nos 3 and 43) are collected and arranged by provenance. 10  See the note published by Susini 1973: 139. 6 

3

Sylloge of  Defixiones from the Roman West sheer number of new discoveries during the last third of the twentieth century. In 1979, just six years after these scholars announced their ambitious project, two large and extremely important caches of defixiones were discovered in the British sanctuaries of Sulis Minerva (Aquae Sulis/ Bath) and Mercury (Uley), where 130 and 140 curses were found, respectively. These two finds alone drastically increased the project that Solin and Jordan had intended to undertake. Fortunately, R.S.O. Tomlin has dedicated years to the study and publication of these large and important collections.11 Only four years after Tomlin’s masterful edition of the curses from Aquae Sulis/Bath, J.G. Gager edited a book containing an interesting selection of mostly Greek curses, which brought the topic to a wider audience.12

Latin defixiones, with a special interest in pragmatics.16 In a different vein, G. Németh has published 86 sketches that Audollent made while editing the North African tablets for DT.17 These sketches, which are housed in the regional archive of Puy-de-Dôme, had previously been unpublished and unedited. Accordingly, Németh’s work has revealed a wealth of new details concerning the layout and iconography of these curses. Most recently, A. Alvar has published a study of the magical practices employed by slaves in the Roman world, while D. Urbanová has recently written a monograph on Latin defixiones, which seeks to distinguish the so-called ‘prayers for justice’ in a compilation of 309 curse tablets.18 This brings us to the present sylloge, which follows in Audollent’s footsteps and collects 535 defixiones written in Latin, Oscan, Etruscan, Gaulish and Celtic from the Roman West. Traditionally, much of the scholarly effort has been dedicated to the study of the formulae and linguistics of the tablets. Nevertheless, in the present volume, these inscriptions are studied with a particular emphasis placed on the defixiones’ archaeological and cultural contexts. Far from being monolithic, the practice of writing curse tablets changed and evolved over a millennium in the area that would become the Roman West. Recognizing this fluidity, this book aims to be a trustworthy source for scholars interested in the topic, offering not only an overview of the phenomenon but also an updated and reliable collection of texts.19 With this purpose in mind, and unlike the majority of scholars working on defixiones after Audollent, I have directly examined the texts whenever possible while compiling this sylloge. Luckily, most curse tablets that were discovered long ago remain legible,20 and in many cases an autopsy (i.e., an in-person examination) has yielded new results that can improve our readings and understandings of the corpus.21 To conclude, the ambition of this sylloge is not to be just another compilation of texts, but rather a tool that clearly presents the evidence and is capable of generating further interest in this fascinating topic.

But happily this was not the end, since then extraordinary discoveries continued: in 1999, excavators unearthed two more large caches of defixiones at both the sanctuary of Isis and Magna Mater (Mogontiacum/Mainz) and the fountain of Anna Perenna (Rome), both of which have been edited by J. Blänsdorf.13 Though smaller than the British collections, these two discoveries have proven exceedingly important. These large discoveries have undoubtedly revolutionized and reinvigorated the study of defixiones, which have been the object of various studies and research projects since 2000. Notable scholarship includes the work of J. Tremel, who has studied agonistic curses, E. Eidinow, who has focused on Attic defixiones from a psychological perspective, and S. Sichet, who has collected the curses from North Africa and studied them in conjunction with the particular magical practices of their social and geographical context.14 B. Mees has analysed (though not always exercising an appropriate level of caution) the Celtic and Gaulish curses from Britain and Gaul, while F. Murano’s masterful study has greatly improved our understanding of the Oscan curses.15 In a study of nearly 400 tablets, A. Kropp has provided a more global perspective on ancient cursing practices, paying attention to the language employed in 11  For Bath, see Tomlin’s masterful edition (1988a). Of the 140 tablets from Uley, which are generally in worse condition, see Tomlin’s brief publication (1993a) as well as the individual editions of various curses which are annually published in the journal Britannia (for references, cf. 355–73). Currently, Tomlin is preparing a monograph dedicated to the curses from Uley. 12  Cf. Gager 1992. The curse tablets, which are translated into English (no original Greek or Latin), are organized by content and accompanied by a brief commentary with relevant bibliography. The only ‘downside’ of the volume is that the editor focuses disproportionately on Greek texts, giving short shrift to Latin curses. 13  For the Mainz collection, see the magisterial edition of Blänsdorf 2012a. For the fountain of Anna Perenna (in general), see Piranomonte 2002 and 2015. Several of the curses have already been published by Blänsdorf in various publications (for references, cf. 19–47). The whole collection is the subject of a forthcoming monograph. 14  Tremel 2004 provides a collection of 100 tablets, which were written in both Greek and Latin and were mostly directed against charioteers, gladiators and venatores); Eidinow 2007 examines a corpus of 170 Greek curse tablets and compares them with the oracular questions and responses from the Dodona oracle; Sichet 2000: 865–939 collects 120 tablets written in Greek and Latin from North Africa. 15  Mees 2009 chapters 1–6 focuses on the Gaulish and Celtic inscriptions (both curses and other types of inscriptions); for the Oscan curse tablets, see Murano 2013.

16  See Kropp 2008. Her catalogue lists 578 tablets, 391 of which are subject to linguistic analysis. 17  See Németh 2013, which includes neither a reading nor transcription of the texts. 18  See Alvar 2017 and Urbanová 2018 (which is an English translation of the original in Czech, published in 2014). 19  Although the sylloge does not include curses published after summer 2018, new scholarly discussions of previously published texts have been included in the commentary and bibliography. For a fuller explanation of the inclusion criteria for the sylloge, see the Note to reader under section II. 20  Contra Urbanová 2018: 13. 21  For the tablets that I have examined personally, not only have many readings of some texts been improved, but autopsy has also provided new important details about the curse’s layout or iconography. Furthermore, this meticulous process has allowed me to discover that certain artefacts that have previously been classified as defixiones have actually been misidentified. Just to mention some examples, in addition to the lead tags included in Besnier 1920: nos 40–50 and 52 (cf. note 7), see also another label currently housed in Florence. The text, considered by Besnier 1920: no. 52, Kropp 2008: dfx 1.1.1/2 and Urbanová 2018: no. 2 as a defixio, is actually a label of an officina plumbaria, whose text reads: M(arci) Ponti Secundi oficina plum(baria) (cf. Paolucci 1994: 106–07, contra Gordon 2019b: 423).

4

2 A Definition According to Philoxenus, the Latin lexeme defixio, derived from the verb defigere, is equivalent to the Greek κατάδεσμος.22 Both of these terms refer to a popular practice in Antiquity, namely, cursing. Documented in the epigraphic and literary records,23 these two terms allude to a magico-religious practice in which an unfortunate individual inescapably became the victim of a spell (see Figure 2.1).

Within epigraphical discourse and debate, defixiones have traditionally been considered a subset of ‘private’ epigraphy. This category is meant to describe inscriptions that were written by private individuals with an ephemeral purpose in mind and were destined for a small audience (e.g., wax tablets, graffiti or texts on instrumentum domesticum). Accordingly, the survival of such texts is the happy result of their being inscribed on non-perishable materials that happened to be preserved almost by accident.28 There are numerous factors that have led to defixiones being categorized as a sort of ‘private’ epigraphy. First off, these texts are certainly hard to classify as ‘public’, since cursing was an illegal act from at least the time of the XII Tables (VIII, 1a, 1c and 4, cf. section I.9.2 below). As a result, these texts were rather secretive as can be seen in several of their common features: not only were the texts themselves at times designed to be inscrutable through the use of coding devices, but they were also deposited in hidden places away from public view that were widely believed to help in their activation (e.g., graves, wells or sanctuaries). In addition, the intended ‘readers’ of these texts were certainly not a wide human audience, but rather were the deities invoked.

In 1985, D.R. Jordan provided the most widely accepted definition of a defixio: ‘Defixiones, more commonly known as curse tablets, are inscribed pieces of lead, usually in the form of small, thin sheets, intended to influence, by supernatural means, the actions or welfare of persons or animals against their will.’24 Jordan’s definition can be further refined to stress that the true essence of a defixio is found neither in the material used25 nor in the inscribed message, but rather in its harmful nature and desire to manipulate. Every curse is born from rancour, envy, anger, fear, desire, love or desperation and seeks to accomplish what a defigens (the person who writes or commissions a curse tablet) cannot do through the available means, due to a lack of knowledge, power or economic/legal resources.26 A curse, then, is a strategy for obtaining ‘individual justice’, which was illegal in Rome and feared throughout ancient societies. Indeed, Pliny captures the widespread fear of being cursed when pointing out that defigi quidem diris precationibus nemo non metuit, that is ‘there is no one who is not afraid of being enchanted with malevolent prayers’.27

Up to this point, defixiones appear to conform to the category of ‘private’ epigraphy. That said, we ought to nuance this view and consider one final and fundamental aspect of curse tablets, that is, their ‘expiration’ date. While private texts are characterized by having ephemeral messages, we can say that the opposite is the case with the messages contained in defixiones. As is the case with public epigraphy, curse tablets were intended to last in time and have an enduring impact: during the process of manufacture, which included a performative ritual, the practitioners would ‘activate’ the spells contained in the tablet. Once activated, these texts were believed to have a nearly inexhaustible agency and to establish a lasting bond between spell and victim that, on certain occasions, could

Philoxenus, Glosas, Lat II, 40. Among the literary sources, we find Pl. (Resp., 364b; cf. n. 27) as well as some passages from the PGM, such as IV, 336 (παρακατατίθεμαι ὑμῖν τοῦτον τὸν κατάδεσμον, θεοῖς χθονίοις...: ‘I entrust this binding spell to you, chthonic gods...’ [translation by E.N. O’Neil, apud Betz 19922]). Defixio in the ablative (cum defictcsione) has been recently attested in a curse tablet against thieves from Abusina/Eining (Raetia) dated to the first century CE (see Blänsdorf 2019, no. 1). For just a few of the examples of the verb defigere in the tablets, see 105 (l. 8: defigo illos quo pereant), 345 (l. 4: Tacita deficta), 338 (A, l. 1ff: T(itus) Egnatius Tyran(n)us deficus est), etc. 24  Jordan 1985a: 151. For a general definition of curse tablet, see also Bailliot 2010: 71f.; Martin 2010: 9–11; Edmonds 2019: 53–64; Eidinow 2019: 351f.; Watson 2019: 57–63. 25  As the author himself pointed out in 2001: 6. 26  In this volume, the terms defigens, principal, practitioner, author, scriptor (‘writer’) are used interchangeably to refer to the person who was writing the curse tablet. 27  Pliny, HN XXVIII, 19 (translated by Bailliot 2019: 190, n. 79). As is well known, this fear affected not only the lower classes but also elites. The widespread nature of this fear is affirmed through examples of ‘famous victims’ of aggressive magic, such as Germanicus (Tac., Ann. II, 69 and Dio Cassius LVII, 18; see section I.6.5 and n. 323–24), the orator Libanius (Libanius, Orat. I, 249; cf. section I.6.5 and n. 325), or the doctor Theodorus of Cyprus (see Sophronius, Narratio, LV; see section I.6.5 and n. 327). To these, we must add the victims of φαρμακεία or venenum that are documented in a series of inscriptions collected by Graf (2007). 22  23 

Figure 2.1. Two daemones shackling a charioteer (From SV 29).

28 

5

Here I follow the definition of Beltrán Lloris 2014a: 89.

Sylloge of  Defixiones from the Roman West Roman West there is no evidence for such professionals writing curses until the second or third century CE in North Africa39 and Late Antiquity in Italy and other provinces.40 Until these later dates, defixiones (with rare exceptions) were written by aggrieved individuals with different levels of magical-religious savoir faire, thus attesting to what R. Gordon has defined as a type of ‘vernacular’ practice.41 The archaeological record has provided good evidence of defigentes with different levels of knowledge and skills: one the one hand, some deposited pseudo-inscriptions (or even uninscribed texts) in specific contexts (thus demonstrating knowledge of the basic mechanics of how curse tablets worked),42 while, on the other hand, archaeologists have uncovered quite elaborate (and at times almost literary) texts. On the whole, the palaeographical, linguistic and textual analysis of the vast majority of curse tablets demonstrates that writers not only knew the ABCs of cursing but also were sufficiently literate to write curses on their own behalf.43 Obviously, in all of these cases (whether tablets were inscribed or not), we cannot forget that there was an important oral component of cursing as well as a ritual action. In most cases, these aspects have been lost to us, but other forms of evidence, such as the PGM, remind us of their centrality in cursing rituals.

not be broken until a conflict was settled to the defigens’ satisfaction.29 The use of a particular materials and deposit context that were isolated or difficult to access ought to be analysed in light of a defixio’s lasting nature: such media and locations were chosen to prevent a spell from being deactivated.30 Given these considerations about the nature and purpose of curse tablets, the traditional binary distinction between public and private epigraphy does not properly capture the social and magical-religious characteristics of defixiones. Traditionally, the spread of defixiones has been closely associated with the processes of territorial expansion and ‘Romanization’.31 This view of a practice simply spreading and being adopted by new populations, however, can be misleading: curse tablets, which were in use for more than a millennium in a wide geographical area, should be seen as a ‘living’ magical-religious technology, which developed and evolved alongside many other aspects of ancient life. In this regard, we should heed to words of Owen Davies, who has claimed, ‘artefacts have life histories that need to be considered and contextualized in order to understand better their meaning, and the societies that employed them, at any point in time.’32 While Davies wrote these words to describe domestic magical practices in modernity, his description is completely suitable for understanding the different ways that curses were used in different ancient contexts. Indeed, defixiones were adopted and adapted by different provincial societies, whose unique characteristics are reflected in their language,33 onomastics,34 the deities invoked,35 their manipulation36 and depositional contexts.37 All of this precludes us from understanding the use of defixiones as a monolithic practice that was passively taken up by new groups and populations. In fact, it is possible to distinguish extraordinarily rich and meaningful regional variations across the Roman West, not to mention the observable changes in the same geographical region over time.

With the exception of North Africa, where we can already detect the presence of magical professionals in the Imperial Age, in the rest of the Roman West there is no generalized evidence for their presence until Late Antiquity. In this period we can detect a new and updated version of this magical technology that originated in Graeco-Egyptian magical practices. At this point, there were authentic magical professionals who specialized in techniques (e.g., the use of charaktêres, iconography or voces magicae). In the same period, there is still evidence for individuals writing their own curses, though to a lesser extent than before.

Another fundamental question is who exactly wrote these texts? While a passage from Plato tantalizingly refers to ‘beggar-priests and prophets’ and suggests that there were magical professionals already in classical Athens,38 in the

incantations [epôidais], to cure with pleasures and festivals any wrong done by the man himself or his ancestors, and that they will harm an enemy, a just man or an unjust man alike, for a small fee, if a man wishes it, since they persuade the gods, as they say, to serve them, by certain charms [epagôgais] and bindings [katadesmois]’ (tranlsated by Ogden 2002: no. 14). 39  The existence of magical professional is supported by series of tablets, whose palaeography, iconography and/or use of charaktêres and voces magicae are identical. For some good examples, see DT 276–85 from Hadrumetum (Németh 2011 has identified a possible officina magica), DT 286–91 (all of which invoke the demon Baitmo Arbitto; cf. Gordon 2005), etc. 40  See the defixiones and containers discovered in the fountain of Anna Perenna (cf. 19–47 and Faraone 2005). 41  See Gordon 2019b. The model that he puts forth for ‘vernacular’ and professionally written curse tablets in Italy could easily be applied to the rest of the Roman West. On this differentiation between ‘vernacular’ and ‘professional’ texts, see also Faraone and Gordon 2019: 320. 42  For instance, see the pseudo-inscriptions from Bath (cf. 317–21) and the uninscribed caches from Roccagloriosa (cf. 76), Rom (cf. 159) or Uley (cf. 355), which were found in the same context as inscribed curses. 43  It is important to note that within the corpus from the sanctuary in Bath (where 120 curses have been discovered), only two tablets appear to have been written by the same author (see 300–01), which attests to the existence of 119 known scriptores at the sanctuary.

29  On this point, see section I.4.1. For the ‘agency’ of these texts, see Sánchez Natalías 2018: 10. 30  The act of deposition was accompanied by the recitation of certain formulae that were intended to help preserve the curse, as we learn, for instance, from PGM VII, 453ff. On this, see section I.4.1, n. 98. 31  On this controversial concept, see Beltrán Lloris 2017. 32  Davies 2015: 394. 33  Cf. section I.4.2.1.2. 34  For just several examples, see those listed in section I.9.3, n. 576. 35  Cf. section I.8. 36  Specifically, see section I.5. 37  See section I.6. 38  On this point, Pl. (Resp., 364b) provides the best known passage: Ἀγύρται δὲ καὶ μάντεις ἐπὶ πλουσίων θύρας ἰόντες πείθουσιν ὡς ἔστι παρὰ σφίσι δύναμις ἐκ θεῶν ποριζομένη θυσίαις τε καὶ ἐπῳδαῖς, εἴτε τι ἀδίκημά του γέγονεν αὐτοῦ ἢ προγόνων, ἀκεῖσθαι μεθ’ ἡδονῶν τε καὶ ἑορτῶν, ἐάν τέ τινα ἐχθρὸν πημῆναι ἐθέλῃ, μετὰ σμικρῶν δαπανῶν ὁμοίως δίκαιον ἀδίκῳ βλάψει ἐπαγωγαῖς τισιν καὶ καταδέσμοις, τοὺς θεούς, ὥς φασιν, πειθοντές σφισιν ὑπηρετεῖν: ‘Beggar-priests [agurtaí] and prophets [manteis] go to the doors of the rich and persuade them that they have the power, acquired from the gods by sacrifices [thusiais] and

6

3 Media44 44 

3.1. Introduction

Unsurprisingly, though there are a few rare exceptions, we know about the use of perishable materials through the literary record. Specifically, the Greek Magical Papyri (PGM) provide recommendations for materials like papyrus, which was seen as especially apt for writing erotic spells.47 The papyrus to be employed ought to be hieratic (that is, of the best quality)48 and new (not reused).49 In a spell directed against an enemy or a woman, for instance, the PGM recommend the use of χάρτην ἱερατικὸν ἢ μολυβοῦν πέταλον (V, 304–5), that is, giving papyrus and lead as equally valid choices. Interestingly, these two materials are found together in an unparalleled curse tablet from Carthage (see Figure 3.1). In this instance, a sheet of lead was inscribed with four lines of Latin text and folded in such a way that the text would be exposed. Preserved within the folded tablet, archaeologists found a papyrus fragment, which, unfortunately, was so deteriorated that it could not be studied. 50

In 1933, Audollent passed judgement on the relative importance of the text of defixiones and their material characteristics, claiming that ‘[c]ertes l’aspect de ces modestes documents importe beaucoup moins que leur contenu. Il n’est cependant, en plus d’un cas, nullement négligeable’ (1933a: 31). In line with Audollent’s appraisal, scholarship has traditionally paid relatively little attention to the study of defixiones from a material point of view that takes the curse’s media into account: the most recent large-scale discussion of curse tablets (Urbanová 2018), for instance, dedicates a scant page to the materiality of these artefacts.45 Despite this general trend, a volume edited by G. Bevilacqua (2010) has sought to analyse the relationship between writing, media and magic in Antiquity and to consider a wide range of objects including some curses, most of which come from a Greek context.46 In a parallel vein, a major concern of the present volume is to draw more attention to the material features of this corpus of curse tablets.

When it comes to the use of non-perishable materials, we can create further subdivisions: those made of stone and those made of metal. Although the first category is never mentioned in the literary record, at least five examples have been preserved archaeologically. Two of them, which come from Braga (146–7) and date to the fourth or fifth century CE, were written on slate plaques. While we do not know the circumstances that led to these curses being written, we must note that slate was a commonly used material in Visigothic epigraphy and began to be used in throughout Hispania in Late Antiquity (cf. Velázquez 2004). The other two examples of curses inscribed on stone (see Figure 3.2) come from Pompeii (72)51 and Emerita Augusta/Merida (120), the first on a piece of slate, the second on a piece of marble. Each curse betrays careful craftsmanship as well as a carefully arranged layout, and

To analyse the materiality of the defixiones from the Roman West, we must draw a fundamental distinction between two different types of material and media. The first contains the defixiones that were inscribed on objects intentionally created for the purpose of writing (specific media), whereas the second consists of objects inscribed on re-appropriated objects that had originally been manufactured for a different purpose (non-specific media). There are subdivisions within each of these groups based on the various materials used and whether they were perishable or non-perishable. 3.2. Specific Media This category includes all objects whose sole raison d’être was to serve as the surface on which a text was to be written. Within this category, practitioners used a range of materials that can be classified as either perishable or non-perishable (see Table 3.1).

47  Currently, the PGM are the only preserved magical handbooks from Antiquity. Consequently, and despite their late chronology, they are taken as a frequent reference point throughout these prolegomena (and the scholarship more broadly). There is no doubt that these books circulated throughout Egypt, and it is quite probable that similar compendiums also circulated throughout the Roman West, where Graeco-Egyptian magical practices were considered to be especially effective. In this regard, we should not overlook a well-known passage from Dio Cassius (LXXV, 13) in which the historian claims that after his journey to Egypt in 199–200 CE Septimius Severus collected and brought home the secret books found in Egyptian temples. According to Mastrocinque (2019), the magical papyri which we can study today are probably substitutes of the older original texts. This would explain the Late-Antique dating of most papyri from the collection. 48  PGM XI c, 1: εἰς ἱερατικὸν βιβλίον. 49  PGM IV, 78; PGM VII, 940, etc.: χάρτην καθαρόν. 50  Jordan 1996: 122. According to the editor (per litt.) ‘the papyrus, folded inside the lead, had become frayed over the years, and when I saw it I didn’t dare to try to touch it. I don’t know whether the papyrus was inscribed’. 51  This piece, which is still in situ, forms part of funerary monument 23 OS in the Porta Nocera necropolis (Pompeii).

44  ‘Medium’ (pl. ‘media’) is an imperfect English translation for the Spanish word ‘soporte’ (Italian ‘supporto’, French ‘support’, German ‘Textträger’), which is used to refer to the surface on which a text was inscribed. 45  See DT, pp. XLVII–XLIX; Cesano 1961²: 1561–62; Preisendanz 1933: 3–4; Gager 1992: 3–4; Graf 1995: 129–31; Ogden 1999: 10–13; Kropp 2008: 80–82; Bailliot 2010: 76–77, Martin 2010: 15–17; Urbanová 2018: 36. 46  See especially Bevilacqua 2010: 21–82, which includes several Latin defixiones, such as 3 and the curse from Thysdrus/El Jem (cf. Foucher 2000). For curse tablets from the Roman West, see Sánchez Natalías 2011b (where there is an earlier Spanish version of this section) and 2018 (for an in-depth study on the materiality of curses and the metaphorical connection between spell and medium).

7

Sylloge of  Defixiones from the Roman West Table 3.1. Materials used to make specific media Perishable Materials Literary record

Non-perishable Materials

Archaeological record

Literary record

Marble: 4, 120

Papyrus Papyrus and lead

Archaeological record Slate: 72, 146–47

Papyrus and lead: Carthage

Figure 3.1. Curse tablet made of lead and papyrus (From Thysdrus/El Jem. Jordan 1996: 122).

Tin

Tin: 212, 304, 318, 331

Copper

Copper: 30, 31, 45

Lead

Lead: 252, 320, 268, ex multis

Figure 3.2. Curse tablet on stone (From Emerita Augusta/ Merida [cf. 120]. Archivo Fotográfico MNAR, A. Osorio Calvo).

they belong to what Audollent called defixiones against slanderers and thieves. In both cases, the defigens opted for a material commonly used in monumental epigraphy. The choice of materials does not seem to have been influenced so much by ideas about the lasting nature of stone,52 but rather by the unusual fact that these defixiones were meant to be displayed publicly (this is undeniable in the case of the Pompeiian curse).53 Finally, a fifth example of a curse written on stone comes from Rome (cf. 4). The text was inscribed on the back of the funerary altar of the child Junia Procula for two main reasons: first, in order not to be seen (although the back side was the only blank space available); second and more importantly, in order to

invoke the deceased child to avenge the author of the curse. In addition, its composition in metre may suggest that the author was trying to hide the true nature of the inscription. Therefore, and by contrast with the inscriptions from Pompeii and Emerita Augusta/Merida, this text seems to have been written on the altar for practical reasons. In short, the use of stone is extremely uncommon and when it is used the choice of material can be explained in terms of practical considerations or as influenced by general trends in the current writing practices. Turning to metal, the prominence of lead is undeniable. That said, there are also examples, attested both in the literary and archaeological records, of the use of tin and copper as the media for a curse. The PGM, for instance, recommend the use of tin sheets for defixiones intended to destroy chariots,54 to be used as erotic curses55 and even those meant to restrain a victim.56 Metallurgic analysis of some curses from the sanctuary of Sulis Minerva in Aquae Sulis/Bath (Britannia) has shown that pure tin was

52  Giorcelli Bersani 2004: 16 maintains, ‘Si scrive perché lo scritto valga da quel momento in poi, in certi casi si scrive per l’eternità. L’incisione su un supporto duro e durevole come la pietra o il marmo, virtualmente eterno, fissava ancor di più il messaggio nella solidità della materia’. Likewise, Susini 1989: 278 affirms ‘La scrittura su materiale durevole (…) quindi su superficie concettualmente eterne, comporta alcuni effetti sul pubblico (…): 1. la persuasione dell’importanza della scrittura (…); 2. di conseguenza, il senso di sicurezza che promana dal monimentum e dalla sua scrittura, proprio perché concettualmente imperituri…’ For the idea that stone is eternal, also see Susini 1998: 105 and 108. 53  Additionally, perhaps we should add an item from Remeseiros, which was inscribed on a piece of granite that has proven difficult to read and interpret (see the note in the introduction to Hispania in section II.2).

PGM IV, 2212: ἐπὶ πλακὶ κασσιτερίνῃ. PGM VII, 459: ἐπὶ λάμνας κασσιτερίνης. 56  PGM VII, 417: εἰς πέταλων κασσιτέρινον. 54  55 

8

Media used in the creation of two curses (cf. 212 and 331) and also formed more than 90 per cent of the alloys used to make another six tablets.57 Unlike the instances from the PGM, the evidence does not suggest that the choice of tin was intentional in these British cases. As Tomlin has suggested (1988a: 82), tin and lead were both available and commonly used to produce other objects.

HN XI, 114, 275), saying that someone had a leaden hue was a not-so-ambiguous way of insinuating that death was quickly approaching. Taking this series of connotations in account as well as the principle of persuasive analogy,63 the PGM recommend that defixiones be made from lead taken from a pipe used for cold water64 or one taken from a frigidarium65 in order to symbolically ‘freeze’ a victim. In a similar vein, in a spell meant to restrain, it is specified that the lead has to be hammered while cold.66 Another type of comparison based on the principle of persuasive analogy is laid out in two spells for subjugating victims, where the use of lead from a yoke is recommended,67 since just as the yoke subjugates oxen, so too will the tablet subdue the defigens’ enemies.

When it comes to copper, Saint Jerome mentioned the use of this metal (aeris Cyprii lamina) in a somewhat controversial passage dealing with an erotic spell.58 Though the veracity of Jerome’s account can certainly be called into question, it is beyond doubt that copper was at times used in the production of curse tablets, as has been confirmed by the discovery of three curses from the fountain of Anna Perenna in Rome (cf. 30, 31 and 45), which were both deposited inside a lamp, where they were inserted in the place where you would normally expect to find a wick. It is entirely within the realm of possibility that the practitioner chose this metal intentionally, since its colour could be seen as a reference to the lamp’s flame.59

Metallurgic analyses have provided important data about the provenance and purity of lead used for cursing. L. Pintozzi’s isotopic analysis of 11 tablets discovered in the circus of Carthage has shown that the metal used came from distant mines located in places like the Ural Mountains, Cyprus and the Iberian Peninsula.68 At the present date, the only published study on the purity of lead used is abovementioned analysis of 75 of the 130 tablets from the sanctuary of Sulis Minerva in Bath.69 In the case of these British curses, only 15 of the 75 defixiones were made from an alloy of more than 66 per cent lead.70 The vast majority of tablets analysed were made from an alloy of tin and lead (or on several occasions, copper and lead) with varying proportions. This variation suggests that the alloys used for making these tablets were produced on a small scale, which is hardly surprising since the process was rather straightforward: once the alloy was made from the metals that can be smelted at low temperatures, it was cast either in a mould71 or on a flat surface to make a larger sheet from which the tablet could be cut after the metal had cooled down. Any irregularities in the sheet were generally corrected by someone who would flatten

Leaving aside tin and copper, let’s turn to lead, which was so commonly used in the production of curses that it came to be used as a metonymy for a curse tablet. As is well known, lead was one of the most commonly used materials for writing, since it was inexpensive, readily available and easy to inscribe. All these factors were instrumental in establishing lead as one of the main media for writing in Antiquity.60 As F. Graf has argued,61 once this metal was employed in the realm of magic, it was endowed with ritual significance and its properties were thought to guarantee its efficacy: associated with the malign star Saturn, lead was thought to bring misfortune and death (cf. Cesano 19612: 1561). Like a corpse, this metal was considered cold and heavy.62 According to Aristotle (Pliny 57  Cf. 318, 304, 317, 325, 326 and 306. For a discussion, see Tab. Sulis 7, 99, 101, 112–13, 120–21 and 126. 58  Jerome, vit. Hilar. XI: …et subter limen domus puellae portenta quaedam verborum et portentosas figuras sculptas in aeris Cyprii lamina defodit. This has proven controversial due to both the veracity of the deeds (see Tomlin 1988a: 81, n. 2) as well as the interpretation of the phrase aeris Cyprii lamina, which some authors have taken as copper (Tomlin 1988a: 81 and Ogden 1999: 10), while others have seen it as bronze (Gager 1992: 261 and Ogden 2002: 230). Here I follow the first possibility, since Pliny refers to copper as aes Cyprium (e.g., HN XXXIII, 29–30 and OLD s.v. aes). In this way, the ambiguity of Latin aes, which can refer to either metal, is resolved through the qualifying adjective Cyprius. On this passage, see section I.6.5.2 and n. 326. 59  In this regard, see the suggestive proposal of Mastrocinque 2007, who connects the flame to the victim’s life as part of a larger ritual of symbolic homicide. Cf. also Sánchez Natalías 2018: 11. 60  Poccetti 1999: 545–61; Graf 1995: 129; Ogden 1999: 11; Gordon 2015: 153–58. 61  Graf 1995: 129–30: ‘c’è stato chi ha affermato che in origine la scelta del piombo come supporto di questi testi fu determinata dalla natura morta e fredda di questo metallo. Ma si tratta di un’opinione ormai insostenibile... Fissare l’attenzione sul piombo e sulle sue qualità, e considerarlo come il materiale meglio adatto a far da supporto alle defissioni, sono sviluppi secondari, una ritualizzazione a posteriori di una pratica corrente’. 62  All of these characteristics are reflected in the texts of several Greek curses, where the defigens asks that the victim be left cold as lead (see DTA 67, 105, 106 and 107) or that the victim’s tongue be turned to lead (e.g., DTA 96 and 97). These metaphors, however, are not found in the

Latin curses, which employ different types of comparisons (see Sánchez Natalías 2018). 63  Cf. Tambiah 1973: 212 and section I.4.3 for the formulae that follow this principle, according to which an aspect of the actio magica is compared with the victim of the spell. 64  PGM VII, 397–98: μόλιβον ἀπὸ ψυχροφόρου σωλῆνος ποίησον λάμναν. 65  PGM VII, 432: πλάκαν ἐς μολιβῆν ἀπὸ ψυχροφόρου τόπου. 66  PGM XXXVI, 1–2: λάμναν μολιβῆν ψυχρήλατον. 67  PGM VII, 925–26 and X, 36–37, respectively: λεπίδα μολιβῆν ἀπὸ ζυγοῦ μούλων... λάμναν . 68  Pintozzi 1990: 113–33. The author argues that ‘geologically, lead is unique in that the isotopic composition of each ore source varies markedly… When the ore is produced, the lead is separated from the parent isotopes which freezes the isotopic composition and leads to a fixed composition within a given deposit… By measuring the isotopic ratio of the ore sources and then comparing those to that of the final metal product, it is possible to determine the origin of a metal artifact’ (Pintozzi 1990: 113–14). 69  See Tab. Sulis, pp. 81–84. 70  Some of these are uninscribed sheets of lead. Their inventory numbers are 690 (100%), 698 (100%), 487 (97.9%) and 20.002 (99.5) (see Tab. Sulis, p. 260). The following were inscribed: 252 (99.9%), 320 (99.9%), 268 (99.6%), 220 (98.3%), 218 (96.9%), 208 (90.5%), 279 (80.5%), 303 (73.9%), 266 (70.9%), 298 (67.4%) and 264 (67%). 71  As was the case with 300–01 and perhaps 124, the reverse of which contains mould markings in which the lead was flipped.

9

Sylloge of  Defixiones from the Roman West

Figure 3.5. Different shapes of lead tablets: disc (From Barchín del Hoyo [cf. 145], Museo de Cuenca). Figure 3.3. Different shapes of lead tablets: irregular (From unknown provenance [cf. 149]. Courtesy of P. Rothenhöfer).

Figure 3.4. Different shapes of lead tablets: tabula cerata (From Pompeii [cf. 71, author’s photograph], Ministero dei Beni e delle Attività Culturali e del Turismo-Soprintendenza Pompei).

Figure 3.6. Different shapes of lead tablets: tabula cum capitulo (From Italica/Santiponce [cf. 127, author’s photograph], Museo Arqueológico de Sevilla).

the sheet with a hammer.72 The X-ray analysis of a defixio from Groß Gerau (Germania, cf. 482) attests to another method of manufacture in which a small piece of lead was hammered into a sheet of the desired proportions to inscribe the curse.

triangles,73 resemble the outline of a foot,74 or are tabulae cum capitulo (with a single handle, see Figure 3.6)75 and even the well-known tabulae ansatae (with two handles, see Figure 3.7).76 These last two forms deserve special attention, since they are frequently found in other forms of common writing. Texts inscribed on tablets with these shapes were originally made of wood before these same shapes were used for bronze inscriptions (e.g., for the display of public documents) and then lead ones (as demonstrated by the curses discussed here).77

When discussing the use of lead, it is always important to discuss the shapes into which it was formed. While most curses were inscribed on rectangular sheets, we cannot overlook the great diversity of shapes that defixiones could take, running the gamut from curses written on irregularly shaped scraps (see Figure 3.3) to those inscribed on carefully constructed sheets. In the latter group, there are curses that resemble tabulae ceratae (writing tablets, see Figure 3.4), are shaped like discs (see Figure 3.5) or 72 

Such as 106. Such as the erotic curse from Saguntum/Sagunto (cf. 140). 75  Such as the curses from Italica/Santiponce (cf. 127) and Baelo Claudia/ Bolonia (cf. 128), in addition to the tablet from Arverni/Chamalières (cf. 163), in which the text runs longitudinally. 76  Such as 337, 463, 479 and 518, among others. 77  For tabulae cum capitulo, see Costabile and Licandro 2002: 25–34. For tabulae ansatae, see Cornell 1991: esp. 23–24. 73  74 

Clearly the case with 259, 300 and 302, among others.

10

Media

Figure 3.7. Different shapes of lead tablets: tabula ansata (From Bodegraven [cf. 463], Museum Het Valkhof, Nijmegen [NL]).

made in sets that generally include three containers of various sizes so that they can be nested like Russian dolls. In seven cases, the smallest container of the group held a small human figure, which (with only one exception)81 was placed upside-down so that the figure’s head touched the bottom of the jar. In order to preserve each group of objects, the largest container was normally closed with a conical or flat lid, which was then either sealed with natural resins or bent along the lip so as to keep the object from being opened.

The PGM do not provide any rules or guidelines about the specific shape that tablets should take; instead, they only recommend that curses should be inscribed on sheets of whatever metal is most suitable for a particular spell. To refer to lead sheets and the forms that they can take, the PGM employ the following lexemes: λᾶμνα, πέταλον, πιττάκιον, πλάτυμμα, πλάξ and λεπίς.78 In the texts of the defixiones themselves, the following Latin lexemes are used to refer to the physical object: defixio, tabella, tabula, charta and plumbum,79 without providing any further specification about the tablet’s shape. Given the silence of the PGM and the defixiones themselves about the possible shapes of curse tablets, it is reasonable to conclude that the choice was left to the practitioners, who were probably influenced by time and technical restraints (e.g., if they themselves made the tablets or had the help of a professional practitioner).

These nesting vessels have been included in the present volume since most of the innermost containers were used as the surface on which a curse was inscribed. Accordingly, on the curved surfaces of these objects, the defigens represented divine powers or the curses’ victims, whose names were handed over to the deities invoked; furthermore, in the bottom of the innermost container belonging to 26, we find a defixio inscribed on the round flat surface. The fact that the curses are found exclusively inscribed on the innermost of the three containers is clearly a sign that the practitioner sought to hide the written message (unlike curses written on lead sheets, these containers could obviously not be rolled up or folded).

Finally, special attention must be paid to a series of containers from the fountain of Anna Perenna (see 24–27 and 35–36) due to their exceptional nature.80 These were Λᾶμνα is the most frequently used lexeme cf. PGM VII, 397–98; 459; X, 36–37; XXXVI, 1–2; 231 and LVIII, 6. Πέταλον appears in PGM V, 304–05; VII, 417 and X, 36–37. For πιττάκιον, cf. PGM IV, 2956 and XV, 9. For πλάτυμμα see PGM IV, 329 and 406–07, while πλάξ appears in PGM IV, 2212 and VII, 432. Finally, λεπίς (λεπ[ίδα μολιβῆν]), has been reconstructed in PGM LXXVIII, 3–4. 79  Defixio is found (in ablative, cum defictcsione) in a curse tablet from Abusina/Eining (Raetia, cf. Blänsdorf 2019: no. 1); while tabel(l)a is found in 51 (A, l. 10 and B, l. 9); tabula in 494 (A, l. 1); chârta in 96 (A, ll. 1–2); carta in 484 (l. 11) and 367 (A, l. 1). Plumbum is used in 157 (A, l. 1) and 470 (l. 2), among others. 80  When it comes to the containers and their respective contents, Piranomonte’s numbers have not been completely consistent, but vary from 9 to 18, 21 or even 24 lead containers and from one to three terracotta containers with lead lids (compare Piranomonte 2005: 99 with 2006: 194, 2009: 257, 2010a: 204, 2010b: 23; 2012a: 167 and 2015: 79 and also Piranomonte and Marco Simón 2010: 8). Blänsdorf (2015b: 22), for his part, mentions the existence of 27 lead containers. It seems likely that the variation depends on the system used for counting the containers (i.e., whether they are counted individually or as sets of two or three). This has led to further confusion (see for instance, Urbanová 2018: 54). In any case, and given this inconsistency, it’s best to follow the first notice about the group, published by Polakova and Rapinesi (2002: 40–45 and 48–52), who specify that the number of (sets of) containers is a total of ten, of which nine are in lead and one in ceramic (Polakova and Rapinesi 2002: 39). Since it has not been possible to consult the full and yet-tobe published catalogue of the materials from the Fons Annae Perennae, 78 

M. Piranomonte has suggested that the Anna Perenna containers were reused inkwells or medicinal jars.82 Her first suggestion should be rejected for various reasons: first, inkwells were not normally fashioned from lead; second, the lids do not have the extremely characteristic double or single opening for the stilus; third, inkwells were normally produced in pairs in a standard size, not in groups of three with different sizes. According to Piranomonte’s second suggestion that the objects were originally medicinal jars, she was presumably referring to the ubiquitous pyxis (a small cylindrical container with a lid), which could the present study contains six of the nine lead containers (cf. 24–27 and 35–36). The container with inv. no. 475541 has been excluded on the grounds that it was not inscribed, even though that it did have two defixiones inside (cf. 33 and 34). 81  The exception is the figurine found inside of 26 (see Polakova and Rapinesi 2002: 41 and 48). 82  Piranomonte 2010b: 28–30.

11

Sylloge of  Defixiones from the Roman West Table 3.2. Non-specific media used for writing defixiones Perishable materials Literary record

Archaeological record

Non-perishable materials Literary record

Hellebore leaf

Shells

Donkey hide

Magnetite

Archaeological record mention in DT 234 Opus sectile: 57

Bat wings Pot for smoked fish Ceramic cup

Jar: 173

Unfired ceramics

Tegulae: El Jem (Foucher 2000) Ostrakon: 113 Cinerary urn: 16 Lamp: 3 Ceramic jar with lead lid: 526 and Rome (Polakova and Rapinesi 2002: 39, fig. 2). Tin disk: 223 Pewter plate: 235

be made from a range of materials and could come in a variety of different sizes.83 However, of all the sets from the fountain, only 24 presents a similar appearance to some pyxides, a fact that makes this argument inapplicable to the other containers from the fountain.84

Although the ritual described in PGM certainly did not have the same purpose as those from the Roman sanctuary, we can nevertheless establish a strong parallel between the two examples. Furthermore, if the Anna Perenna containers were truly miniature lead cinerary urns, these would have served the same function as the coffins from the Kerameikon: to bury the victim symbolically.

It is more likely that the containers from Anna Perenna were created ex professo as an active part of the magical ritual and hence that they served a double function: first, to provide a place where the figurine representing the victim (always found in the innermost container) could be deposited; second as a surface on which a magical spell could be written. As Piranomonte has rightly pointed out,85 the only archaeological discovery that parallels the sophisticated ritual carried out at Anna Perenna comes from the Kerameikon in Athens, where moulded lead figurines were placed inside inscribed miniature coffins that were also made of lead. But the Anna Perenna curses may also find a literary parallel of similar chronology. In a recipe from PGM XIII meant to keep a man from being cuckolded, the text recommends first moulding a crocodile out of earth, ink and myrrh and then placing it inside of a small lead cinerary urn (εἰς ορίον μολιβοῦν)86 upon which the practitioner is then told to inscribe the proper cursing formula.

3.3. Non-specific media This category contains all objects that were taken up and used in new or complementary ways to inscribe a curse. This includes non-manmade objects, such as shells or magnetite. Within this larger group, there are two large categories: perishable and non-perishable objects (see Table 3.2). 3.3.1. Perishable non-specific media Within this group the PGM mention a range of suitable objects for aggressive-magical purposes that are not found outside of the literary record. Such objects include hellebore leaves (φύλων χαλπάσ(ου), cf. PGM XIX, b1) and a donkey hide (δέρμα ὄνου, cf. PGM XXXVI, 362), both of which were recommended for agōgai (erotic curses of attraction). Furthermore, we find recommendations for using the wings of a living bat (νυκτερίδαν ζῶσαν ἐπὶ τῆς δεξιᾶς πτέρυγος (…) ἐπὶ τῆς ἀριστερᾶς),87 which were proposed for a curse to induce insomnia ‘until she gives her consent’.

In addition to this parallel, a formal comparison of some of the Anna Perenna containers and certain types of lead cinerary urns dating to the Imperial Period provides intriguing results: the containers, with their cylindrical bodies and flat or conical lids, are striking similar.

3.3.2. Non-perishable non-specific media

Hilgers 1969: 265–67, ‘Pyxis’. For a wider discussion, see Künzl 1982. For two bronze containers that come from Albalate de las Nogueras (Cuenca), see Fuentes Domínguez 1987: 259, fig. 7, no. 3 and 4 as well as fig. 8 no. 1 and 2. 84  One of the containers (cf. 24) resembles the inkwells published by Bilkei 1980: 89–90 tables III–IV, no. 22, 149 and especially 121. Likewise, it recalls the vessels recently found in a woman’s tomb in Aquincum (see Lassányi 2008: 68, fig. 6). 85  Piranomonte 2010a: 207 with useful bibliography in n. 51. 86  PGM XIII, 322. 83 

This is an extremely heterogeneous group in which we find objects ranging from shells, magnetite and pieces of metal to lamps, tegulae and ostraka. 87  PGM XII, 376–77. For another recipe to banish sleep that tells the practitioner to write a curse on bat wings, see PGM VII, 652.

12

Media To start with seashells (ὄστρακον ἀπὸ θαλάσσης),88 the PGM contain four different curses (three being erotic) that are supposed to be written on this material with very particular inks that were certainly aimed to reinforce the power of the spell.89 Another source for the use of seashells (with a clearly agonistic purpose) is found at DT 234, a curse tablet written against the charioteers and the horses of the blues and the greens. In ll. 6–7, the text states that the names of the enemy factions’ horses have been written on seashells and then deposited alongside the defixio in a tomb. Unfortunately, the archaeological report of this nineteenth-century excavation does not record its findspot (for a discussion, see Faraone 2019: 170). Drawing on the magical principle of persuasive analogy (similia similibus) and the well-known properties of magnetite (cf. Pliny, HN XXXVI, 127),90 the PGM prescribe the use of this mineral (λίθον μάγνητα τὸν πνέοντα, cf. PGM IV, 1723–24) in erotic curses in order to attract the beloved. In its natural form, magnetite can be found in octahedral crystals, a shape which provides good surfaces for etching images and inscribing texts, which would then be accompanied by the recitation of a longer formula. As was the case with seashells, we only have literary evidence for the use of magnetite in the manufacture of defixiones.

Figure 3.8. Non-specific media: tegula from Thysdrus/El Jem (From Foucher 2000: 58, fig. 1).

small pitcher and a tegula, respectively, before the pieces were fired (see Figure 3.8). In Italy, there are also examples of curses inscribed on fired ceramic. The first is a curse aimed to silence its victim found on an ostrakon from Neapolis/Terralba (cf. 113). The other two extent Italian examples, from Rome (cf. 3 and 14), were painted on a lamp and the lip of a cinerary urn, respectively (see the catalogue for images).

Within the category of non-perishable media, there is a sizable group of ceramic objects. The use of this material is documented in both the literary and archaeological records. Accordingly, to those seeking to separate two lovers, the PGM recommend inscribing a curse on a pot for smoked fish (εἰς ταρίχου ὄστρακον, cf. PGM XII, 366) with a bronze stylus. This procedure would give rise to ‘odiousness, enmity, just as Typhon and Osiris had’.91 For those seeking to attract a lover, on the other hand, the PGM suggest using a ceramic cup (ποτήριον, cf. PGM VII, 643)92 upon which the defigens ought to recite the spell seven times.

On a few rare occasions, ceramic objects have been found together with lead tablets. This is the case with the curse from Faviana/Mautern (Noricum, cf. 526), where a defixio inscribed on a lead sheet was used as a ‘lid’ for a small ceramic jar, inside of which archaeologists found organic remains. There is a parallel case from the fountain of Anna Perenna (Rome), where excavators discovered a ceramic jar filled with the remains of bone and parchment. The vessel was covered with a lead sheet that apparently bears no inscription.95

In a different love spell of attraction, the defigens is told to inscribe the curse on unfired clay (εἰς στρακον ὠμόν, cf. PGM XXXVI, 187).93 In this case, the instructions found in the PGM appear to find direct parallels in the archaeological record: there are two erotic curses, from Maar (Gallia Belgica, cf. 173) and from Thysdrus/El Jem (Byzacena, cf. Foucher 2000),94 which were inscribed on a

In addition to the ceramic media, it is worth noting as well an unparalleled case from Fundi/Fondi (cf. 57). The curse was painted as a titulus pictus on a reused opus sectile plaque made of marble. The choice of this media, which is certainly unparalleled, seems to be related both with the context in which the curse was finally deposited (i.e., an abandoned pool located in the so-called grotto of Tiberius) and with the content of the text. By the time the curse was composed, the cave still preserved the remains of its once luxurious decoration and was probably the place from which the practitioner took the object on which the curse was written, alluding to the destruction of the

88  This same lexeme (ὄστρακον ἀπὸ θαλάσσης) appears at PGM IV, 2218, though the type of spell is not explicitly stated; VII, 300ª (ὄστρακον θαλάσσιον); VII, 374 and VII, 467. 89  For these inks, see PGM VII, 300ª ff.; VII, 467 ff.; and IV, 2218 ff. 90  According to Pliny, trahitur namque magnete lapide, domitrixque illa rerum omnium materia ad inane nescio quid currit atque, ut propius venit, adsilit, tenetur amplexuque haeret. 91  Translated by R.F. Hock, apud Betz 1992. 92  PGM VII, 643. 93  PGM XXXVI, 187. 94  In Sánchez Natalías 2011b: 86, I wrongly included a tegula from Wilhering (Raetia), whose text had traditionally been interpreted as a defixio (see Kropp 2008: no. 6.2/1 [with previous references], Urbanová 2018: no. 102). However, the re-reading of the text found in Thüry 2001 has made it clear that this is not a magical document.

95  On this, see Polakova and Rapinesi 2002: 39 and 49. Jordan (1996: 115, n. 4), for his part, highlights the existence of cinerary urns with inscribed lead lids, which may provide an important Greek parallel.

13

Sylloge of  Defixiones from the Roman West

Figure 3.9. Non-specific media: curse on a pewter plate from Aquae Sulis/Bath (cf. 235; The Roman Baths).

space in which the defixio was deposited. In addition, the content of the text was also related with this idea, since it aimed to destroy the house and spirit of two individuals through a series of comparisons based on famous passages of the Bible. Therefore, the choice of this medium was probably aimed to reinforce, through persuasive analogy, the purposes pursued in the curse.96 Finally, we must discuss the use of metal objects as nonspecific media, for which all of our evidence comes from the archaeological record. There are two important curses from the sanctuary of Sulis Minerva (Aquae Sulis/Bath) which both list the names of cursed enemies. The first of these curses, 223, was inscribed on a tin disc that was coated with bronze. The disc has a small protrusion that could have been used to attach some sort of cord, suggesting that it might have been a pendant. The second curse, 235, was inscribed on a pewter plate, which was folded in two after the curse was inscribed (see Figure 3.9). In both cases, it appears that the media of the curses were everyday objects that were appropriated for writing a curse.97 We are left to wonder whether these objects were chosen because they were typical offerings to the goddess (this would especially be the case with the plate) or because they were somehow connected to one or several of the curse’s victims. Nevertheless, and given that both texts are just simple lists of names, neither hypothesis can be confirmed.

96  See Alfayé 2019: 269 and in general, for magical practices among ruins. 97  Discussing the tin disc, Mees (2009: 35) notes ‘Its place of deposition and the metal was made of suggest that the Bath tablet was specifically created for the purpose of cursing (…) perhaps the pendant form of the first Celtic Bath find was supposed to suggest the magical quality of the item, if not, say, be a symbolic representation of an object which had been stolen from the curser.’

14

4 Inscribing the Defixiones 4.1. Introduction: the written and spoken word

endeavour. Consequently, we would be remiss to give the written part of magical acts more importance.100

In determined contexts and due to their performative value, speaking and writing were two actions that were thought to have a certain agency in Antiquity. This means that the contents of a verbal or written message could come to fruition in the physical world if a series of specific rules and procedures were adhered to.98 Once the material on which a curse was to be inscribed was chosen, its content would be transmitted through two different modes of communication: the written and the spoken word. Accordingly, magical acts were performed on two separate but closely connected levels—even if today only one has been preserved. The important point that is easy to overlook due to the state of our evidence is that defixiones were both written and recited aloud.

Despite the inherent impossibility of preserving an oral utterance in the archaeological record, it is extremely likely, as the PGM clearly demonstrate and the XII Tables strongly suggest, that an oral recitation accompanied the deposition of the vast majority of defixiones. This is especially likely in the case of the non-inscribed tablets,101 such as the tablets found by a metal detectorist near Lincolnshire, upon of which the likeness of the emperor Valens (from a coin) was impressed (see Figure 4.1). In a similar vein, it is logical to hypothesize that the pseudoinscriptions found in the British sanctuaries in Aquae Sulis/Bath and Uley, which were inscribed by illiterate defigentes seeking to emulate the act of writing, were accompanied by oral cursing formulae that fleshed out the meaning and details of the inscribed ‘texts’.102

The oral aspect of the actio magica, few traces of which have been preserved, was undoubtedly a fundamental part of any cursing ritual. As P. Poccetti has put it, oral recitation ‘rappresentava il requisito essenziale per lo svolgimento e l’efficacia del rito magico’ (2002: 16) and played a role both in the manufacture of a tablet as well as its deposition. Several recipes found in the PGM make this clear: after the inscribing of the tablet, practitioners are generally advised to recite various formulae that were designed to invoke (and urge on) the powers that were tasked with bringing the curse to fruition and also provided the defigens with the opportunity to reiterate his or her demands.99 Here it is worth recalling the prohibitions enshrined in the XII Tables (dated to the mid-fifth century BCE), where cursing appears to have been conceived of as primarily an oral

What we know about the written portion of the ritual is, of course, more detailed, since in the majority of the cases this has been preserved. That said, and as is made clear in the catalogue, just because a text has been preserved does not mean it is legible, comprehensible or able to be deciphered. As various authors have pointed out, writing is a practice ‘a cui la condizione esoterica in società non profondamente alfabetizzate conferisce poteri magici’.103 These powers arise from the fact

98  See Rodríguez Mayorgas 2010: 37, who explains in a discussion of orality: ‘se le concede una fuerza creadora a la articulación sonora de las palabras, en especial de los nombres, como si tuvieran el poder de dar vida o de quitarla… Los ejemplos más elocuentes, sin duda, aparecen en la Biblia donde… Dios crea el mundo asignándole nombres’; for writing, see pp. 227–28. Likewise, Cardona 2009: 120 claims: ‘Tutta la storia della scrittura ci mostra… come si sia sempre ritenuto possibile agire sul reale a partire dalla manipolazione dei simboli, e come anzi l’uomo sia giunto a nutrire un terrore sacro di quei simboli e del loro potere, quasi che, ormai tracciati, essi potessero da sé soli, e senza intervento di altri, scatenare la loro azione’. 99  A good example can be found in PGM III, 30ff. (which recommends the recitation of formulae during the deposit of the tablets, which are to be placed inside of a ritually killed cat); PGM IV, 332ff. (which recommends reciting the formula written on the defixio and pronouncing others while depositing the tablet) and 1747ff. (which urges the reader to inscribe the formula and then recite it); 2235ff. (where the text is supposed to be recited after being written); PGM V, 319ff. (which recommends pronouncing a series of formulae, while piercing the tablet and a different series of formulae during the deposition); PGM VII, 453ff. (where the reader is told to recite a formula while depositing the tablet in order to increase the text’s longevity); 474ff. (which tells the reader to recite a formula while depositing the curse); PGM XII, 366ff. (which says the recite the formula after inscribing the tablet); and finally, PGM LXXVIII, 4ff. (which recommends inscribing a figure while saying a specific name and then later reciting a formula to invoke the deity).

Figure 4.1. Defixio from Lincolnshire (From: The Portable Antiquities Scheme). See Poccetti’s suggestive comparison of the Greek and Roman contexts (2002: 16–17, 2005 and 2015: 378–82); for the Hellenic world, see Faraone 1991a: 4–5. On this, see also section I.9.2. 101  E.g., those from Saguntum (see the introduction to Hispania, in II.2) or Rauranum/Rom (see the commentary to 159). 102  See 317–21, to which we could add 382 and 440. 103  Poccetti 2002: 15. Also see n. 98 above and Piccaluga 2010: 13–14, Graf 2015 and Chiarini 2019. 100 

15

Sylloge of  Defixiones from the Roman West that through the act of writing the person named in the curse (i.e., the victim) along with his or her belongings, actions, life and destiny were thought to be irreversibly connected to the curse tablet and hence constrained by its spell. Writing was considered a dangerous practice for a spell’s victim, as is made explicit in several formulae that establish a persuasive analogy between the victim and the act of writing.104 Furthermore, guaranteeing the longevity of a curse was often one of the practitioners’ main objectives:105 defigentes could produce spells that would dog a victim quoad vixerit106 or they could ask for its power to be renewed (rediviva)107 in an attempt to perpetually prolong its power.

That said, Table 4.1 clearly shows that the inclusion of such elements within the corpus of curses from the Roman West are actually quite rare. Provinces such as Africa or the Italian peninsula, which scholars have argued were more susceptible to the influence of Graeco-Egyptian magic,116 stand out for an elevated use of such inscribed elements. Here, it is worth stressing that the use of such Graeco-Egyptian elements is also more common within a certain time frame: during the High Empire, these features are present in North African defixiones, though it was not until Late Antiquity that these new ‘updates’ to the technology of cursing reached the Italian Peninsula and other Mediterranean provinces. Furthermore, we should point out that the presence of these items has generally been connected to the rise of magical professionals, who were especially active in large urban centres, such as Carthage, Hadrumetum and Rome. There is less evidence to suggest that these features were employed by individual defigentes writing curses on their own behalf.117 Yet even the higher rate of such elements in these areas does not counter the larger statistical trend: such tablets remain in the minority (or simply are not attested) in some northern provinces. Despite the fact that there are examples of curses that have unusual layouts already in the fourth century BCE, it is not until the High Empire and Late Antiquity that the combination of various languages (transliterations, mixed language texts, etc.) or the use of iconography and voces magicae are most widely attested. The fact that these aspects are concentrated in certain areas and times is yet a further reminder that we would be remiss to think of cursing as a monolithic category that was stable across time and space: rather, it was an evolving technology that could be adapted to changing social circumstances.

The PGM, for their part, allude to the act of writing a curse with the lexemes γρὰφω108 and ἐγχαράσσω,109 sometimes without any further specification. When it is specified that a curse should be inscribed on a lead sheet (except for rare exceptions), the PGM urge their readers to write their curses using a bronze stylus or nail.110 When the curse is written on other materials, such as papyrus, seashells or the wings of a bat, however, the PGM recommend the use of ink, which is spread with a pen111 and could be made of a wide range of ingredients (e.g., cinnabar,112 myrrh113 or even different types of blood114). 4.2. The act of inscription As has been mentioned in the previous section, the actio magica was carried out on two levels that were tightly linked: the oral and written. Within the second category, the importance of the inscribed word is undeniable. In addition, other Graeco-Egyptian elements that were incised on the tablets played an extremely important role and were thought to have a close symbiotic relationship with the inscribed word. The most common of these elements include the deployment of iconography, charaktêres and certain voces magicae (such as palindromes, alphabetic sequences and onómata barbariká). Sometimes such elements are found in conjunction on a single tablet.115

4.2.1. The text: layout, language and writing Text is the only element found in the vast majority of the tablets, which generally lack more ‘elaborate’ features, such as iconography or charaktêres. For the most part, text is written and runs in the customary fashion (i.e., from left to right), though depending on a pracititioner’s ability and knowledge, we occasionally find unusual layouts and/or noteworthy combinations of different languages and ways to organize writing.

104  See Faraone and Kropp 2010 and Chiarini 2019: 140–46. See also Poccetti 2002: 38–39. 105  Piccaluga 2010: 16. Cardona (2009: 141) claims, ‘se la forza magica evocatrice della parola detta si spegne quando l’ultimo suono è stato pronunciato, la potenza della formula scritta rimane intatta nel tempo, e non la si può spegnere se non distruggendo il supporto’. For a similar argument, see Ogden 1999: 26. 106  Cf. 250 B, l. 4. Analogous phrases are found in 486 (A, l. 8: quandius vita vixerit) as well as the British curses 350 (ll. 3–4: usque die`m´ qụọ moriatur) and 459 (ll. 6–7: (ad) diem mortịs). 107  Cf. 205, l. 1. 108  See PGM IV, 304, 2144; V, 360; VII, 459 and 462, etc. 109  See PGM VII, 433 and 459. 110  For the stylus, see PGM VII, 398, 417 and 926; XII, 366 (defixio written on a pot for smoked fish); XXXVI, 2, 186 (curse written on unfired clay) and 232. For the nail, see PGM LXXVIII, 4. 111  PGM V, 308. 112  PGM III, 18–19. 113  PGM XXXVI, 103 and perhaps also XIXb, 6. 114  The examples as for the blood of a black donkey (PGM VII, 301), a donkey (PGM XXXVI, 73), Typhon (PGM LXI, 61) or ‘in the blood from the womb of a silurus after mixing in the juice of the plant Sarapis’ (translated by E.N. O’Neil, PGM XXXVI, 362–63). 115  Examples include the defixiones 117–18 (dated to the fourth or fifth century CE), whose texts are written in a mix of Latin and Greek. The

4.2.1.1. The layout118 Eighty-eight curses (or 14 per cent of the corpus) contain an unusual layout (or ordinatio),119 the majority of which tablets also contain iconography, a charaktêr, a magical symbol and a sequence of nearly identical voces magicae. 116  Cf. Gordon 2002. 117  A defixio from Eccles (cf. 446) may be one of the few exceptions, on the reverse of which the Chnoubis symbol is found. 118  For an extended version of this section, which includes the Latin and Greek defixiones from the Roman West, see Sánchez Natalías 2020d: 103–08. 119  See Poccetti 2002: 47–49. For the layout of the text as a list, see Gordon 1999 as well as Centrone 2010 (though more focused on the Greek world). Also see Urbanová 2018, sections 7.6.1, 8.4, 9.4, 10.4, 11.3 and 12.4 (where the author briefly comments on the layout of several Latin texts as features such as voces magicae, iconography and signa magica).

16

Inscribing the Defixiones Table 4.1. Defixiones grouped by provenance (It. = Italia, Afr. = Africa, Hisp. = Hispania, Gall. = Gallia, Brit. = Britannia, Ger. = Germania, Raet. = Raetia, Nor. = Noricum, Pann. = Pannonia) and their characteristics (O = ordinatio, L/W = language/writing, I = iconography and C = charaktêres. Varia refers to tablets that contain fewer common elements, such as palindromes, magical symbols, voces magicae, etc. The total number of curse tablets in each province is tallied in the right column (that said, even if a defixio has several characteristics, it is counted as a single piece in the province’s total) Prov. It.

O

L/W

I

C

20

17

16

13

22, 117 and 118: voces magicae. 5: Greek vowels. 6, 20, 22, 24–26, 29–30: magical symbols. 26–27, 30: palindrome.

40/119

DT 218, 227, 233, 243, 244, 255: voces magicae. AE 1907, 165 and 1933, 234-235: voces magicae. DT 243: magical symbols, “wings”, palindromes. DT 247, 258 (?) and 260: magical symbols. DT 251: voces magicae transliterated with the Latin alphabet. DT 251 and 253: some Greek letters. DT 253: palindromes, voces magicae. DT 264 and 266: voces magicae and magical symbols. DT 265 and 268: voces magicae. DT 272-274: voces magicae and identical with respect to C. DT 286, 288-289, 291-293 and 295: voces magicae. DT 286 and 292: some Greek letters. AE 1907, 68-69: magical symbols. DT 270: voces magicae.

62/88

Varia

Total

Afr.

20

28

23

13

Hisp.

10

2

–

1

–

11/30

4

166: voces magicae (?) 156: alphabetic sequence. 183–184 and 189: text distributed in three rectangular boxes. 183–184 and 190: magical symbols. 196–197: iconography (?). 160: voces magicae.

11/55

–

207: alphabetic sequence (?). 259: sign (cross, side A). 446: magical symbol (SSS). 448: sign (side A).

21/256

Gall.

3

12

3

Brit.

17

Ger.

14

–

–

–

–

14/59

Raet.

3

–

–

–

–

3/6

Nor.

–

–

–

–

–

0/1

Pann.

1

1

–

–

530: magical symbols.

2/9

3

–

hail from Africa, Italia, Britannia and Germania. With the exception of the North African defixiones, the majority of the rest of these texts were by individual practitioners, who display quite different levels of literacy. For these individuals, writing in a non-standard manner appears to have been a means of making their curses even more effective. Instead of being an attempt to encrypt a message, the distortion of the standard layout, which appears to have always been intentional, has been explained along the lines of persuasive analogy, according to which just as a text is deformed or out of whack, so too shall be that which the text describes.120 A good example of this can be seen in 19, a curse from the sanctuary of Anna Perenna in Rome, which is directed against a certain Antonius. On side A, a small anthropomorphic figurine has been etched. 120 

This figure is clearly identified as the victim, whose name is written according to normal conventions. On side B, however, the name appears in a distorted fashion, with the syllables placed in the wrong order. The scrambling of these syllables, or membra if you will,121 should be analysed as a symbolic dismemberment of the victim. Among the ‘atypical’ layouts, the most common perversion consists of writing the text in the opposite direction. Of these, the oldest are 12 Oscan tablets, which mostly date between the fourth and third centuries BCE122 and which hail from southern Italy: the texts are written from left to 121  For membrum as a synonym for syllaba, see TLL s.v. membrum (8.0.634.30). 122  The tablets 75–77 are dated to the fourth century BCE, while 78–81 have been dated between the fourth and third centuries BCE. Curses 63– 66 have been dated to the Late Republic (first century BCE).

Cf. Ogden 1999: 30.

17

Sylloge of  Defixiones from the Roman West

Figure 4.4. Different types of right-to-left layouts: Celti/ Peñaflor (cf. 131; From Stylow 2012: 151).

Figure 4.2. Different types of right-to-left layouts: Fectio/ Vechten (cf. 461; From Bramach 1867: no. 57).

Figure 4.3. Different types of right-to-left layouts: Colonia/ Cologne (cf. 464; From Blänsdorf et al. 2010: 272).

right, instead of right to left, which is standard in Oscan epigraphy. A similar case is documented in Volaterrae/ Volterra (cf. 91), where, unlike the Etruscan usage, the text runs from left to right. The remaining cases, all of which are in the Latin language, are written from right to left. These tablets mostly hail from Germania, Britannia and Hispania.123 From a chronological point of view, there are three early examples from Hispania (first century BCE),124 but it was not until the High Empire that this sort of layout really took off (with 23 examples); by Late Antiquity, the trend of writing a text backwards seems to have waned with only 5 known examples. Geographically, this type of layout spread along the same lines as cursing practices did generally: the first examples are found in southern Italy (probably under Greek influence)125 and then spread to the northern provinces. In some examples of texts running from right to left, the orientation of the individual letters was also ‘mirrored’ (i.e., written from right to left), such as the curse from Vechten (cf. 461, Figure 4.2);126 that said, in most of the tablets the orientation of the letters remained standard, as can be seen in the Colonia/Cologne tablet (cf. 464, Figure 4.3).127 Another rare but important variation of the normal layout can be found in texts that are double reversed and run from right to left and bottom to top. In other words, to read the text, one must start at the end of

Figure 4.5. DT 233: ‘Fence’-style layout (From: CIL VIII, 12504).

the final line, as can be seen in the Celti/Peñaflor curse (cf. 131, Figure 4.4).128 Another layout that merits mention is the ‘frame’ or ‘fence’ type in which a portion of the text follows the edges of the tablet (see Figure 4.5). Such a layout is found in 20 known Latin curses dating from the High Empire and mostly discovered in the North African cities Carthage and Hadrumetum. As Gordon has noted, it is quite possible that this layout was ‘derive[d] from the praxis of magical-amulet design, where divine images are regularly enclosed by a text’ (2005: 78–79). As I was able to confirm during an autopsy of several North African curses,129 the inscription of such texts begins with the ‘fence’ and then continues inward to fill the defined space. Without a doubt, the text chosen to form the ‘fence’ was

These provinces have yielded 12, 8 and 7 items respectively. To this group we must add the three found in Gaul, one from Italy and another from Raetia. 124  Specifically, 125–26 and 133, to which 139 (dated to the end of the first century BCE or beginning of the first century CE) could also be added. 125  Such as the five tablets from Selinunte (mostly dated to the mid-fifth century BCE), in which the lines run from right to left, though the letters maintain their normal orientation. See Bettarini 2005: nos 2, 4, 12, 24 and 25. 126  The tablets 59, 133, 164, 249, 442 and 462 could be added. 127  Other examples are 85, 125–26, 139, 167, 266, 341, 345, 350, 450, 479–80 (these two only in part), 481, 483, 485, 489, 501 and 503–04. 123 

128  129 

18

To which 134–36, 149, 303, 465 and 524–25 ought to be added. Specifically, DT 276–80 and 282–84.

Inscribing the Defixiones (cf.  145), for instance, the circular shape of the curse’s media functioned as a guide for its authors, whose texts took on the form of a spiral. In these cases, the layout of the text seems to have been determined by the object on which it was inscribed and hence can be understood as practical rather than a decision dictated by ritual. There are also two tablets from Aquae Sulis/Bath (cf. 308) and Brigantium/ Bregenz (cf. 521), in which there is a combination of lines running from left to right and right to left (perhaps these are attempts at boustrophedon?). This layout is taken to the next level in a curse from Poetovio/Ptuj (cf. 527), a tablet in which letters are additionally written mirrored and upside down.136 But it is in two British curses from London (cf. 340 side B) and East Farleigh (cf. 458), written in capitals with very irregular layouts, that things are taken to an extreme: text runs from right to left, left to right, in ascending and descending fashion; furthermore, letters are sometimes oriented forwards, sometimes ‘mirrored’ and sometimes upside down (see the commentary in the sylloge for a fuller description).

not left to chance; on the contrary, all evidence shows that the professional practitioners that were producing these curses carefully chose the text to occupy this important and symbolically significant space. Accordingly, in three cases the ‘fence’ consists of the names of the daemones invoked,130 while in another six curses it is composed of voces magicae.131 In the majority of such texts, however, the ‘fence’ is reserved for the cursing formula proper.132 While several of this curses should be categorized as juridical (cf. DT 218), erotic (cf. DT 227) or commercial (cf. AE 1933, 234–35), the bulk of them are agonistic in nature and deal with chariot races.133 In such cases, it is possible that the ‘fence’ can be interpreted as a schematic representation of the circus or even of the race itself. Whatever the case may be and irrespective of the content of any particular curse, it seems clear that the intentional layout should be understood in terms of a persuasive analogy, since the ‘fence’ is an excellent means of symbolically ‘confining’ the spell’s victims, who are accordingly trapped and cannot escape the text’s powerful sway.134

When a tablet also includes iconography, the inscription is fitted around the images almost as a frame (cf. with 22 and 24, from Rome). On certain occasions labels are used to identify the represented figures (cf. 6 and 117, from the ‘Sethian’ and the Bologna collection, respectively).

The defixio against Prima Aemilia (from the sanctuary of Isis and Mater Magna in Mogontiacum/Mainz, cf. 484) is a particular ‘fence’-style curse that deserves special mention since the layout seems to be a blunder on the writer’s part. Though the author clearly had a good deal of experience with writing practices, he or she did not properly judge the amount of space needed for writing the curse: the author wrote the first four lines following the edges of the tablet (i.e., turning it 90 degrees between each line); then, (s)he continued with the curse beginning just above the centre of the tablet. However, the author miscalculated the space needed to finish and hence had to finish writing on the upper part of the tablet; as a result, there is a space of about 8 mm between ll. 6 and 13. Despite an effort to use the fence-style, it is clear that the result was not quite as planned: the author simply misjudged the needed space for writing the entire text. Something similar can be observed in a defixio from Cruciniacum/Bad Kreuznach, (cf. 473), in which the defigens had so many enemies that (s)he had to use the margins of the tablet to squeeze in the final cursing formula. This technique of using the margins to finish a message has been commonly observed in ancient letter writing as well.135

4.2.1.2. Language and writing After the distortion of the layout, the most attested innovation in the curse tablets from the Roman West is the combination of different languages and/or scripts. This is found in almost 10 per cent of the corpus of defixiones, which mostly date to the High Empire. Scholars have generally explained the practice of combining different languages and/or writing systems as an attempt on the authors’ part to increase the text’s magical power. Nevertheless, this is only clear when certain significant words or phrases have been affected.137 Up to what point these sorts of combinations were thought to ameliorate or facilitate the defigens’ communication with the powers invoked is a worthwhile question. The topic could be understood in light of the High-Empire trend of using obsolete or at least extremely erudite language in religious writings.138 But this was not always the case. As we will see in the following paragraphs, some of these cases can also be explained in terms of practicality, like in instances of linguistic borrowings or of an author’s insufficient knowledge of the Latin alphabet. Whatever the case may be, it appears that most of the authors who wrote the

Finally, it is necessary to mention other layouts, which, though less common, are quite eye-catching instances. In two curses from Rome (cf. 26) and Barchín del Hoyo 130  Documented in DT 284 (from Hadrumetum), and AE 1933, 234–35 (from Carthage). 131  Like in the Carthaginian tablets DT 218 (juridical in nature), DT 227 (possibly erotic), DT 233, DT 244, DT 253 (agonistic) and DT 255 (written for unknown reasons). 132  Among these, we find DT 275–84 and AE 1907, 68–69 (all agonistic curses from Hadrumetum; for DT 275, see Cagnat’s commentary apud Audollent 1904: 383; Piccaluga 1983: 122–23; Gordon 2002: 92–93). 133  Cf. DT 233, 244, 253, 275–84 and AE 1907, 68 and 69 (for the charaktêres found in DT 276–85 and AE 1907, 68 and 69, see Németh 2011). 134  Cf. Sánchez Natalías 2020d: 105–06. 135  Sarri 2018: 112; an analogous case is found in 494.

We can also add 461 (cf. Fig. 4.2) or 479. Poccetti 2002: 28; Marco Simón 2012: 128. 138  Good examples of this are the Carmen Arvale (CIL VI, 2014, written in archaic Latin during the reign of Augustus), the so-called ‘Pillar of the Boatmen’ (RIG II.1, L-14), which the nautae Parisiaci dedicated to Jupiter during the reign of Tiberius. This piece consists of stone blocks which have been engraved with Roman and indigenous deities, which are identified with Latin and Gaulish theonyms. Another example is the small group of Lusitanian inscriptions from Lamas de Moledo and Arroyo de la Luz (cf. Alfayé and Marco Simón 2008). For more examples of linguistic recuperation in religious contexts, see Mancini 1988: 208–09. 136  137 

19

Sylloge of  Defixiones from the Roman West tablets discussed in this section of the prolegomena were well-trained professionals who were capable of mixing various languages and writing systems within the confines of a single text. In several cases, this sort of sophistication was combined with other noteworthy features.139

inclusion of such Greek language elements undoubtedly constitutes a conscious decision on the part of the author, who sought to imbue his or her text with a powerful boost that would ensure its efficacy. On certain occasions, the actual invocation of the divinities or even an entire cursing formula were written in Greek, while the rest of the inscription was put in Latin.147 This is the case with a curse from Lilybaeum/Marsala (cf. 109), the oldest of this group (dated between the fourth and third centuries BCE), where the author employs a long Greek invocation to summon Persephone, the Titans and the atélestoi; the practitioner ask that they take action against a group of victims, whose names are listed in Latin. Another interesting case is found in a pair of Carthaginian tablets discovered in the amphitheatre and dated to the High Empire (cf. DT 248–49). In the first, the text is written in Latin except for the final lines, which contain the classic formula ἤδη ἤδη ταχὺ ταχὺ (DT 248 B, ll. 8–9), while the second curse is completely in Greek except for the brief mention of the victims (DT 249 B, ll. 1–3). Given that these two curses target the same individuals, it appears that they were written by the same magical professional and meant to work together in conjunction.

Following the typology proposed by J.N. Adams in his study of bilingualism in Antiquity, we can place the relevant defixiones into three categories: bilingual, mixed and transliterated (2003: 29–70). While texts can usually be neatly placed in one of these three categories, there are certain cases which could be classified in either the second or third categories.140 Bilingual curses are those that use two different languages to duplicate the same message. There is only one known instance of such a curse, which comes from Barchín del Hoyo (cf. 145; dated to the end of the first century BCE or beginning of the first century CE) and written in Greek and Latin by an individual who appears to have been a native Greek speaker.141 Adams’ mixed type of text is much more common. In this instance, a curse employs various languages to elaborate different content. The causes that give rise to these hybrid texts can be explained by that fact that, as Poccetti explains, ‘alla scrittura sovente il potere magico deriva dalla condizione di allografismo e/o di allografia. Si attribuisce, cioè, forza magica alla scrittura o alla parola che appartengono a tradizioni grafiche e/o linguistiche “altre”’ (2002: 28). In the Roman West, 48 tablets belong to this group, the majority of which hail from Africa and Italia and were written in a mixture of Latin and Greek.142 Here, it is important to note that most of these texts were written by professional magical practitioners, who were showing off their Graeco-Egyptian magical know-how in order to impress their audience (human clients and gods invoked). Generally speaking, Greek is used for writing certain ‘technical’ magical phrases, such as the names of daemones or deities,143 certain voces magicae,144 palindromes145 and the seven vowels of the alphabet.146 The

Even though the majority of tablets within the mixed categories combine Latin and Greek, there are other noteworthy mixtures of languages. A good example comes from Cumae (cf. 64): the curse, which dates to the end of the first century BCE and was inscribed by a Latin-speaking practitioner, uses Latin to name the spell’s victims148 but switches to Oscan for the cursing formula proper.149 As Mancini has argued, the linguistic diversity of this curse, which is ‘intenzionalmente distante dall’uso quotidiano’ (1988: 208), was meant to guarantee the spell’s efficacy. A similar case can be found in the curse tablet from Rauranum/Rom (cf. 159), which mixes Latin and Gaulish.150 This tablet, which dates between the third and fourth centuries CE and whose interpretation has proven to be extremely controversial, contains various Latin lexemes (pura [A, ll. 7 and 9], poteat [B, l. 6] and decipia [A, l. 6]) alongside a series of Gaulish verbal forms (e.g., uoraiimo [B, ll. 1 and 10], atepriauimo

139  E.g., 5, 117–18 or the North African curses DT 251, 253, 286, 289 and 292–94. 140  Such curses are 5, 117, 130 and the curses from Hadrumetum DT 295 and 269. All belong to the mixed type (since they are written in Latin and Greek) and to the category of transliteration (since the first four have Greek texts transliterated into the Latin alphabet, while the final one contains a Latin fragment transliterated into the Greek alphabet). 141  According to the editors (Curbera et al. 1999: 283), the text’s palaeographical characteristics (such as the use of ‘Δ’ for ‘D’ and the final -n to mark the accusative in the Latin portion) support this hypothesis. 142  From Africa are the Carthaginian curses DT 218, 227, 233, 243, 244, 248–51, 253, 255, AE 1933, 234–35 and Jordan 1988: no. 3; the tablet from Naro/Hammam-Lif (Audollent 1910, no. 1) and those from Hadrumetum DT 264, 266, 293–95 and AE 1911, 6. Most of these African texts were written by professional practitioners. From Italy are 64, 109, 117–18, and the curses written by professionals discovered in Rome 5, 24–28, 33–36, 38 and 40. To these, we can also add 150–55, 166, 157, 159 and 530. 143  Such as in the Carthaginian curses DT 230, 243, 250, 253, AE 1933, 234–35; Jordan 1988: no. 3; the curses from Hadrumetum DT 264 (?), 293–94 and AE 1911, 6; the curses from Rome 5, 25–27 and 35–36. 144  Such as in the Carthaginian curses DT 218, 227, 233, 243–44, 253 and 255; that from Naro/Hammam-Lif (Audollent 1910, no. 1); the curse from Hadrumetum AE 1905, 171; the curses from the fountain of Anna Perenna in Rome 20, 24–28, 33–34, 38, 40. 145  Such as in 26 and DT 253. 146  Such as in 5 and Jordan 1988: no. 3.

147  For the Greek invocations, cf. 109, 117–18, 166 and the curse from Hadrumetum DT 295. Complete formulae in Greek are found in 109, 530, DT 248, 249, 253 and perhaps 266. 148  The victim’s names, which are in the nominative, use different terminations: the first three employ Oscan morphology (ending in –is, ll. 1–2), whereas the last two use the Latin termination (ll. 4–5). 149  As Mancini (1988: 222–23) points out, ‘dove è stato possibile il trascrittore ha operato nel senso di latinizzare la formula originariamente osca. In quest’ ambito si spiegano le forme sint, sit, recta. L’ovvia sovrapponibilità formale e funzionale ha causato una sorta di riverniciatura latina di queste parole, che in osco dovevano suonare rispettivamente sins, *sid, *rehtas’. 150  Lambert (apud RIG II.2) further adds the tablets from Le Mans (cf. 171) and Les Martres-de-Veyre (cf. 161). In the first, he interprets DM as an abbreviation of the Latin formula D(is) M(anibus), though we must remember that this proposal is extremely speculative due to the complexity of the text, in which only several Celtic personal names can be identified. In 161, Lambert has taken DIVOS (A, l. 2) as Celtic diuos, though it could certainly be Latin divos.

20

Inscribing the Defixiones [B, ll. 7–8] and bicartaont [A, l. 5]). Finally, we should perhaps add to this group of mixed curses the six defixiones discovered in Amélie-les-Bains (cf. 150–55, today their whereabouts are unknown) that were written in a language whose precise identification is not agreed upon151 but are clearly interspersed with a series of Latin words like rogamos and depecamụs (‘Vulgar’ forms of ̣ rogamus and deprecamur).152 Generally speaking, there are various factors (our partial understanding of the Gaulish language, the tablets’ poor state of conservation and their disappearance) that have impaired the analysis of these texts and kept us from determining to what extent and to what end Latin was mixed with Gaulish, not to mention what role each language played within these curses.

onómata156 or iconography157 to accomplish that purpose), but rather point to the identity of the magical practitioner from whom the tablet was commissioned. These professionals were probably native Greek speakers who had not achieved mastery of the Latin alphabet and had not found a way to render certain Greek characters in a different alphabet.158 The third and final of Adams’ categories deals with transliterated texts, that is inscriptions written in one language, but using the writing system of a different language. Among the defixiones from the Roman West, those that belong to this category come from the Mediterranean provinces and tend to combine different languages and alphabets. Good examples can be found in five erotic curses that were discovered in Hadrumetum and Carthage: while they are in the Latin language, the magical practitioners who wrote them opted for the Greek alphabet.159

Two other curses should be mentioned in which a text almost entirely in Latin includes several specific words in a different language.153 The first comes from Montfo (Gallia Narbonensis, cf. 157), which was found inside of a well and dates to the first century CE. In this text, the defigens references the masitlatida (B, l. 5), which appears to be a Gallic ritual of unknown characteristics (cf. Lejeune 1981: 52). The second example is a curse tablet from the sanctuary of Sulis Minerva in Aquae Sulis/Bath (cf. 206), dated to the second century CE, which reads in part: ‘qụ[i] mihi vilbiam iṇ[v]olavit…’ (l. 1). Russell has interpreted vilbiam as a Celtic word that referred to the stolen object mentioned in the curse, perhaps a tool similar to a gouge (2006: 364). In these two cases, the inclusion of an indigenous word is eminently practical and should be considered instances of linguistic borrowing and hence be differentiated from the previous examples.

Less common are tablets written in the Greek language using the Latin alphabet. In fact, there are only five known curses that fit this category and only partially so. The first example is a Carthaginian curse that contains the onómata barbariká as well as some transliterated formulae;160 the second comes from Hadrumetum and invokes to autem Domina Canpana,161 whereas the third example, from Rome (cf. 5), ends with [ed]e ede/ tacy tacy (ll. 21–22). Perhaps the second fragment from a curse from the Bologna collection, 117, in which an imaginary(?) language whose meaning is undecipherable but phonetically resembles Greek, should be counted in this group. Furthermore, a fifth curse that hails from Aratispi/Villanueva del Cauche (Baetica, see Figure 4.6 and cf. 130) should be added. After eight lines of charaktêres, it reads hutos apoleson M(arcum) Domitiu(m) Nigrum Firmanae filium. In these cases, the use of transliteration has generally been understood as a ritual attempt on the magical practitioners’ part to increase the magical power of the spell.

Now we must turn our attention to a final series of mixed tablets that employ what Adams has called ‘alphabetswitching’, which is securely attested in seven curses from North Africa. These curses, written in the Latin alphabet, include some Greek letters, such as Chi (χ)154 or Zeta (ζ),155 ‘for which there was no exact equivalence in Latin’ (cf. Marco Simón 2012: 125). As in the previous two examples, these cases do not provide evidence for using a different alphabet as a means of augmenting a tablet’s magical potency (these tablets actually employ

Celtic and Oscan tablets need to be mentioned separately, since they do not belong to the category of transliterated texts: in these instances, the curses were written in the alphabets that their users had taken up for writing their native languages. The use of these writing systems, then, had become traditional and therefore their use in defixiones cannot be attributed any sort of magical function.

151  Lambert (apud RIG II.2, *L-97) interprets it as a mix of Gaulish and Latin, while Coromines (1975 and 1976) identified it as Sorothapic (the language spoken by the so-called Urnfield culture). For a general discussion, see RIG II.2, p. 250. 152  Rogamos et depecamụs appear in 150, I, ll. 2–3, and perhaps in 154, ̣ ll. 2–4, while rogamus is found on its own in 151, A, l. 2 (perhaps this is a case of a mixed language similar to that from La Graufesenque?). 153  In addition to these curses, a tablet from Orosei (cf. 110) should be added. Caprara (1978: 152) has interpreted the word nurgo (ll. 2 and 5) as a Sardinian word derived ‘della radice nur –nella sua accezione di “voragine”, e quindi probabilmente di luogo infero’. 154  Cf. DT 251 (in which some onómata maintain the final c, as in Psarchyrinχ (Col. 1, l. 2) and in which the Greek letters π and ω also appear) and DT 292–94 (where final χ is maintained in the vox magica anoχ). 155  See DT 253, where the author wrote Vicentζus for Vicentius (ll. 10, 11, etc.). Poccetti has defined this as ‘ricorso a grafemi greci che segnalano processi di affricazione’ (2002: 46). Also see DT 286 and 289, where the vox magica basagra is written as baζagra.

Found in DT 251, 253, 286, 289 and 292–94. As is the case in DT 286, 289 and 292. 158  As Adams put it ‘If someone writing (e.g.) Latin in Latin script adopts Greek letters from time to time, he was not in total command of Latin script and is likely to have been a Greek with imperfect Latin literacy’ (2003: 72). 159  Specifically, DT 267 (ll. 15–25), 269–70 and 304, and the Carthaginian curse DT 231. Additionally, we must add 436, an exceptional curse from the sanctuary of Mercury in Uley (Britannia) that was written in the Latin language and Greek alphabet (for a preliminary note on the curse, see Tomlin 2002: 175). 160  Cf. DT 251, I, ll. 2–10, II and III, l. 2. 161  See DT 295, l. 20–21. 156  157 

21

Sylloge of  Defixiones from the Roman West since the publication of curse tablets (which in some cases were edited more than a century ago) sometimes lack visual representations. This omission surely needs to be remedied and further study of these defixiones’ iconography is sorely missing. In all of these tablets, iconography is accompanied by the text. That is, there are no defixiones with only images, while there are obviously many curses with text without an image. That said, in tablets that have both text and image, where there is always a close connection between the two, it would be inaccurate to claim that the image is somehow secondary. In fact, in these curses it is the text that is fitted around the image, not the other way around. Texts with iconography, then, contain a carefully planned layout which serves to reinforce the power of the image. The text, therefore, can be understood as a secondary element that appears to be dependent on the image. This, of course, is not to say that the text is somehow rendered unimportant. The result of combining these two elements is a clearer, more compelling defixio that is better equipped to carry out the practitioners’ desires, since his or her message can reach the powers invoked through both image and text.

Figure 4.6. Detail from the Aratispi/Villanueva del Cauche defixio (cf. 130 From: CIL-Archiv, II_5, 0729 beta).

With the exception of five curses from Mazan and Augusta Treverorum/Trier (Gallia), all of the curses with iconography hail from the places near the Mediterranean, such as Rome, Carthage, Hadrumetum, Cirta/Constantine and Thysdrus/El Jem and were composed by professional practitioners. The earliest examples date to the second and third centuries CE and were discovered in the abovementioned North African cities. It is only at the beginning of the fourth century CE that defixiones with iconography appear in the Roman archaeological record. In the fourth and fifth centuries CE such defixiones are well attested in the imperial capital, as can be seen in the Late-Antique collections from the sanctuary of Anna Perenna and the so-called ‘Sethian’ defixiones.

4.2.2. Other components linked to the text As Gordon (2002: 17) puts it, ‘it was precisely in malign magic, generally considered the most difficult to perform, that special efforts were felt necessary to increase the force of the text’. Among these special efforts, and as mentioned previously, individuals (or better, professional practitioners) could include a range of other components that coexisted in a symbiotic relationship with the curse’s text, such as iconography, charaktêres and certain voces magicae. As can be deduced from Table 4.1, these elements appear (save in rare exceptions) in curses written by professional practitioners coming from North Africa and Italy and mostly date to the High Empire or Late Antiquity, respectively. Iconography, charaktêres and voces magicae are the most widely attested of such additional components, while palindromes and alphabetic sequences are less common.

The drawings found on these tablets generally represent either the deity invoked or the curse’s victims. There are a few rare instances in which both the deity and victim are represented on the same tablet; while it is even rarer to find other sorts of images.164 Instances in which ‘the images concretize the divinity or power (or a metamorph) whose intervention is required by the spell or rite’165 are especially common in African curses. In said province, there is an important series of (mostly) agonistic defixiones (dated between the second and third centuries CE and composed by professional practitioners) on which the demon invoked is depicted.166 Such demons tend to be anthropomorphic167 —though sometimes they

4.2.2.1. Imagines magicae162 Within the collection of curses from the Roman West, the inclusion of iconography is a relatively rare phenomenon. So far, there are only 42 known examples of defixiones that are accompanied by imagines magicae.163 In almost half of the instances, the knowledge that we have of these imagines is often limited to an editor’s verbal description, 162  An extended version of this section, which includes the Latin and Greek defixiones from the Roman West, can be found in Sánchez Natalías 2020d: 108–21. On this same topic, see also Sánchez Natalías 2013c. 163  See Gordon 2002 and 2018 (for Late-Antique curse tablets with elements of the Graeco-Egyptian tradition, such as those from Anna Perenna or the Sethian collection); Viglione 2010: 119–24 (though she is mostly focused on Greek texts); Németh 2013 (which includes the unpublished sketches that Audollent made while compiling DT). For the connection between images, oral speech and writing in the PGM, see Crippa 2002 and 2010.

See the classification proposed by Cesano 19612: 1568. Gordon 2002: 98. 166  These are the Carthaginian tablets DT 229, 244, 247–48 and 260 and those from Cirta/Constantine (DT 300), Naro/Hammam-Lif (Audollent 1910: 136–41) and Hadrumetum (DT 286–92 and AE 1911, 6). 167  Cf. DT 229 and 244 (= Németh 2013: 152, 164–65, where it looks like a bust) as well as those from Cirta/Constantine (DT 300; see fig. 16), Naro/Hammam-Lif (Audollent 1910: 137–41) and Hadrumetum (DT 286–92= Németh 2013: 200–09 and AE 1911, 6). 164  165 

22

Inscribing the Defixiones From Italia (though their precise archaeological context is unknown), there are other noteworthy defixiones, such as 117–18, on which there are possibly representations of Hecate-Selene. In favour of such an identification would be the series of onómata composed of φωρ- and the snakes that emerge from her head (which recall a description of the goddess in PGM IV 2800–02: ‘[you] who shake your locks of fearful serpents on your brow’).174 The collection from Anna Perenna also contains some curses with noteworthy iconography. The deity invoked is depicted on various containers from the fountain (cf. 25–27 and 35–36) and takes the form of a figure with the crested head of a rooster that carries a lamp or a torch in its hand. Furthermore, this figure bears an inscription on its abdomen, which, according to G. Németh (2012b and 2015), is an abbreviation (with slight variations) for the Christian formula Ἰησοῦς Χριστός Ναζωοραῖος, Ἰησοῦς Χριστός Ναζωοραῖος, καὶ Θεός, Θεός, Θεός. Like the abovementioned case of Baitmo Arbitto or the other North African demons, this formula would have served to invoke and identify the divine power tasked with bringing the spell to fruition. Dating to Late Antiquity and probably made by a single professional practitioner, this series of containers allows us to glimpse the rich environment of religious syncretism in which elements from various traditions came together. Accordingly, and although the rooster-headed figure is reminiscent of Abraxas,175 the formula inscribed on his abdomen signals a Christian context in which the palindrome Ablanathanabla, which is attested in some Gnostic gemstones that contain all these elements, was also used.176

Figure 4.7. Imagines magicae: daemon Baitmo Arbitto (From: DT 286).

have the heads of a snake,168 bird169 or horse170— unshod (as a way of marking their divinity) and carry attributes or magical symbols. Sometimes the deity’s name is inscribed on his chest. In some tablets, the demons are shown seated,171 though it is more common to see them standing or on a ship, as can be seen in the famous demon Baitmo Arbitto, who is depicted in an important series from Hadrumetum172 (see Figure 4.7). We should also mention two curses from Carthage on which the defigentes depicted beasts and fantastical creatures, which may be representations of the powers tasked with carrying out the curses’ requests.173

Blänsdorf, the editor of the Anna Perenna curses, has identified another deity invoked in a container (cf. 28): Seth. The identification, however, is quite suspect, since a new reading of the container determined that the deity is not named in the text, but rather the victim: Seberinus.177 Thus, the representation would match the curse’s text, alluding to the victim of the spell. This leads us to the other predominant type of iconography found on defixiones: representations of the victim.178 Such representations broadly take the form of portraits or busts, depicted frontally. Even if this model could have been inspired by funerary portraits of the deceased (commonly found in cemeteries), a better way to think of them is as beheaded victims (cf. Martín Hernández 2021). Good examples of this ‘bust’ style can be seen on a Carthaginian defixio, which is probably agonistic

Cf. DT 247, where Audollent described this representation as ‘Homo stans (Typhon-Seth?) serpentis capite insignis, hastam dextra fulmen (?) sinistra’. See Németh 2013: 168. 169  Cf. DT 260, which Audollent described as ‘Daemon cuius caput avis (?) capiti simile quasi petaso tegitur duplici crista ornato et brachia extenduntur, nescio quid dextra, tridentem fortasse sinistra tenens’. See Németh 2013: 179. 170  Cf. DT 248, which Audollent described as ‘Daemon stans longis auribus et asinino (?) capite insignis, pateras (?) duas, sinistra gerens, fascem (?) dextra’. See Németh 2013: 168. 171  Cf. DT 292, which Audollent described as ‘Daemon sedens in structura caementicia obliqueis lineis adumbrata, flagellum tenens sinistra. Ante pedes daemonis humana protome’. Sichet (2000: 190, fig. 24), provides a drawing based on the photograph published in Cat. Alaoui, tab. XXVIII. See Németh 2013: 207–09. 172  Specifically, DT 286–91. See Gordon 2005: 69–76 and Németh 2013: 200–05 and 230. 173  Cf. DT 255 and 259 (= Németh 2013: 176 and 178–79, respectively). Audollent described the first as ‘imago beluae ingentis delineata est, quam quo nomine adpellem nescio; nam etsi hippopotamo similis est corpore et cornu in naso erecto, elephanta in memoriam revocat hirsuta proboscide (non lingua, ut ait W.), a fronte ad pedes promissa; praeterea alterum cornu et fortasse galli crista in capite eminent, ac tandem cauda draco fuisse videtur; saetae per totum corpus horrent’. The second is DT 259, in which the editor identified a cross and ‘Ensis aut crux ita 168 

figuratur cum canis aut bestiae cuiusdam imagine in summa hasta’. See the discussion in Németh 2012a: 145 and figs 8–11. 174  Translated by E.N. O’Neil apud Betz 19922. On this iconography, cf. Sánchez Natalías 2013a, with further references. 175  For the identification as Abraxas, see Blänsdorf 2012e: 619, 622, 624– 25 and 2015b: 22f.; Piranomonte 2012a: 167–71 and 2012b: 618 and 625 (with further references). 176  See Spier 2007: 81–85. 177  Cf. 28 and Sánchez Natalías 2020b. 178  For the representations of the victims in the tablets, see Marco Simón and Sánchez Natalías 2022.

23

Sylloge of  Defixiones from the Roman West

Figure 4.9. Imagines magicae: tied charioteer (From Jordan 1988, 130, fig. 2).

Figure 4.8. Imagines magicae: Saturnius (From Delatte et al. 1906, 323).

in nature179 (see Figure 4.8), or on a curse from Augusta Treverorum/Trier (cf. 180) that appears to be directed against a gladiator. Occasionally, there is a full-body representation of the victim, who is shown standing, such as we find on two curses from the sanctuary of Anna Perenna. The first, 19, targets Antonius, who is shown facing the viewer in standing position with his arms placed against his sides. He looks stunned, like someone that is exposed to and helpless against a magical spell. In a similar fashion, although with less expressiveness, is the abovementioned case of Seberinus, depicted in 28.

Figure 4.10. Imagines magicae: victim encircled by snakes (From DT 245, apud Németh 2013, 166).

Some practitioners, however, were much more explicit: in certain cases, we find representations of the terrible punishment that victims are meant to suffer. The first examples of this come from Carthage and date to around the third century CE. In an agonistic curse against an auriga, the victim is represented with his hands and feet crossed and tied with a rope, in such a way that he surely could not drive his chariot, let alone win a race. The depicted ropes served as a symbolic representation of the desired constraint of the curse’s victim180 (see Figure 4.9). In addition to ropes, we can also find other restrictive elements such as large snakes. In another defixio from Carthage of similar date, we see a victim trapped in a sort of wooden coffin from which his head sticks out, entrapped within the coils of a huge snake. Although the head of the reptile has been lost, Audollent’s drawing preserves the

bifurcated tip of the snake’s tongue, which is just shy of touching the victim’s face (see Figure 4.10). These large winding serpents with their pointy ears, pronged tongues and scaled bodies are clearly found in classical literature and iconography (just think of Aesclepius’ snake, the agathos daimon or the myth of Laocoön). Nevertheless, during Late Antiquity this reptile enjoyed a new resurgence thanks to Egypt’s influence on magical-religious practices throughout the Mediterranean. As proposed elsewhere,181 it is possible that the image of Osiris wrapped in the coils of a snake could have served as a model for the magical practitioners who drew them on Late Antique curse tablets. Unlike in Egyptian religion where the serpent protected Osiris in the underworld, in these curses the snake becomes a symbol of danger and constriction that is associated with demonic forces.

See Delattre and Monceaux 1906: 222–23 (= AE 1907, 165). Cf. Jordan 1988: 129ff. It is worth stressing that the representation of victims with tied hands and feet is quite common in the collection of ‘Sethian’ defixiones, as can be appreciated in SV 20–22, etc. 179  180 

181 

24

Sánchez Natalías 2020d: 115.

Inscribing the Defixiones There are several examples from the fountain of Anna Perenna that match this new conception of the snake. The curse tablet against Sura (cf. 22), for example, contains a noteworthy representation of its victim with snakes along with other eye-catching features: the judge Sura, whose face is superimposed on a vulva (cf. Gradvohl 2009). This image is connected to the variation of a common matronymic formula, which in this case reads qui natu(s) est de vulva maledicta (ll. 14–16). Sura is shown trapped inside of a rhombus, flanked by snakes and hence completely confined by the spell’s powers. Another noteworthy case is a curse that appears to target a certain Petronius Cornigus (cf. 33), if we accept that the curse was targeting the same individual as the magical figurine that was found together with the tablet in the innermost container. In the curse tablet, the victim is shown standing with his arms flat against his sides. In addition to being tied up with dozens of ropes, the victim is flanked by two snake-like monsters that represent the powers meant to destroy him.

victim Porcellus is shown mummified and with his arms crossed or tied, whereas in the tablets from Rome (cf. 5 and 6) we see three victims schematically represented. In the other instance from the ‘Sethian’ collection (cf. 6), we see two figures: the first is bigger, has a large insignia with a charaktêr on his chest and is surrounded by magical symbols, seven stars and charaktêres. All of this seems to be a means of representing the figure’s supernatural status. The second figure is represented upside down and is shown swaddled, which strongly suggests that this is the unborn child mentioned as one of the tablet’s victims. Finally, we must briefly mention the tablets in which neither the victim nor the deity is represented, but rather we find other elements that tend to be related to the spell itself. A good example is found in an erotic defixio from Hadrumetum: in the centre there is a representation of intertwined cords that are pierced with nails or swords, which seems to be a visual representation of the union of the defigens and his victim.185 Another noteworthy example can be seen in an agonistic curse from Carthage (Figure 4.5), on which the outline of the circus has been inscribed.186 Three more defixiones (one from Carthage and two from Augusta Treverorum/Trier) where we see a sort of basket or fence (perhaps meant to entrap the victims?) also belong to this category. Finally, a curse from Mazan (cf. 156) represents what the editors have identified as either a snake or ship, which closely resembles an image found on a curse from Corinth.187

Another noteworthy motif that is found several times in Late Antique curses is the mummification of the victim. As Gager has put it, the function of these images was ‘to anticipate and enact the desired outcome for the spell itself, to bind or in some other way harm the target’ (1992: 11). The mummification of the target, which was surely influenced by the representation of mummified Osiris, served as the ultimate way of depicting the immobility and constriction of a victim, who was left totally helpless and vulnerable to the spell’s powers. This is seen in 117, which finds interesting parallels among the ‘Sethian’ collection (cf. SV 17, fig. 4.7, left) and in a magical figurine from the fountain of Anna Perenna, which are all dated to the same period (fourth–fifth centuries CE).182 As argued elsewhere,183 these images are better understood as depicting the victim of the spell. In favour of such a hypothesis, consider the general layout of the curses (with the deities on top and the victims on the bottom of the tablet) as well as the expressions κατὰ κράβατον τιμωρίας (‘on (the) bed of punishment’) and κακῷ θανάτῳ (‘evil death’), documented in these same ‘Sethian’ curses.

4.2.2.2. Magical symbols and charaktêres The use of magical symbols and charaktêres is somewhat more frequent than the inclusion of iconography in the corpus of curses from the Roman West: they are securely found on 53 tablets, which constitutes 8 per cent of the total collection. Audollent grouped magical symbols and charaktêres together under the heading signa magica.188 Today in many cases we cannot easily distinguish one from the other, due to the lack of published drawings or photos of many curse tablets. As was mentioned in the previous section, for texts that were published well over a century ago, information about visual components is often scarce or even non-existent. Today, any type of unusual symbol found in a defixio or other magical object tends to be included in the group of charaktêres.189

As mentioned above, there are rare occasions in which the victim and a deity are both represented in an attempt to create a narrative structure that served as a sort of script for the divinity tasked with carrying out the defigens’ wishes. Hence, through his or her creativity, the practitioner shows the deity what it should do and provides it with a modus operandi.184 The curses that securely belong to this category are 117 (against the veterinarian Porcellus) and 5 (which targets the miller Praeseticius), to which we should possibly add 6 (against an unborn child, among other victims). In the first two curses, the divinity is seen in standing position and is placed above its victims, who occupy a lower plane near the deity’s feet. In 117, the

Charaktêres find their origin in the imitation of various writing systems (such as Egyptian hieroglyphs190 or letters Cf. DT 264 (for an image, see Németh 2013: 182–83). Cf. DT 233 (with an image in CIL VIII, 12504). 187  Cf. Barruol and Barruol 1963: 109. For the Corinthian parallel, see Jordan 1994: 120–21. To these, we must add a recent find from Tongres (see Bélanger Sarrazin et al. 2019). 188  See DT: p. LXXII: ‘Plures autem multo se praestant litterae singulae, sive mudae sive globulis in extremis hastis ornatae’. 189  As Gordon has maintained (per litt.): ‘at any rate, in the meantime it has become usual to use the term “charakteres” for all deliberately estranged signs, whether or not they closely resemble ordinary alphabetic signs’. 190  Cf. Zago 2010: 848, where the author claims ‘sembrano costituire un alter ego della scrittura geroglifica egiziana, di cui sopravvalutano però 185  186 

182  For the ‘Sethian’ curses, see SV 16–18 and for the magical figurine, cf. 24. 183  Cf. Sánchez Natalías 2013c: 10 and 2020d: 113–15. 184  Cf. Gordon 2002: 101–03.

25

Sylloge of  Defixiones from the Roman West from the Greek or Latin alphabet)191 and the use of a series of basic shapes, such as rectangles, rhombi, stars, bars and crosses. Generally, the main feature that distinguishes them from other magical symbols is that they have small circles on the end of lines. As Poccetti has shown, these symbols constitute ‘una simbologia grafica allo stato puro, che implica né una lettura né una recitazione, ma solo una visualizzazione’.192 Their presence in defixiones, amulets and ‘magical’ recipes attests to their range of functions and purposes.193 One of their defining characteristics is their divine nature, which several defigentes invoke. Indeed, they were thought to be sacred elements containing significant power, as is seen in the various invocations documented in several agonistic curses and phylaktéria.194

Figure 4.11. Charaktêres attested more than 15 times (Gordon 2011: 28, fig. 3).

are also instances in which charaktêres are used to divide a text into several parts.200 As mentioned above, the earliest tablets containing charaktêres come from Aratispi/Villanueva del Cauche (Hispania Ulterior, cf. 130 and Figure 4.6) and Tongres (Gallia Belgica), which have been dated to the end of the first century CE. By the second century CE, there are other examples attested in Gallia and Africa.201 In the latter, the use of charaktêres flourished until the third century CE. The proliferation of this trend is in fact so strong that it is possible to identify specific series that point to the existence of prolific magical professionals in Hadrumetum during the High Empire.202 By the end of the third century CE and the beginning of the fourth century CE, the first curses containing charaktêres begin to appear in Italy, though it was not until the fourth and fifth centuries CE with the appearance of the ‘Sethian’ and Anna Perenna collections that this visual element truly reached the height of its popularity.203 Furthermore, during the same period there are other curses with charaktêres or magical symbols from Gaul (cf. 183–84 and 190) and even further afield in Britain (cf. 446) and Pannonia (cf. 530).

Most curses containing charaktêres date between the second and fourth centuries CE (cf. Gordon 2002: 90, n. 75), though two tablets from Aratispi/Villanueva del Cauche (Hispania, cf. 130 and Figure 4.6) and Tongres (Gallia Belgica), which have been dated to the first century CE,195 expands the normal chronological limits associated with the use of charaktêres. In a careful analysis of the charaktêres found in a considerable number of magical objects (the PGM, gems, defixiones and phylaktéria), Gordon has catalogued around 1,000 different symbols (2011: 27ff.). Within this large collection, Gordon has singled out two important groups, the first of which consists of 12 symbols that are attested more than 15 times (see Figure 4.11); the second consists of 30 symbols that appear around 10 times. The rest of the charaktêres, he concludes, ‘were simply produced on one occasion. That is why there could be so many of them’ (2011: 30).196

4.2.2.3. Other voces magicae On rare occasions, defixiones from the Roman West contain other voces magicae alongside the use of iconography or charaktêres. This less common category includes items such as the onómata barbariká, palindromes or letter sequences containing the vowels of the Greek alphabet. Although they may have been incomprehensible, these elements could be pronounced and hence (in contrast to the charaktêres) played an important role in the oral portion of the actio magica.

As is the case with defixiones that contain iconography, those that are adorned with charaktêres or magical symbols are mostly found in Africa and Italia; indeed, examples from elsewhere are few and far between (see Table 4.1).197 When used in curse tablets, charaktêres are generally given a place of prominence, like the heading (i.e., before the actual cursing formula begins), alongside imagines magicae198or even alone on one side of the tablet.199 There

As is the case with charaktêres, the onómata barbariká are not only found in defixiones, but are also used in other magical objects, such phylaktéria, gems, papyri and ostraka. The onómata barbariká are first attested in the Mediterranean in the first century BCE and then reached the height of their popularity during the High Empire.204 According to the Neo-Platonic philosopher Iamblichus (Myst. VII, 4–5), the onómata barbariká, which were

il carattere ideografico rispetto a quello fonetico’. Likewise, Frankfurter 1994: 207 states that ‘the heavenly “letters” in fact denote Egyptian hieroglyphs’. For a contrary view, see Mastrocinque 2004: 98 and Gordon 2011: 27, 30–31 (who notes some exceptions). 191  Cf. Gordon 2011: 28 and 30. 192  Cf. Poccetti 2002: 42; Gordon 2011: 26–27. 193  Cf. Zago 2010: 847, n. 60. 194  Good examples include the tablets mentioned by Gager (1992: 11) from Apamea, Beth Shean and Hebron (= Gager 1992, no. 6, 77 and 106, respectively). We can also add the curses from Hadrumetum DT 273 (ll. 11–13) and 274 (ll. 11–13; cf. Németh 2011: 101), among others. For the phylaktéria, see the examples discussed by Mastrocinque 2012: 537–38. 195  To the editor’s surprise, who writes, ‘Litterarum formae saec. I vix posteriores videntur, quamquam charakteres saec. II demum frequentiores apparuerunt’ (CIL II²/5, 729). For the tablet from Tongres, see Bélanger Sarrazin et al. 2019. 196  Cf. also Gordon 2002: 90. 197  There is one defixio from Hispania (130), five from Gaul (166, 183– 84, 190 and 203) and one from Pannonia (530). 198  As in the cases of 6, 25–26, 117–18 (where they appear on the chest and pubic region of the deity invoked), etc. 199  E.g., 20, 166, 184, 259 (?), 446, and 448.

E.g., 30, 183 and the curses from Hadrumetum DT 275–83. Cf. 166 and AE 1907, 68 and 69 (from Hadrumetum). 202  The following Carthaginian curses share the same date: DT 243–44, 247, 258, 260, Jordan 1988: no. 3 and AE 1996, 1717 and 1718 as well as the following from Hadrumetum DT 264–65 and 272–83. For the sequence of charaktêres as evidence for the existence of officinae magicae, see Németh 2011. 203  See 6, 18, 20, 22, 24, 29–30 and 117–18. 204  For a discussion of dating and diffusion throughout the Mediterranean, see Zago 2011: 115f., esp. n. 20. For the onómata barbariká in general, see Graf 1995: 211–15. 200  201 

26

Inscribing the Defixiones immutable and eternal, had a divine origin, since they were consubstantial with the gods. As M. Zago has argued, these onómata are the product of ‘una babele linguistica, che ricorre a diversi sistemi di lingua (ebraico, egiziano, greco, aramaico, babilonese, copto) e li fonde in neologismi’ (2010: 835). It is precisely in this otherness that the power of the onómata resides. As Poccetti argues, these onómata were intentionally distinct from common languages (or even human language) to assimilate themselves to superhuman beings (2002: 35). Accordingly, the inclusion of the onómata in defixiones (something which was always intentional) was meant not only to increase the potency of a given spell, but also to improve communication between a defigens and the deities invoked. Thus, the use of this extraordinary special language was thought to guarantee a spell’s efficacy.

remained largely unchanged over a long period of time, are hardly found in the corpus of curses from the Roman West. In fact, δαμναμενευς only appears in three tablets from Hadrumetum and (so far) one from Gaul.212 Another category of voces magicae that ought to be mentioned separately consists of palindromes, which, of course, are the same whether read from left to right or right to left. Gordon has suggested that the mysterious origin of these words should be located in the Greek world (2002: 86). So far, in the corpus of curses from the Roman West there are only a few examples of such palindromes: αραρακαραρα and (its variant?) αραραχαραρα are attested in two curses from Carthage that date to the High Empire.213 The classic αβλαναθαναβλα, for its part, appears on a series of containers from Anna Perenna (dated between the fourth and fifth centuries CE; cf. 25–27, 35–36).214

In the Roman West, the onómata barbariká are securely found in 30 tablets that mostly hail from Carthage205 and Hadrumetum.206 Additionally, there are several tablets of interest from Italy (cf. 22, 117–18) and Gaul (cf. perhaps 160). Besides the three Italian examples (dating to the fourth and fifth centuries), the rest of the relevant cases belong to the High Empire. Generally speaking, the onómata are written using the Latin alphabet, just like the curses that they accompany.207 That said, there is a small, but noteworthy, group of nine tablets in which the onómata barbariká were written in the Greek alphabet, while the remainder of these texts was written in Latin with the Latin alphabet: by switching the alphabet, these professional practitioners had found a further way to difference between the onómata and the language found elsewhere in the same curses.208 While the majority of these onómata come from unknown languages or have been so distorted from their ‘original’ form that they are now utterly unintelligible, specialists have determined that some of the words that are found (mostly) on defixiones from Carthage have an Egyptian origin.209

Even less common is the use of alphabetical sequences in the defixiones from the Roman West. In this group we must highlight the vowels of the Greek alphabet. Due to their association with the seven planets, it was believed that the ritual recitation of these seven sounds summoned all the sounds of the cosmos.215 Such sequences are found on two agonistic tablets (that also have iconography): the first comes from Carthage and dates to the third century CE:216 the tablet contains the partial sequence ααα/εεε/οοο. The second curse, from Rome (cf. 5), belongs to the so-called ‘Sethian’ collection (fourth or fifth century CE). Along with the visual representation of the deity invoked, we find the Greek vowels arranged in sequences of seven and more or less forming a plinthion.217 Finally, another alphabetical series that deserves mention is the Latin alphabet itself. According to Ogden, the complete or partial recitation of the alphabet was thought to be powerful. There are two such examples found on curses from the Roman West. The first, an erotic curse from Maar (cf. 173, Gallia Belgica), contains the entire alphabet written on a small clay jar

As is well known, the ephesia grammata belong to the larger group of onómata barbariká. Although they are often associated with Ephesus due to their name, scholars such as D. Ogden have argued that these grammata should actually be linked to the Babylonian word epesu (‘enchant’).210 According to Androcides (apud Clem. Al., Strom. VIII, 45.2), there are six of these powerful nomina: ἄσκιον, κατάκιον, λίξ, τετράξ, δαμναμενεύς and αἴσια.211 It must be noted that the ephesia grammata, which

Cf. DT 267 (l. 7), 268 (l. 8), AE 1968, 620 (l. 4) and the curse from Autun (166, col. III, l. 2). 213  They are, respectively, DT 243 (ll. 11–26) and 253 (ll. 2 and 67). For the first, we must mention that this palindrome as well as other onómata barbariká are arranged in decreasing triangles. As R. Martín Hernández (2010: 495) has maintained in a discussion of certain PGM recipes, ‘most spells in which a word… have been written in a decreasing triangle… are spells in which the intention is to make something disappear.’ This idea jibes well with the aims of DT 243, which is directed against a group of agitatores. For the performative value of triangles, see Mastrocinque 2010: 5ff. For the carmina figurata, which involve the construction of triangles (klima), rectangles (plinthion/pterugion) and wings (pterygion), see Frankfurter 1994: 199–200; Gordon 2002: 85; Poccetti 2002: 53–56, who has underscored the interesting combination of the acoustic effect with the visual impact generated by the layout of the writing. 214  Cf. 26 (container c, correctly written in Greek: Αβλαναθαναλβα), 27 (container b, in the Latin alphabet: ablatanabla (!)), 35 (container b, in the Latin alphabet: ablatanbla (!)) and 36 (container b, in the Latin alphabet: ablatanabla (!)). If the editor’s reconstruction is correct, we should add 30, which reads abl[anatabla] (l. 7; cf. Blänsdorf 2010c: 46–49). For this theonym of Hebraic origin and disputed interpretation, see Mastrocinque 2004: 99–100. 215  See DT: p. LXXIIIf.; Frankfurter 1994: 200–01; Zago 2010: 846, n. 57. 216  Cf. Jordan 1988: no. 3. 217  For the plinthion, cf. n. 213 above and Mastrocinque 2010: 7, who maintains that the magical power of this shape rests in the multiplication of the letters. 212 

Specifically, these are DT 218, 227, 233, 243–244, 251, 253 and 255; AE 1933, 234–35 and Jordan 1988: no. 3. 206  Cf. DT 264–67, 270, 272–74, 286, 288, 289 and 292–95. 207  Like in the following from Carthage DT 244, 251, AE 1933, 234–35 and Jordan 1988: no. 3, the following from Hadrumetum DT 264–67, 270, 272–74, 286, 288–89, 292–95, 22 and 160. 208  Like in the following curses from Carthage: DT 218, 227, 233, 243, 244, 253 and 255 as well as in 117–18. 209  Zago 2010: 831, n. 2 identifies those from DT 243, 251 and 253, to which we should add those found in 22. 210  Cf. Ogden 1999: 47. For its origin and relation to Orphism, see Bernabé 2003. 211  See Gager 1992: 5–6 with further bibliographical references, and Zago 2011: 116–17, esp. n. 22. 205 

27

Sylloge of  Defixiones from the Roman West and upside down when compared to the text of the curse proper.218 The alphabetical sequence found in Aquae Sulis/ Bath (cf. 207), however, is partial and reads abcdefx. ̣

or she would pierce or fold the tablet. If any divinity is mentioned in the text, it was in the role of a witness who was meant to supervise the actio magica (cf. Faraone 1991a: 5). 2. ‘Prayer Formula’ refers to instances in which the defigens uses second-person imperatives (e.g., κάτεχε) to spur the deities invoked to take action and carry out the defigens’ request. The deities found in this type of curse tend to be chthonic and referred to with their respective epithets. By the High Empire, other components could be included, such as iconography or voces magicae (see section I.4.2.2). As Faraone has argued, this type of formula is securely attested from the fourth century BCE onwards. As was the case with direct binding formulae, prayer formulae could include the ritual manipulation of the physical curse. 3. ‘Wish Formulae’ (or similia similibus formulae), attested in the Attic curses starting in the fourth century BCE, are traditionally linked to the principle of sympathetic magic; more precisely, he connects it to the principle of persuasive analogy (cf. Tambiah 1973: 212). This type of formula establishes a comparison between a characteristic or phase of the actio magica and the victim due to ‘a strong belief in the persuasive power of certain kinds of formulaic language’ (Faraone 1991a: 8). Often, similia similibus formulae are followed by an explicitly stated wish, in which the victim is the grammatical subject.

4.3. Formulae defigendi In a chapter dedicated to the writing of curses, it is necessary to save space for a discussion of the formulae found in defixiones. Since this is one of the questions that has attracted the greatest amount of scholarly attention, it is not my intention to propose a new system for classifying these formulae, as others have done. Thus, the following pages outline the main taxonomies established by C. Faraone, A. Kropp and D. Urbanová, which have been the most recent and/or influential approaches to the issue.219 The oldest and also most simple cursing formulae consisted simply of inscribing the victim’s name on a lead sheet, which could be further specified with the addition of several other identifying details, such as the patronymic, matronymic or specification of the victim’s profession).220 This technique is likely based on the principle that names were considered to be fundamental parts of the people to whom they refer, with all the magical implications that such a tight connection implies. The first attestations of names (or lists thereof) are found in Greek curse tablets dating to the fourth or fifth centuries BCE and continued to be used until the fourth century CE, though at a lower rate.221 That said, at the same time that these minimalistic ways of cursing enter the archaeological record we also have evidence for the use of more complex formulae, which Faraone has divided into three categories. It is worth laying out this classificatory system in some detail.222

Kropp, for her part, has built upon the research of Kagarow and Faraone and has deeply analysed the formulae defigendi from the perspective of speech-act theory.223 As a result she has proposed four main categories:

1. ‘Direct Binding Formula’ refers to cases in which a defigens uses a first-person singular verb (e.g., καταδῶ, παραδίδωμι or καταγράφω) to curse a victim (occasionally also making pointed reference to vital organs or possessions). In such instances, the defigens also carried out a ritual act of manipulation in which he

1. Manipulation Formulae (Manipulationsforme):224 this category, which is analogous to Faraone’s first category (‘Direct Binding Formula’), includes formulae in which there is explicit reference to the ritual manipulation of the tablet, which would have physically taken place during the actio magica. For Latin texts, common verbs in this type include defigo, implico, describo, immergo and ligo as well as its compounds alligo, colligo, deligo and obligo. 2. Committal Formula (Übergabeformel):225 this typology, also derived from Faraone’s first category, includes both formulae for consigning victims to the deities invoked as well as formulae that refer to the deposition of the defixio (with verbs like commendo, devoveo, do, dono, mando, trado, etc.). The ‘consignation’ could be carried out through either explicit or implicit performative expressions, which would immediately take effect after being pronounced.226 The first kind depends on

In full, though with an error in the sequence p r r, for p q r. See Faraone 1991a; Kropp 2008: 137–79 and 2010. Additionally, we must add the contributions of Audollent 1904: pp. LVIff. (‘verba devotoria’) and LXXIVff. (‘formulae defigendi’); Kagarow 1929: 28–49 (for Greek curse tablets); Cesano 1961²: 1570–76; Tomlin 1988a: 63–74 (for the formulae founded at the sanctuary of Aquae Sulis/Bath; this classification can be applied to the whole corpus of curses against thieves from Britannia); Graf 1995: 117–25; Ogden 1999: 26–28; Curbera and Jordan 2007; Murano 2010 and 2012 (for Oscan formulae); Urbanová 2018: 60–77 (for the structure, auxiliary formulae and a summary classification of the formulae founded in Latin curse tablets) and 102–26 (where the author explains her own typology, which results from the mix of Faraone’s and Kropp’s work). 220  This should be understood to a certain extent as a magical version of the proverbial idea pars pro toto, according to which the name that is used for referring to a certain person is essentially linked to him/her. Therefore, and given that the name is a vital part of the individual, it is possible to harm someone through their names (cf. Rossenblat 1977: 27). Ogden has also noted that ‘initially, the very act of writing a name, of “freezing” it permanently in lead, could in itself have been considered a way of tying it down by comparison with the transience of uttering it’ (1999: 10). Also see Gager 1992: 5; Piccaluga 2010: 13–20. 221  E.g., 19, 178, 180, 360, 362 and 368. 222  See Faraone 1991a: 4–10, who based his taxonomy on that of Kagarow (1929: 28–49) for Greek defixiones. 218  219 

Kropp 2008: 137–79 and 2010. Cf. Kropp 2008: 145–46, table 4.1 and 2010: 361. 225  Cf. Kropp 2008: 146–49, table 4.2 and 2010: 362–64. 226  Cf. section I.4.1, esp. n. 98. The recitation of this sort of expression in a determined context could lead to concrete effects in the real world. A classic contemporary example is the phrase ‘I now pronounce you man and wife’ which has an immediate legal effect. Cf. the classic volume by Austin 1962. 223  224 

28

Inscribing the Defixiones first-person singular verbs in the present indicative that are followed by an accusative. A good example is found in the tablet from Baelo Claudia/Bolonia: Isis Muronem tibi c ̣ọnmendọ furtu(m) meu(m) (cf. 128, ll. 1–3). Implicitly performative expressions are those that betray a more reserved and humble attitude on the defigens’ part, who seeks to ‘shield’ him- or herself from the dangerous consequences of directly communicating with a deity.227 These formulae are often in the passive voice, as can be seen in a curse tablet from Ratcliffe-on-Soar (cf. 349 A, ll. 1–2), which reads: donatur deo Iovi Optimo Maximo. 3. Request Formula (Aufforderungsformel):228 this category is similar to Faraone’s second type, though it differs in so far as the stress is placed on the act of asking for something (which does not entail any relationship between the person who asks for help and the one who is asked). This replaces the idea of ‘prayer’, which implies a hierarchical relationship.229 This category includes all formulae that are used to solicit the intervention of the evoked deities and that offer up instructions.

consists of formulae that use a performative expression to directly reference to the act of cursing itself. So far, there is only one known example from Aquae Sulis/ Bath: cf. 304, 1. 1: ẹxẹcro qui involaverit.... Recently, Urbanová has based her classification on Faraone’s and Kropp’s typologies, bringing together the pragmatic, semantic and syntactic aspects of the various formulae attested in Latin curse tablets and ‘prayers for justice’.231 Following the work of R. Risselada (1993), she has distinguished three main categories within which there are up to nine different types of formulae. Furthermore, these individual formulae can be found on their own or in conjunction with another type, forming ‘simple’ or ‘combined’ curses. 1. ‘Simple nominal list of cursed people’ (or formula 0): this merely consists of a list of the name of the victim(s), which can be accompanied by further details (e.g., matronymics, professions, etc). A good example would be 99: Secundula aut qui sustulet. 2. ‘Direct cursing formula’: this is equivalent to what Faraone called ‘Direct binding formula’ and Kropp referred to as ‘Manipulation formulae’ and also includes what Kropp called ‘Committal formulae’. This kind of formula contains a first-person verb in the present indicative followed by: a. A ‘predicate of cursing’ or ‘formula 1’, in which we find verbs like defigo, ligo, deligo, obligo, describo, together with the name, body parts or properties of the victim(s); such as in 339, ll. 1–2: Tretia(m) Martia(m) defico et illeus vita(m) et me(n)tem... b. A ‘predicate of committal’ or ‘formula 1a’, in which we find verbs for handing over the victims of the curse (such as do, dono, voveo, mando, trado, commendo, desacrifico, defero), together with the name and/or body parts of the victim(s). On occasion, we also find the mention of divinities. A good example is 56, ll. 4–5: Dii i(n)feri vobis com(m)e(n)do illius memra colore(m)... 3. ‘Invoking formulae’: this category partially corresponds to Kropp’s request and committal formulae (which were based on Faraone’s first two categories). ‘Invoking formulae’ include any formula in which invocations take on a more substantial role. For Urbanová, there are various subtypes, depending on the types of verbs, their moods and grammatical constructions (e.g., verbs of requesting and giving, with purpose clauses, imperatives, wishes or similia similibus formulae). 232

Generally, orders are given directly or through performative expressions (again either explicit or implicit). The first option entails the use of a second person imperative (e.g., 118, ll. 8–13: Fistu occidit̂e inicat̂e). In performative expressions, we normally find verba rogandi (like rogo, oro, etc.), as can be seen in a curse from Emerita Augusta/Merida 120, ll. 1–6: Dea Ataecina Tur̂ ibrig(ensis) Proserpina per tuam maiestatem te rogo oro obsecro uti vindices quot mihi furti factum est. In instances when the text does not directly name the deity to whom it is addressed (who, in all likelihood would have been named orally during the ritual deposition of the curse), practitioners could solicit a deity’s intervention either directly or by means of an analogy; these two strategies are equivalent to Faraone’s third category (‘Wish Formula-Similia Similibus’). A direct petition is expressed as a desire with a subjunctive verb and the indication of the desired results (e.g., 210, ll. 3–6: qụi illas involavi(t) ut mentes sua(s) perd[at]). Analogies are spelled out with similia similibus formulae, which are usually introduced by quomodo and meant to set in motion a process that would lead to a favourable outcome (e.g., 157, ll. 1–3: quomodo hoc plumbu(m) non paret et decadet sic decadat aetas membra vita). 4. Curse Formula (Fluchformel):230 this category, which does not have a clear analogue in Faraone’s typology,

As I mentioned before, it is not my intention to propose yet another system for taxonomizing every single type of formula according to linguistic categories, an endeavour that at times can over complicate things. What is clear is that the work Faraone did back in 1991 remains extremely influential for later authors, such as Kropp and Urbanová, whose discussions fruitfully take further considerations

227  Cf. Kropp 2008: 148–49 and 2010: 366, where the author notes that ‘the defigens articulates his awareness of the status difference between him and the god invoked, insofar as he gives his request a non-peremptory character and, at the same time, qualifies his own position as dependent and inferior’. 228  Cf. Kropp 2008: 149–55, 174–76, table 4.3 and 2010: 365–69. 229  Kropp 2010: 365, notes, ‘the term “prayer” implies a hierarchical relationship between orant and addressee, with the latter occupying the dominant position. I adopt the more general term “request”… [because it] has the advantage of not excluding any kind of communicative setting’. 230  Cf. Kropp 2008: 155 and 2010: 369.

See specifically, Urbanová 2018: 127–36. On the similia formulae, see Franek and Urbanová 2019a and 2019b.

231  232 

29

Sylloge of  Defixiones from the Roman West into account without radically changing the picture. While Urbanová’s classificatory system is by far the most detailed, it is also the least user friendly. Furthermore, the scholarly desire to classify and organize can also make us forget that, as Gordon has recently pointed out in relation to the curses from Italy, most of the texts do not follow a general model (2019b: 421). This same idea can be further extended to the entire Roman West, where most of the texts were written by individual practitioners. Consequently, instead of thinking strictly in terms of formulae, Gordon has proposed that ‘it seems ... more useful to distinguish between different degrees of religious competence, drawing upon different kinds of cultural resources in an attempt to gain illocutionary authority vis-à-vis the (implied or expressed) addressee(s)’ (2019b: 421–22).233 Given the general priorities of this volume to elucidate the cultural and archaeological contexts of curse tablets, in the following catalogue I resist the urge to fit each text into a neat box based on linguistic and formulaic features. That said, I do use useful phrases such as ‘similia similibus or persuasive analogy’, ‘all-inclusive’ formulae, as a useful shorthand to refer to common features across the corpus.

233  On this, see also Gordon 2019a on religious competence and the pragmatics of cursing.

30

5 The Manipulation of Defixiones 5.1. Introduction: technical considerations

places that were inaccessible to ‘mortal’ eyes). Another possibility is that practitioners folded the tablets as part of an attempt to protect the text from being physically damaged (either scratched or otherwise harmed). However, the archaeological record also yields examples that complicate this hypothesis: uninscribed curse tablets which were nevertheless folded before being deposited.239 In these instances, the practitioners had no text to protect or conceal; perhaps they were just performing what was perceived to be ‘the usual’ ritual procedure: orally performing the curse and a series of physical gestures, with the folding of the tablet being the only one visible in the evidence. Finally, we are left to wonder if folding could simply have been a pragmatic means of easily transporting a defixio from where it was manufactured to where it was finally deposited.

On certain occasions, after inscribing the tablet, the practitioner would physically manipulate the curse, either as part of ritual or for more practical reasons. There are three main types of manipulation that are found in the archaeological record: folding, rolling and piercing. Additionally, tablets could be folded and pierced or rolled and pierced.234 Although these are by far the main ways of manipulating defixiones, Table 5.1 shows that there were other modi operandi of great interest. As the data collected in the table shows, over half of the known defixiones from the Roman West were manipulated after being inscribed. Unfortunately, these results do not give us the complete picture, since many editiones principes do not comment on the absence or presence of any signs of ritual manipulation. As a result, we simply do not know whether the tablets were deposited ‘open’ (i.e., without having been deliberately manipulated)235 or whether they had been manipulated, but archaeologists did not record such details, which has resulted in the loss of important information.236

Generally speaking, tablets were folded in half longitudinally, as can be seen in the defixiones from Nomentum/Mentana (cf. 49, where the tablet was folded only once) or Cremona (cf. 105, folded twice) to give just two examples. Sometimes after being folded longitudinally, a tablet was then folded transversally, thus being reduced to an even more compact unit. A curse from London (cf. 342, folded twice) and one from Savaria/Szombathely (cf. 531, folded seven times) provide good examples of tablets being folded multiple times.

5.2. Folded tablets With a total of 172 examples, folding is the most frequently attested type of manipulation (see Figure 5.1). Folded tabulae come predominantly from Italia, Gallia and Britannia.237 This technique is already attested in the fourth century BCE (with the Cirò Marina curse, cf. 75)238 and endured until the fifth century CE (with the Giuncalzu tablet, cf. 114). As Tomlin has pointed out, ‘its purpose may have been… to conceal what had been written’ (1988a: 84). That said, the fact that other inscribed tablets were simply left ‘open’ without undergoing any type of manipulation suggests that folding was not meant to hide the texts (which, after all, were quite often deposited in

While these are the most common types of folding, the archaeological record contains other, less conventional ways of folding a tablet. A curse tablet from Aquae Sulis/ Bath (cf. 244) for example, was folded four times in towards the centre, while another from Old Harlow (cf. 353) was folded in the ‘accordion’ style; another from Dalheim (cf. 203) was folded thrice, first in half before each side was folded in towards the centre. The number of folds attested ranges from one to nine, and we cannot give any general rule concerning how many times a text would have to be folded. Presumably, the size of a particular lead sheet played a role in influencing each practitioners’ choice in this regard.

234  In general, see Gager 1992: 18; Ogden 1999: 13–14; Bailliot 2010: 77–81; Martin 2010: 20. More specifically, see the analysis of binding gestures found in McKie 2018: 118–20. 235  Ten per cent of tablets are ‘open’. Some of these defixiones were left ‘open’ due to the characteristics of the medium. The marble plaques or altars (cf. 4, 57, 120) the slate plaques (cf. 72, 146–47), lamps (cf. 3), busts (cf. 15), cinerary urns (cf. 16), lead containers (cf. 24–27 and 35– 36), ostraka (cf. 113), jars (cf. 173) and tegulae (such as the one from Thysdrus/El Jem, see Foucher 2000) could obviously not be folded. 236  This is the case for over a quarter of the tablets, some of which were found during nineteenth-century excavations. 237  Of these, 134 were just folded, 25 were folded and pierced and 12 were folded and further manipulated (see Table 5.1). These 171 tablets comprise 27 per cent of the total corpus. 238  We could also add the tablets from Laos/Marcellina (cf. 78–80), which likewise belong to the group of fourth- or third-century Graeco-Oscan curses.

As mentioned above, a tablet could both be folded and pierced. Thus, some curses (cf. e.g., 49–51, 56 and 73) were clearly pierced before being folded and then deposited. Conversely, we have examples of tablets that were folded 239  Cf. for instance, 159, which was accompanied by other uninscribed but folded curse tablets. Also, at the sanctuary of Mercury in Uley, 140 tablets were found, mostly ‘rolled and then flattened’ (Woodward 1993: 113). Of these, only 86 were uninscribed. That said, it seems clear enough that the practitioners still felt that it was necessary to roll them in order to comply with all the steps of the cursing ritual.

31

Sylloge of  Defixiones from the Roman West Table 5.1. Defixiones classified according to provenance (It. = Italia, Afr. = Africa, Hisp. = Hispania, Gall. = Gallia, Brit. = Britannia, Ger. = Germania, Raet .= Raetia, Nor. = Noricum, Pann. = Pannonia) and type of manipulation (F= folded, R= rolled and N= nailed). The total number of manipulated curse tablets in each province is tallied in the right column. Prov.

F

R

N

F&N

R&N

Alia

Total

1

1: F and tied. 48: attempt of F and N. 52: F and tied. 53: R around a nail. 61: F and tied. 88: F and tied. 92: attempt of N. 97–98: R and crushed. 103: F, cut, R and N. 109: R around a nail.

67/119

2

AE 1907, 69: F or R.

44/88

It.

14

13

Afr.

8

33

Hisp.

4

5

3

1

1

132: F and closed with an iron pin. 139: cut. 148: F and cut.

17/30

Gall.

18

7

4

–

–

175: thrown in a fire. 188: hammered.

31/55

–

206: cut. 269 and 309: F and cut. 338: N and thrown in a fire. 432: punctured with a pointy object. 433: hammered. 441: R, with label.

104/256

39/59

Brit.

72

6

14

12

14

7

Ger.

14

11

4

2

–

483: R and stuck to a nail. 488–489, 498: R and thrown in a fire. 493: R, F and thrown in a fire. 504: F and cut. 508: F or R. 512: beat with a blunt object.

Raet.

–

1

–

–

–

–

1/6

Nor.

–

–

–

–

–

–

0/1

Pann.

4

–

–

1

–

528: thrown in a fire.

7/9

530: R around a nail.

5.3. Rolled tablets After inscribing a defixio, the PGM recommend to ‘roll up the lamella in the usual fashion’,240 a habit that scholars have generally analysed as a reflection of epistolary practices.241 In this regard, we must mention papyrus rolls more generally, which were normally rolled up. In a different vein, authors such as Ogden have interpreted this type of manipulation as a symbolic act that was meant ‘to achieve a sympathetic binding and perhaps a sympathetic confusion of the tablet’s contents’ (1999: 13). While his proposal is doubtlessly suggestive, the lack of concrete evidence means that it remains little more than a tantalizing theory. Figure 5.1. Folded defixio (cf. 485, before opening. Landesarchäologie Mainz).

and then pierced. The best examples of this practice are the five first-century BCE tablets from Rome that are now housed at Johns Hopkins University (cf. 10–14). These curses were all folded up together and then transfixed with a single iron nail.

PGM XXXVI, 234 (translated by R.F. Hock [apud Betz 19922]), which recommend τὴν λάμναν ἑλιξας κατὰ τρόπον and also PGM VII, 463–64, which advise to ‘roll (the curse) up and throw it into the sea’ (ἕλιξον καὶ βάλε εἰς θάλασσαν; translation by E.N. O’Neil [apud Betz 19922]). On these passages, see also section I.6.3. 241  Thus, Gager 1992: 18 and Ogden 1999: 13. 240 

32

The Manipulation of  Defixiones The curse from Roccagloriosa (cf. 76), for example, was first pierced and then rolled up.245 Though it itself was not pierced, the curse from Caistor St Edmund was rolled up and then deposited with an uninscribed lead label that had been pierced with an iron nail (cf. 441).

Figure 5.2. Rolled defixio (cf. 486, before opening. Landesarchäologie Mainz).

5.4. Pierced tablets Nails have undoubtedly become an emblematic part of defixiones: no element of these tablets better embodies the ideas of binding, joining or constricting. The piercing of a tablet was a symbolic gesture through which the spell itself materialized, as the etymology of the verb defigo (‘to fix, fasten down, render immovable’) makes clear. In fact, as Cesano has maintained, ‘il chiodo trapassante la tabella significava che doveva di necessità avvenire ciò che vi era scritto; come dal fautore della defixio il chiodo era confitto nel piombo così il defisso venina indissolubilmente legato’ (19612: 1562).246 Only one recipe from the PGM gives specific information about what kind of nail should be employed, recommending the use of a ‘copper nail from a shipwrecked vessel’.247 In this case, the nail’s provenance (i.e., from a context related to the biaiothánatoi [see section I.6.2.1]), it would undoubtedly have been endowed with considerable magical power. In the archaeological record, however, we most commonly find iron nails with unknown origins and of various lengths. At times, they are still found transfixing the defixio,248 while at others they are simply found nearby.

Figure 5.3. Defixio rolled around a bird bone (Courtesy of GDKE, Landesarchäologie, Mainz).

Based on the available data, there are 92 confirmed instances of tablets being rolled up in the Roman West,242 dating from the second half of the fourth century BCE (the GraecoOscan curse from Castiglione di Paludi [cf. 77]) to the fourth century CE (a curse from Hamble [cf. 451], for example). This type of manipulation was especially important in the North African cities of Hadrumetum and Carthage. Most curses from these two sites were inserted into cinerary urns via the libation-tube. To do so, the practitioner had to roll the tablet up ‘comme un cigare’.243 It seems that in the case of these North African tablets, the rolling up of a defixio should not be understood in terms of sympathetic magic but rather be seen as practical solution to assist with the goal of depositing the tablet in a certain context.

Despite the nail’s iconicity, in the Roman West there are only 72 known instances of defixiones closely associated with nails, which amount to a relatively small percentage of the overall corpus.249 These examples date from the fourth century BCE (the Graeco-Oscan curse from Roccagloriosa [cf. 76]) to the fourth or fifth century CE (the defixiones from Augusta Treverorum/Trier). Most of the pierced tablets date to either the Republic or High Empire. While there are examples of pierced curses from each province in the Roman West, the greatest concentration is found in Italia and Britannia.

Generally speaking, tablets were individually rolled without anything placed inside before being deposited in a chosen place (see Figure 5.2). When two curses dealt with a similar theme, however, they could be rolled up together, as happened with two defixiones from Corduba/Cordoba (cf. 125–26), which were found deposited inside a cinerary urn. On rare occasions, defigentes rolled their tablets around things like iron nails or bones, objects which could have been used during the cursing ritual (see Figure 5.3).244

We can distinguish between three different categories of curses associated with nails: those that were deposited alongside nails,250 those that were still transfixed with the nail in situ at the time of discovery251 and, finally, cases in

Since it was physically difficult to pierce a defixio that had already been rolled up, piercing would precede rolling.

245  For other examples, see the North African curses DT 253 and 264 and perhaps the defixio from Carmo/Carmona (cf. 129). 246  Cf. Piccaluga 1983; Gager 1992: 18; Ogden 1999: 14; Gordon 2015: 158–61. 247  PGM VII, 466: ἥλῳ κυπρίνῳ ἀπό πλ[ο]ίου νεναυαγηκότος (translated by E.N. O’Neil, apud Betz 19922) 248  Such as those found with the tablets 10–14, 119, 124, 345, etc. 249  Of these, 37 were pierced, 25 folded and pierced, 4 rolled and pierced, and 6 belong to the group of otherwise manipulated defixiones. In total this is 11 per cent of the entire corpus. 250  Such as a curse from Groß Gerau (cf. 483), which was rolled up and deposited together with a nail, though it does not appear to have been pierced. 251  Such as 119 (which still has both nails) and 345 (pierced five times with four nails still preserved in situ). In addition to these, see Gager 1992: 19 for a discussion of tablets nailed up in the Carthaginian circus (also see section I.6.5.3 n. 336), and also the Carthaginian tablets ‘fixées

242  Specifically, there are 76 rolled tablets, 4 rolled and pierced, as well as 12 in which rolling is combined with another type of manipulation (see the preceding table). This amounts to 14 per cent of the total corpus. 243  Delattre 1898: 218. 244  The tablets rolled around a nail come from Lilybaeum/Marsala (109, third century BCE), Ostia (cf. 53, second century CE) and Carnuntum/ Petronell (530, fourth century CE), while one of the curses found in the sanctuary of Isis and Magna Mater (Mogontiacum/Mainz) was rolled around a bird bone (see Fig. 5.3). Here we should also include Choppard and Hannezo’s observation that some curses from Hadrumetum: ‘On trouve quelquefois [lamminam]… roulée sur un ossement’ (apud DT, p. CXV, n. 3).

33

Sylloge of  Defixiones from the Roman West

Figure 5.5. Pierced defixio from Pompeii (cf. 72, detail; Ministero dei Beni e delle Attività Culturali e del TurismoSoprintendenza Pompei).

Figure 5.4. Pierced defixio (cf. 119; Getty Museum).

which a tablet was pierced before or after being inscribed but now the resulting hole provides the only evidence for this ritual manipulation (i.e., we do not have the nail). This final category is the largest.252

on both sides of the tablet. The curse from London was pierced nine times: the largest of the holes comes between the victim’s name and the word vita(m) (l. 2), which is certainly no accident.

Within this larger practice of piercing tablets, there are two motivating reasons that can explain the practice: magical considerations linked to ritual or more practical reasons. In the first group, the holes coincide with significant words or formulae found in the text as a way to stress the symbolical and unavoidable bond formed by the victim(s) and the spell. A good example of this phenomenon is the curse from Mediolanum Santonum/Villepouge-Chagnon (cf. 160), in which we find a similia similibus formula referring to the victims that reads: sic tra(n)specti sin quomodi ille (Tab. II, ll. 4–5). This phrase, referring to the moment when the puppy was harmed (or perhaps even killed) is followed by the hole left by the nail. Similarly, a defixio of unknown provenance and dated to the first century CE (cf. 119, Figure 5.4) refers to the victims with the expression quorum nom[ina] hic sunt perea[nt] (col. II, ll. 3–5). In this phrase, the hole coincides with the word nomina. Given that victims’ names were considered a vital part of an individual, like an additional body part, piercing the word for name enacts the insoluble link between the victims and the spell.

Currently there is only one known case of a curse tablet that was pierced for magical and practical reasons: the Pompeiian curse against Faustus (cf. 72, Figure 5.5) which was written on a slate plaque that was pierced in its four corners in order to attach it to the front of tomb 23OS in the Porta Nocera necropolis. The tablet, however, also has a fifth nail, which is larger and placed between the words meae innocentiae (l. 6). In this case, and as Elefante noted (1985: 433), the fifth nail serves as a means to stress and reinforce the importance of this phrase. As mentioned above, certain tablets were only pierced for more practical reasons. There are instances of tablets being pierced at their corners, perhaps so that they could be mounted and ‘displayed’. This practice can be clearly seen in a curse from Caere/Cerveteri (cf. 95), dated to the second or first century BCE. The defixio, which contains a list of personal names, was found in the dromos of a tomb in the Banditaccia necropolis. The tablet was pierced in its four corners, as a means of mounting it on the wall (there is no indication of a magical reason for these holes). In a similar vein, two other first-century BCE curses from Hispania (cf. 129 and 133, both without an exact archaeological context) were pierced in their corners. Finally, it is also worth noting a different curse tablet from Pompeii (cf. 71) that was written on two tablets that resemble tabulae ceratae. This curse was pierced twice in order to make it into a type of diptych. In all these instances, rather than piercing the tablets with magical purposes in mind, it seems that the practitioners were imitating the usual practices of displaying inscriptions employed in the realm of public epigraphy (e.g., in the cases of Caere/Cerveteri and Hispania) or of common writing practices (e.g., in Pompeii).

We must also add two British curses, one from Aquae Sulis/Bath (cf. 213) and the other from London (cf. 339), to this group. As Tomlin has noted,253 the former is an opisthographic tablet that contains lists of names on both sides, though they are ordered differently. This allowed the practitioner to pierce the name Anniola (A, l. 6 and B, l. 1) aux parois du cippe à l’aide d’un clou de cuivre qui en transperçait tous les plis’ (Delattre, apud DT, p. CXV) as well as the two from Hadrumetum that were ‘trouées par un gros clou, encore adhérent à la plaque, comme si on avait voulu les maintenir appliquées contre une des faces du tombeau’ (Cagnat, apud DT, p. 361). 252  If the nails used in rituals were as valuable as is implied at PGM VII, 466 (cf. n. 247), it is possible that professional practitioners chose to use such nails on later occasions instead of just depositing them. 253  Cf. Tomlin 1988a: Tab. Sul. 8.

34

The Manipulation of  Defixiones 5.5. Other modi operandi As Graf has aptly put it, the inscribed text of a defixio is itself the representation of the victim (1995: 128), a sort of effigy that could be harmed and injured in various ways. Admittedly, examples of this are rare, but those that we do have are undeniably fascinating. In all likelihood, the cruelty shown to the physical tablets took place while the author of the text recited verbal curses. This type of ritual performance should be interpreted in terms of sympathetic magic: based on the principle of persuasive analogy, the victims would suffer the same damage that the defigens inflicted upon the physical curse tablet.

Figure 5.7. Melted defixio (cf. 493, before opening; Landesarchäologie Mainz).

of the sanctuary of Magna Mater so that the victims would melt just as the lead melts (488, B, ll. 2–5: sic illorum membra liquescan(t) quatmodum hoc plumbum liquescet). This modus operandi (throwing the tablets into the altar) was the usual in this sanctuary.

Two curses from Classis/Classe (cf. 97–98), which only contain the names of their victims were discovered rolled up and crushed inside cinerary urns. Other curses (such as 103, from Ateste/Este, see Figure 5.6, 139, 148 and 504) were cut with a sharp object after being inscribed, which led to different types of damage. The lead tablet from Emporion (cf. 139), however, was so mutilated that only a small fragment of the tablet has been preserved. This extreme act of chopping up the tablet could be seen as a symbolic method of dismembering a victim.

Finally, we must turn to a small group of tablets that were tied up before being deposited. All such examples hail from Italian necropoleis and have been dated between the second century BCE and the first century CE. A curse tablet from Volaterrae/Volterra (88) was discovered with two other defixiones (89 and 90), which, according to Lanzi, were inserted in the first one, which is folded as a dyptich and tied up with a lead strip.254 A similar case can be observed in a tablet from Rome (cf. 1), which was folded and ‘filoque ferreo clausa’ after being inscribed.255 An already mentioned curse from Pompeii (cf. 71) was written on two lead sheets in a way that is reminiscent of tabellae ceratae; after being inscribed, these two were closed like a diptych, connected with two small nails and then tied together with a strip of inscribed lead (which today is lost).256 In a similar fashion, two curses, one from Cumae (61) and the other from Ostia (52), were folded in half, pierced (once and five times, respectively) and then tied with a metallic(?) thread.257 In these instances, the authors of the texts were probably thinking of the curses as letters to the underworld; accordingly, they employed the same system for sealing their texts.258

In the same vein, other texts were beaten after being inscribed. A curse from Uley (cf. 433), for example, was stabbed with a pointed object until it was totally disfigured, while a text from Mogontiacum/Mainz (cf. 512) was pounded eight times with a blunt object. A defixio from Augusta Treverorum/Trier (cf. 188) was beaten with a hammer with the result that today its text is barely legible. Throwing texts into a fire is also a well-documented type of manipulation. Good examples of this practice are seen in a curse from London (cf. 338) and another from Emona/ Ljubljana (cf. 528), whose edges show tell-tale signs of smelting). Additionally, we ought to add three defixiones from Mogontiacum/Mainz (cf. 488, 493 [see Figure 5.7] and 498), all of which were thrown into the fire on the altar

Apud CIE 52 a. According to CIL I², 1013. 256  For a possible formal parallel, cf.CIL IV, 9252. 257  These threads have not survived. However, based on documented parallels, it seems likely that they were made of lead or iron. 258  Letters written on wood (such as tabulae ceratae) were usually notched and enclosed with a string. See Sarri 2018: 83. 254  255 

Figure 5.6. Cut defixio from Ateste/Este (cf. 103, detail; Soprintendenza per i Beni Archeologici del Veneto. Museo Atestino di Este).

35

6 Deposition Contexts: Where the Curses have been Discovered 6.1. Introduction: putting ‘context’ into context

The second question to be considered is the cultural significance attached to the various contexts. Any analysis must be carried out by taking culturally and historically relevant frameworks into account, since the principals who wrote (or commissioned) the curse were people of their time, who were likely—as we shall see—able to perceive and interpret these contexts from different viewpoints. In this vein we can examine the tablets found in two British amphitheatres (from Isca Silurum/Caerleon and London, cf. 337 and 343, respectively), which can serve as a good counterpoint to the agonistic tablets from the Carthaginian amphitheatre. In these tablets, which are both defixiones in fures, Diana and Nemesis are urged to punish the guilty party. Given the content expressed in these curses, the choice of depositional context does not seem to be linked too much with the proximity to the victims (in both cases the identity of the thief was unknown) or with the morbid undertones of the amphitheatre; instead the choice seems to have been taken along different lines: the amphitheatre was a sacred space dedicated to the very divinities whom the defigentes evoked. In short, we can see how these categories can quickly breakdown and, in the process, lose much of their interpretative value.

Once the curse tablet had been manufactured, ‘the final stage of a tablet’s activation was its deposition’, as Ogden (1999: 15) has put it. As was the case with other parts of the ritual, defigentes did not choose a place to deposit a curse at random, but rather took a series of considerations (e.g., the desired outcome or evoked divinity) into account for increasing the tablet’s efficacy. Relying on literary and archaeological sources, traditional scholarship has differentiated four main contexts where defixiones could be deposited: necropoleis (often taken as the archaeological context par excellence for curse tablets), sanctuaries (both dedicated to chthonic and uranic deities), aquatic spaces or areas somehow linked to the curse’s victim.259 As is the case with my typology, this classificatory scheme is problematic and upon further scrutiny can be woefully simplistic. By neatly separating every single deposition into one of these four categories, scholars can overlook two key questions: the possible overlapping of these categories and the cultural significance that these contexts could have in different times and places throughout the history of the Roman West. Let’s start with the first question. The proposed contexts do not necessarily fit into neat categories, and therefore, they cannot be completely differentiated from one another. At times, deposits are superimposed, creating new, mixed and polysemic contexts that were conducive to aggressive magic for various reasons. This is certainly the case with amphitheatres or circuses, since both of these were closely connected to the victims (who are often explicitly named as gladiators, charioteers or even horses), but also were spaces inhabited by the ‘restless dead’ and stained with miasma, a characteristic that is also shared by funerary contexts.260 The professional practitioners who wrote a series of curses directed against the venatores from the amphitheatre in Carthage were in all likelihood aware of this double valence: they deposited their defixiones in a cella within the amphitheatre complex where, according to Audollent, ‘caesorum gladiatorum trahebantur corpora’ (1904: CX, cf. DT 246–54). Accordingly, this space is doubly conducive for depositing a curse since it was extremely close to its victims and was also linked to the spirits of the aoroi and biaiothánatoi.

Accordingly, the present study of contexts seeks to analyse the different places where defixiones were deposited, while taking these interpretative problems into account. I do this while continuing to work within the traditionally sanctioned categories (i.e., funerary, aquatic, sacred and spaces linked to the victim). The goal is to examine each of these spaces through the information gleaned from the sylloge along with the different temporal and geographical data about the curses. As hinted at above, however, there are several cases (which are clearly marked) in which is difficult or even somewhat arbitrary to assign a defixio to a specific type of context, either because it can comfortably be placed within two categories (‘mixed’ deposits) or because a more culturally sensitive analysis of the context can change its meaning and distinguish it from superficially similar cases. 6.2. Funerary contexts Funerary contexts are understood as those spaces that were permeated with death and inhabited by the so-called restless dead (cf. Johnston 1999), all of which meant that these spaces were suffused with miasma. While within this context, tombs and necropoleis undoubtedly play the leading role, we cannot overlook other spaces like certain cellae in amphitheatres (see section I.6.5.3) or specific aquatic spaces (see section I.6.3).261

259  For a general discussion of magical contexts, see Wilburn 2013: 40–53; for curse tablets in particular, see Audollent 1904: CX–CXVII; Cesano 1961²: 1587–89; Preisendanz 1972: col. 5, Vb and col. 20, IV; Gager 1992: 18–21; Graf 1995: 123; Ogden 1999: 15–25; Kropp 2008: 90–94; Bailliot 2010: 81–83; Martin 2010: 25–28; Urbanová 2018: 58–60. 260  See Tremel 2004: 31–33; Le Glay 1990: 222; Gómez Pantoja 2007: 69.

On curse tablets in funerary spaces, see a preliminary Spanish version of this section in Sánchez Natalías 2012b. For another discussion of 261 

37

Sylloge of  Defixiones from the Roman West Table 6.1. Defixiones discovered in funerary contexts, categorised by date and provenance (It.= Italia, Afr.= Africa, Hisp.= Hispania, Gall.= Gallia, Brit.= Britannia, Ger.= Germania, Raet.= Raetia, Nor.= Noricum, Pan.= Pannonia). The final line shows the total number of tablets from funerary contexts against the total number of tablets from that province Provenance

Date

It.

Afr.

Hisp.

Republic

21

2

5

High Empire

23

64

5

Late Antiquity

6

Unknown

3

6

53/119

72/88

Total

Gall.

Brit.

10 1

10/30

The practice of using necropoleis to deposit defixiones has traditionally been explained through recourse to two different theories, which in my opinion are not mutually exclusive. The first theory views funerary spaces as sites impregnated with miasma. Given that one’s name was considered a fundamental part of any individual in Antiquity, by engraving someone’s name on the tablet and depositing it in a space tinged with miasma, the victim would automatically become infected with that very miasma from the necropolis.262 The second theory stresses that funerary environments were privileged sites of communication with the chthonic powers and the spirits of the deceased who could thus be recruited to act as a practitioners’ assistant.263 The validity of this second theory has been questioned by J. Curbera, who has argued that, although the oldest curses from Sicily originate in necropoleis,264 invocations to the gods or epistolary formulae only appear at a later date. Therefore, for Curbera this suggests that any attempt to communicate with supernatural forces was not a top priority for practitioners at that time (1999a: 160).

Ger.

Raet.

Nor.

Pan.

13

3

1

4

3/6

1/1

4/9

1

1

1

1

12/55

1/256

15/59

Therefore, it stands to reason that despite the fact that the earliest curse tablets in the Roman West do not preserve such invocations,266 these may have been uttered by the practitioners during the ritual act in an attempt to communicate with the deities and spirits of the ‘restless dead’ who presided over the site. Something similar is at play in the sanctuaries where some, but certainly not all, curse tablets explicitly mention the deities to which these spaces were consecrated (see Table 8.1 on page 64). Nevertheless, we still can assume that the defigens was addressing the same deities that presided over the site. All that is to say that Curbera’s observation may not invalidate this second theory about communicating with the deities or the restless dead. 6.2.1. Geographical and temporal data about funerary deposits In the Roman West, necropoleis and tombs are among the most well-attested depositional contexts. As Table 6.1 shows, nearly a third of the curse tablets studied in this volume come from such contexts.

It should be pointed out, however, that from its earliest stages, the actio magica was performed both through writing and oral recitation (see section I.4.1). In these rituals, the recitation of certain formulae is recommended both for the invocation of the numina and for the enumeration of the practitioners’ desired outcomes.265

The geographical distribution of these defixiones is even more noteworthy, seeing that most of them originate in major cities in Italy and North Africa, such as Rome, Carthage and Hadrumetum. Over one third of all the curse tablets found in necropoleis come from these two North African cities.267 The popularity of this context is later reflected in the PGM, whose recipes recommend the use of the graves of biaiothánatoi as a place for

funerary contexts that takes into account Greek curses from the Roman West, see Alfayé and Sánchez Natalías 2020: 45–48. 262  See Ogden 1999: 16 and Curbera 1999a: 160. In Antiquity the name was considered a fundamental part of the individual. By virtue of the performative power of script (see section I.7.2), etching a name on a curse tablet automatically led to that individual becoming infected by the miasma. 263  Faraone 1991a: 9 claims that ‘The rationale behind the placement of the tablets in graves and chthonic sanctuaries has been similarly misinterpreted. It is true that contact with the coldness and inertia of corpses provides the motivation of some similia similibus formulae, but … these formulae seem to rationalize … the practice of communicating with the gods or the dead.’ 264  In particular, see Curbera 1999a, nos 15 (dating from between the sixth and fifth century BCE or to the fifth century BCE), 1, 4–7, 12 (whose chronology goes back to the fifth century BCE), 2–3 (dated to the fifth– fourth century BCE), 8 (from the late fourth century BCE), 47 (dating from the third century BCE), etc. For an updated corpus of Sicilian curse tablets, see Sommerschield 2019. 265  Here I provide several good examples: PGM III, 30 ff., which provides the recitation of formulae during the deposit of the tablets; PGM IV, 332 ff., where it is recommended to recite the formula written on the defixio and to utter another one during its deposit; PGM IV, 1747 ff., which

urges the practitioner to engrave the formula and to recite it afterwards; PGM IV, 2235 ff., where the recitation of the text takes places after being written; PGM V, 319 ff., where a series of formulae are uttered while piercing the tablet and another series are spoken during its deposit. 266  For just a few examples, see 49–51, 55, 71, 78, 94, 95, 125–26. 267  Thirty-five per cent of the total number of defixiones found in a funerary context originate in this province. When it comes to Hadrumetum, according to Audollent (1904: 360), ‘inter sepulcra coemeterii romani iuxta viam quae ducit Kairouan effosi diversis temporibus erutae sunt omnes hadrumetinae tabellae (263–98)’. In Carthage, 35 of the 43 defixiones were found in a funerary context: 27 were found in graves and 8 (DT 247–54) were found in the cella of an amphitheatre where the corpses of gladiators would have been stored. This, then would constitute a ‘mixed’ archaeological context, undoubtedly linked to the funerary world (see section I.6.5.3 and n. 341 below). Beyond these cities, three other curses deserve mention: one from Cirta/Constantine (see DT 300), one from Thysdrus/El Jem (see Foucher 2000) and one from Naro/ Hammam-Lif (see Audollent 1910).

38

Deposition Contexts deposition.268 The preference for using funerary contexts as sites for depositing curse tablets is also reflected in Italy and Hispania, where a third of these curse tablets were found. Nevertheless, it should be pointed out that further north this type of deposition is not as common as sacred spaces. Thus, just one tenth of the curse tablets found in funerary contexts originate in provinces like Gallia or Germania, the figure drops to almost zero in the case of Britannia, where (so far) we only know of a single curse tablet that was deposited in a grave (cf. 345). This is all to say that practices and preferences were not uniform throughout the Roman West. In fact, this diversity deserves explanation.

archaeological evidence. The evidence is organized in a crescendo, beginning with the deposits furthest away from the dead to those closest to the corpses. Within the PGM, at least four different formulas refer to the time and method of depositing the tablet in a funerary context. The Great Magical Papyrus of Paris contains three of these. The first urges the practitioner to deposit the tablet next to the grave of a person who had died prematurely or violently.270 This necessarily involved a certain proximity with the burial site (and probably depositing the curse in the dead of night so as to avoid detection), though without direct contact with the deceased. And indeed, this is documented in numerous cases where the tablets were placed near funerary monuments, mausolea or graves; in such cases, however, the defigens always maintained some distance between the deceased and him-/herself. For some examples of this, we can cite some defixiones, such as a curse from Rome (against Caecilia Prima, found next to a mausoleum in Via Ostiense, cf. 48), one from Copia Thurii/Sybaris (discovered in the outside wall of a mausoleum, cf. 82), a tablet from Aventicum/Avenches (found, still folded, behind a funerary monument, cf. 481) and another from Celti/Peñaflor (found about five metres away from two graves in a cavity wall, cf. 131).

As shown in Table 6.1, chronologically, funerary contexts were used for depositing curse tablets from a very early period: in the Republican period, Italia, Africa and Hispania are already represented. Even if it is true that in Africa and Hispania there are not many early cases, we should stress that in Italy there are more tablets from funerary contexts dating to the Republic than to the High Empire. During the High Empire, funerary contexts were used as a site for deposition throughout the Roman West, except for in Britannia where sanctuaries were by and far the favoured context. Unless archaeology should prove the opposite, during Late Antiquity the use of necropoleis fell out of favour and sacred and/or aquatic deposits became much more prevalent.

Yet, when approaching a burial area, practitioners could employ more imaginative strategies. For example, in Mazan (Gallia Narbonensis, cf. 156), next to grave no. 1 in the necropolis of Saint-Andéol and near the head of the deceased (but outside the tomb), an urn was found containing the remains of a small bird under which a Latin curse tablet had been placed. According to the editors of this defixio and taking the characteristics of the deposit into account, this ensemble may have been buried at the same time as the deceased. While the tablet in question, of unknown date, contains a fragmented text that is still unpublished, the presence of this small bird appears to be an offering to the chthonic deities addressed to by the practitioner. Another interesting case comes from Roßdorf (Germania, cf. 480), where a curse tablet was found together with organic remains and a small iron nail inside of a cinerary urn. Perhaps, as Nuber has suggested (1996: 241 and 243, n. 3), the remains found inside of this urn might have been ousiae (i.e., something that had belonged to the curse’s victims). Be that as it may, the most interesting aspect is that the urn was deposited upside down (hence ‘spilling out’) at the very limits of the necropolis, a liminal area that was conducive to magical practices, since these border zones were often reserved for burying more marginalized members of society (e.g., prostitutes or gladiators; cf. Hope 2000: 116–18).

6.2.2. The various modi operandi The fact that tombs and necropoleis were spaces presided over by the deities of the underworld, populated by all kind of spirits and filled with miasma made these spaces into an ideal environment for depositing curse tablets. Thus, necropoleis are used for not only depositing the oldest defixiones from the Roman West (cf. 68, 74–75 and 78), but there are also a good deal of tablets dated to the High Empire, irrespective of the presence (or lack thereof) of invocations to chthonic deities or references to Hades. Although depositing a curse in a funerary environment was thought sufficient for achieving the desired outcome without the need of more ‘invasive’ practices, the archaeological record gives examples of defigentes going above and beyond.269 In the following paragraphs, we will examine these different practices and comparing them to the recommendations found in the PGM, which are later in date, but the only source with which we can compare the 268  See, for example, PGM IV, 333 and 2215–16 or PGM V, 330–40. Also note other relevant rites: PGM VII, 466, where the practitioner is told to write a defixio using a nail from a shipwreck; PGM XV, 8–10 and 16–19, where it is suggested that the curse be placed in an aquatic context since those who died prematurely inhabited this space; PGM XXXVI, 370, where the mouth of a dead dog is used as the place for depositing the tablet. There are other such examples. 269  Despite the information that we do have, we must also stress that there is lamentably little information about the archaeological context of many curses that were discovered during nineteenth-century excavations. For many of these, even if we know that they were found in necropoleis, we have no further details about many important questions, such as the type of tomb, the exact position of the curse or whether it was touching human remains.

Even though depositing curses near graves was both effective and feasible for the activation of the tablets, some PGM IV, 333, where it is recommended τίθεσαι ἡλίου δύνοντος παρὰ ἀώρου ἢ βιαίου θήκην (to place it ‘as the sun is setting, beside the grave of one who has died untimely or violently’ [Translated by E.N. O’Neil, apud Betz 19922]). 270 

39

Sylloge of  Defixiones from the Roman West 1755 in the entrance to the lower chamber of a hypogeum and among broken urns and ashes. Similar instances are documented in other tombs, such as no. 336 from the necropolis of Banditaccia (Caere/Cerveteri, cf. 95), where a defixio was fixed to the wall in the dromos. Another interesting case comes from tomb no. 10 from the Pompeiian necropolis located outside the Porta Stabia, where a defixio which looked like a diptych (cf. 71) was found on the ground at the entrance to a grave.

rites in the PGM actually recommend burying them inside the grave of a person who had died prematurely.271 As if that were not enough, there are some occasions in which practitioners were given even more precise instructions: sometimes we learn the exact moment when the act of deposition ought to take place (‘when the moon stands in opposition to the sun’)272 or even the ideal depth for the deposit (‘four fingers deep’).273 While these prescriptions are hard to document archaeologically,274 it should be pointed out that over half of the defixiones originating in necropoleis were found inside a grave (both burial and incineration graves). Access to such graves not only depended on the characteristics of the site but also on the skill (and at times, the audacity) of practitioners. As a result, there are a wide range of practices that deserve a more minute analysis.

There are, however, other examples in which there was even distance between the practitioner and the deceased: curses were at times deposited next to the grave goods, as in Maar (Gallia Belgica), where a defixio was found in 1893 written on a clay pot next to the incineration urn, remains of oil lamps and three coins (cf. 173). There may be another instance of this from Faviana/Mautern (cf. 526) if we accept Scherrer’s reinterpretation of the context (1998: 77–79). According to this scholar, the area where the tablet was deposited was not a sanctuary (as Thaller 1948 had proposed) but a grave, as is suggested by the presence of copious ashes, 2 pieces of pottery, 13 iron nails and some 40 coins. The defixio in question was used as the cover of a small jug containing the remains of organic materials, sand and some human hair-like fibres, which have been explained by Faraone and Kropp (2010: 388, n. 27) as the remains of a burnt magical figurine. Although this hypothesis cannot be proven, it is doubtlessly suggestive.277

The safest of all the options is broadly documented in North African necropoleis, where libation tubes connecting cinerary urns with the outside world were used to ‘send’ the curses inside a grave. Thus, according to Chopard and Hannezo, defixiones from Hadrumetum were rolled up and then slid down through these conduits. Sometimes this did not work so well: in fact, there are cases in which cinerary urns had been stuffed with so many tablets that defixiones got stuck in the tube itself.275 This modus operandi is also attested at the necropoleis of Bir-ez-Zitoun and Bir-elDjebbana (Carthage), where defixiones were rolled up ‘like cigarettes’ (Delattre 1898: 218). It must have been so common to deposit defixiones in this manner that even a cippus was found whose tube had deliberately been sealed with a perforated lead sheet through which only fluids could be introduced. This device demonstrates an effort on the part of the deceased’s relatives to prevent this deeply rooted praxis from affecting the grave of their loved one.276

Nonetheless, in an attempt to create a close link between the remains of the deceased and the tablets, the latter were at times deposited inside the cinerary urns themselves. This modus operandi was unquestionably riskier than others, given that it not only involved entering the columbarium or opening a grave, but also the direct handling of an urn (at least when it lacked a libation tube for ‘sending’ the lead tablet inside). Of the total number of curses found in funerary contexts, about 20 were discovered inside cinerary urns. These defixiones mostly come from Italia, although we also know of examples from Africa, Hispania, Germania and Gallia. These tablets are generally found mixed together with the cremated ashes and bones. There are, however, some noteworthy exceptions, such as the defixio from L’Hospitalet-du-Larzac (cf. 158, Gallia Narbonensis): in 1983, during the excavations of grave no. 71 at the necropolis at La Vayssière, which included an impressive collection of 40 vases, a Gaulish defixio was discovered. The curse, inscribed on two tablets, had been carefully placed on top of a cinerary urn serving as a sort of a lid that ‘protected’ the charred human bones within. The fact that the tomb was not ransacked during Antiquity strongly suggests that the tablets were deposited at the same time as the burial, as was the case from Mazan mentioned before (cf. 156).

If there was a lack of libation tubes, practitioners at times opted for bolder procedures, breaking into the funerary monument and depositing curses at a grave’s entrance. This riskier method was used in the case of three curses from Volaterrae/Volterra (cf. 88–90) that were found in 271  PGM IV, 2215–16: κατορύξεις δὲ ἐπὶ ἀώρου θήκην τὴν λεπίδα ἐπὶ ἡμέρας γ´. 272  PGM IV, 2220–25 (translated by H. Martin, Jr. apud Betz 19922): καὶ καταχώσεις εἰς ἀώρου μνῆμα σελήνης οὔσης διαμέτρου ἡλίου. 273  PGM V, 333 (translated by M. Smith, apud Betz 19922) ὄρυξον ἐπὶ δ´ δακτύλους καὶ ἔνθες. 274  It is obviously impossible to determine the exact moment when a defixio was deposited. Likewise, it is only in rare instances that we can know whether someone was deemed to have died prematurely. See, for just some examples, 96 (deposited next to the burial of a child), 125– 26 (where two tablets were found inside a cinerary urn containing the charred remains of a child), or a curse tablet from Naro/Hammam-Lif (found next to a child’s grave; cf. Audollent 1910, 136–41). 275  Chopard and Hannezo 1894: 194: ‘étaient au préalable roulées et que l’on glissait pour les faire pénétrer jusqu’aux ossements contenus dans le récipient; souvent même, en raison de l’encombrement du récipient et du tube, la lamelle de plomb était arrêtée dans le tube et mélangée aux os calcinés’. In the case of the defixiones found in Hadrumetum, DT 275–84 are believed with a high degree of certainty to have been found in this context. 276  Delattre 1898: 218: ‘…était fermé par une lamelle de plomb percée de petits trous formant passoire et ne permettant que l’introduction des liquides’.

In this regard, it is worth noting an interesting parallel from the sanctuary of Anna Perenna (Rome), where a small ceramic jug was found with the remains of bones and parchment and had been sealed with an apparently uninscribed lead sheet (Polakova and Rapinesi 2002: 39 and 49).

277 

40

Deposition Contexts At times, when curse tablets were deposited in burial graves, the practitioner deliberately sought to put the curse in direct contact with the remains of the deceased. This intentional and morbid contact has been documented on rare, though noteworthy, occasions when the skull or the chest of the deceased was chosen as the place to deposit a curse tablet. Indeed, some defixiones were found placed directly under the deceased’s skull, such as in Minturnae/Minturno where the tablet was actually nailed to the back of the cranium (cf. 56). According to Delattre (1888: 158), in Carthage, some curses were found on top of the skulls of two decapitated individuals. The chest of the deceased could also be a good place to deposit a curse tablet. There is an interesting case from Durocortorum/Reims (cf. 174), where a Gallo-Roman burial was found in 1894 that contained the remains of a person whose skull was malformed (hence, probably, the use of this grave for magic purposes) and on whose chest (next to the left shoulder), a coffer was found which contained 14 coins and a rolled defixio (Jadart 1901: 66). Another case from Paris (cf. 165) could be added, where according to Vacquer (1879: 111) ‘une lame de plomb mince, pliée en deux et reposant sur la poitrine du mort’ was found.

coldness of the water would symbolically ‘freeze’ the victim of the curse (Ogden 1999: 23).279 This theory is quite compelling for the defixiones found in the wells from the Athenian agora, among which we frequently find formula like ὡς ταῦτα τὰ ὀνόματα ψύχεται, οὕτω καὶ ᾿Αλκιδάμου ψυχείσθω τὸ ὄνομα... (‘as these names grow cold, so too let Alkidamos’ name grow cold’).280 Here, as the water cools the tablet, the practitioner asks that the name and hence victim himself be frozen and paralysed. Even though this explanation is clearly correct for describing curses from this specific context (the Athenian wells), it would be a mistake to try to give the same explanation to the curses from the Roman West, where (so far) no similar formulae have been attested. In a different vein, other authors have linked aquatic contexts to Hades. As M. Martin (2010: 27) states, water can come from the depths of the earth, where, thanks to its perpetual movement, it was thought to be linked with the chthonic deities who lived below the earth’s surface.281 There is another theory put forth by authors like Wünsch, who suggested that defixiones were deposited in aquatic settings since shipwrecked sailors resided there (1898: IV, col. 2).282 This interpretation is also recorded in the PGM, where some recipes recommend depositing tablets in water, thanks to the widespread belief that the souls of drowned men wandered the bottoms of the sea since they were unable to reach Hades.283 In the PGM, it is recommended that curses be thrown into rivers or the sea ‘late in the evening or in the middle of the night’284 or ‘before sunrise’.285 Another possibility was to ‘glue it to the dry vaulted vapor room of a bath’,286 to ‘throw it into the heating chamber of a bath’,287 or deposit it ‘where there is a stream or the drain of a bath, having tied a cord [to the plate] throw it into the stream or into the sea’.288 At least in these recipes from PGM VII, it is clear that different

As can be seen, the archaeology attests to a great heterogeneity of practices. This diversity can be explained in terms of the defigens’ attempt to do his/her best to deposit a curse in a necropolis. That said, during the High Empire, in places like Carthage and Hadrumetum, the existence of professional practitioners provides us with a different pattern, that is, the repeated use of the very same tombs over and over as places for depositing defixiones (something similar also took place in the Late Antiquity in cities such as Rome with the ‘Sethian’ corpus). In general, even if the recommendations recorded in the much later PGM are attested in the archaeological record, the imagination of actual practitioners went well beyond suggestions found in the literary record.

In a similar vein and for a discussion of the verb καταψύχω, see Guarducci 1978: 255 and Jordan 1985b: 241, n. f, where he argues that in curses from the Athenian wells this verb ‘may refer to the chilling effects of the waters in the wells’. That said, in SV 16 (from a necropolis) the verb alludes to ‘the chill of the lead itself’. For the idea of depositing a defixio in a well so as to ‘freeze’ the victim’s name, see the brief mention in Jordan 1990: 437. 280  Translated by Jordan 1985b: no. 6, ll. 27–29. The same analogy is found in Jordan 1985b: nos 1 (ll. 16–18), 2 (ll. 13–15), 3 (ll. 13–15), 4 (ll. 21–25), 5 (ll. 10–12), etc. 281  See also Guarducci 1978: 242 and Graf 1995: 123. 282  On this, also see Fox 1912b and Cesano 1961²: 1589. 283  See Stramaglia 1999: 194–95, where there is a discussion of the explicit passage of Achilles Tatius V, 16, 2: λέγουσι δὲ τὰς ἐν ὕδατι ψυχὰς ἀνῃρημένας μηδὲ εἰς Ἅιδου καταβαίνειν ὅλως, ἀλλ’ αὐτοῦ περὶ τὸ ὕδωρ ἔχειν τὴν πλάνην, that is: ‘They say that the souls of those who have met their end in the deep never go down to Hades, but wander in the same spot about the face of the waters’ (Translated by S. Gaseke, Loeb 1984). 284  PGM VII, 435. Translated by M. Smith, apud Betz 19922. 285  PGM VII, 420. Translated by M. Smith, apud Betz 19922. 286  Translated by E.N. O’Neil, apud Betz 19922, PGM XXXVI, 75: κόλλα εἰς τὸν ξηρὸν θόλον τοῦ βαλανίου. 287  PGM VII, 469: βάλε εἱς ὑποκαυστήριον βαλανείου. Translated by E.N. O’Neil, apud Betz 19922. 288  Translated by M. Smith, apud Betz 19922, PGM VII, 436: ὅπου ῥοῦς ἐστιν ἢ παραρέον βαλανείου (...) βάλε φέρεσθαι εἰς τὸν ῥοῦν (ἢ εἰς θάλασσαν). 279 

6.3. Aquatic contexts The phrase aquatic context is generally used to refer to any space that is composed of water, such as rivers, seas, fountains, wells, etc. Even if scholars have traditionally used this category as if it were straightforward and easily separated from the other contexts in which defixiones were deposited, it is actually one of the more problematic contexts, not only because it intersects with other categories but also because the coherence of the category itself is challenged by the diverse cultural meanings that water could have in Antiquity.278 Scholars have explained the use of aquatic contexts in three different ways. The first one links this space with the principles of sympathetic magic. In other words, the 278  A different version of this section, which included Greek curse tablets and magical figurines from the Roman West, has already been published (see Sánchez Natalías 2019b).

41

Sylloge of  Defixiones from the Roman West aquatic contexts (i.e., the stream, the bath, the sea) were seen as equivalent places for depositing tablets.289

that we cannot determine a date for these texts means that we do not know whether the practitioner deposited the curses in a space that was still in use (perhaps linked to the victims?) or had been abandoned (perhaps inhabited by ghosts who could help with the curses’ activation?).293

The reason that the PGM recommend practitioners to use this depositional context can be deduced from erotic spells (agōgai) collected in PGM VII, which can be dated to either the third or fourth century CE. One curse tells the reader to the curse ought to be written ‘with a copper nail from a shipwrecked vessel’290 and subsequently thrown into the sea. In this case, the nail used to write the curse is an element that links aquatic and funerary contexts. The second relevant spell urges the reader to use a lamp whose wick is made ‘of [the hawser from] a wrecked ship’.291 As was the case with the previous recipe, the material used in the ritual is an element that is linked to both funerary and aquatic contexts, each of which brings its own magical valence, as Wünsch and his contemporaries had already noted.292

As far as the aspects that could potentially favour communication with infernal deities, we ought to highlight (though not without some reservation given that difficulty involved in deciphering the text) a Gaulish curse from Arverni/Chamalières (Gallia Aquitania, cf. 163) which was deposited in the ‘Source des Roches’, a Gallo-Roman sanctuary where there was an aquatic cult. The text invokes andedíon uediíumi diíiuion risun-/ artiumapon arueriíatin’ (ll. 1–2), which Lambert translates as ‘Au nom de la bonne force des divinités chtoniennes, j’invoque Maponos d’Avernion’.294 That said, the fact that the tablet was deposited in a space that could have been a sanctuary dedicated to Maponos could also explain this invocation.

Provocative though these three hypotheses may be, they are all difficult to verify archaeologically. Theories concerning the value of aquatic settings in terms of sympathetic magic or their ability to summon the infernal gods or restless dead can only be judged by examining the textual remains of the defixiones, which generally omit any reference to this issue. A few exceptions, however, do exist. Thus, 206, deposited in the sacred spring of Sulis Minerva in Aquae Sulis/Bath, does attest to the magical valence of the aquatic contexts in terms of sympathetic magic. Written in response to a theft, this tablet asks the goddess to make the thief ‘as liquid as water’ (l. 2), thus directly alluding to the dissolution of the victim. This same idea of dissolution is found in the persuasive analogies found in three curses from the sanctuary of Isis and Mater Magna (Mogontiacum/Mainz), where the victims and their property ought to dissolve ‘like salt in water’ (cf. 490, 492–93). We must note that in these three instances the analogy is not connected to the context in which the curse tablets were actually deposited (they were tossed into the fire on the altar at the back of the temple).

6.3.1. Geographical and temporal data about aquatic deposits With the aim of clarifying the use of aquatic spaces for the deposition of curse tablets, it is fruitful to turn to a more archaeological (and less theoretical) perspective and to study the various contexts in which tablets have been discovered (see Table 6.2). 6.3.2. Aquatic contexts as sacred spaces To start, I would like to suggest that we ought to divide aquatic contexts into two separate categories: sacred and non-sacred spaces. In the first category, we find sites such as the British sanctuaries of Sulis Minerva in Aquae Sulis/ Bath and Mercury in Uley, where water played a key role, since the spring itself was the centrepiece of cultic activity. The idea that bath complexes were infested with demons is well attested in the Imperial period, which could explain why the tablet was deposited in this context (see Bonner 1932a and Stramaglia 1999: 191 ff.). On these tablets, see Alfayé 2016, who does not address their date. The same author suggests that all the curses from Bath/Aquae Sulis could be seen as coming from the same context (2016: 28). This, however, does not take into account that these tablets were thrown along with votive offerings into the sacred spring). Alfayé also identifies four other British tablets as coming from baths (336, 456–57 and 446); since the archaeological context of these four items is uncertain, I have excluded them from the present analysis. Among the defixiones that evoke the aquatic context in the terms of sympathetic magic, a curse from Salerno has traditionally and incorrectly been included (cf. 69): discovered in a grave, it reads: Locus · capillo-/ribus · expect-/at · cap-/ut · su-/um. Several authors (Mommsen, apud CIL X, 511; Fox 1912b: 305; Ogden 1999: 24; Marco Simón and Velázquez 2000: 271, n. 44) have ignored the interpuncts and singled out the letters ribus, which they have incorrectly analysed as a Vulgar form of rivus, hence falsely understanding the tablet in terms of sympathetic magic: just as hair is submerged, so too should the victim be submerged. As Mancini already noted (1884–86: 81), the reading capilloribus (por capillis) invalidates this hypothesis. Accordingly, the place that awaits the victim of the curse is the tomb where the hair and tablet were found, not the supposed brook mentioned by the above-mentioned authors. 294  According to Lambert (apud RIG II.2, L-100), -mapon(on) (l. 2 formed from *ma(k)kʷo- ‘young man’ and the suffix -ono-) is Maponos, while the following word, arueriíatin (l. 2), ought to be identified as an epithet referring to the place of god’s cult worship, in this case Maponos ‘of the underworld’. 293 

Although to date there is no further textual evidence that connects aquatic contexts to the principles of sympathetic magic, the findspot of two British curses could possibly point in this direction. More specifically, there are two defixiones from Bravonium/Leintwardine (cf. 347–48) that were discovered inside a small drain in the frigidarium of the local bath complex. Since the texts of these curses consist only of lists of names in the nominative case, we cannot determine whether the practitioner sought to ‘freeze’ symbolically the victims. Furthermore, the fact 289  Although it is possible that many North African tablets ended up being deposited in the sea or other aquatic contexts, so far we only know of four defixiones from the ‘Fontaine aux Milles Amphores’, in Carthage (on this, see Audollent 1933b). Of these, the two Latin curses were written for economic reasons. 290  Translated by E.N. O’Neil, apud Betz 19922, PGM VII, 466: γράφε ἐν ἥλῳ κυπρίνῳ ἀπό πλ[ο]ίου νεναυαγηκότος. 291  Translated by E.N. O’Neil, apud Betz 19922, PGM VII, 594–95: ποίησον ἐλλύχνιον [ἀπ]ὸ πλοίου νεναυαγηκότος. 292  See n. 282 above.

42

Deposition Contexts Table 6.2. Defixiones discovered in aquatic contexts, organised by provenance. The number of pieces from this sort of context is compared to the total number of pieces from each province Provenance

Archaeological context

SD/site

Total

Italia

19–47 57 83 92 104

Fountain of Anna Perenna. Grotto of Tiberius (abandoned pool). Lacum Fucinum (?). Fountain. Channel.

Africa

Carthage (Audollent 1933)

Fountain ‘aux milles amphores’.

2/88

Hispania

121 137

Sanctuary (pool). Beach (?).

2/30

Gallia

150–155 157 159 163 164

Fountain ‘Le Gros Escaldador’. Fountain ‘Source des Roches’. Well. Well. Fountain ‘Chaude’.

10/55

Britannia

206–335 340 347–348 354–440 441 449 451

Sanctuary of Sulis Minerva (spring). Riverbank (Thames River). Thermal complex (frigidarium). Sanctuary of Mercury (pool). Riverbank (Tas River). Little Ouse River (metal detector). Hamble’s estuary (metal detector).

224/256

Germania

479

Fountain (?).

1/59 (?)

Raetia

–

–

0/6

Noricum

–

–

0/1

Pannonia

529

Kupa River.

1/9

In these sanctuaries, curses were deposited together with other votive offerings, such as coins, lamps and paterae. While in Aquae Sulis/Bath these were deposited in the sacred spring, at Uley, they were originally deposited inside a pool located in the cella.295

33/119

Besides these manmade sites, we must mention the ‘natural’ ones, which were typical of ‘Celtic’ religion. Within this category of natural sacred spaces, we ought to mention flood plains, rivers and beaches. All these contexts have a common denominator: they were thought to be sacred places that were consecrated to aquatic divinities such as Neptune,297 Niskus298 and Savus.299

Another series of sites that fall into this category includes fountains such as the one of Anna Perenna in Rome but also at places like Arretium/Arezzo, Amélie-les-Bains and Arverni/Chamalières. These sites were cult centres for the Nymphs (cf. 19–47 and 92), the Niskas (cf. 150– 51 and 155) and Maponos (cf. 163), respectively. In these fountains, the presence of deposited coins as well as other cultic objects (e.g., altars, votives, ceramics and organic remains) suggest that all these places were indeed sacred. Probably, the ‘Fontaine aux 1000 amphores’ (Carthage296) should also be added to the list of such sacred aquatic spaces, even though the curses found there do not invoke aquatic deities. Ritual wells formed yet another subset of sacred aquatic spaces. Among these, we can point out two examples from Gallia (Montfo and Rauranum/Rom; cf. 157 and 159), where different cultic offerings were also found.

When it comes to spaces that were apparently not consecrated but were nevertheless used to deposit curses, we can mention a canal (in Altinum/Altino, cf. 104)300 and baths (at Bravonium/Leintwardine, cf. 347–48).301 The absence of any written invocation of gods connected to aquatic contexts makes it difficult to determine why these places were chosen to deposit a curse, although the invocations could also have been uttered while performing the ritual. A defixio found in Neptunus is invoked in the following texts: 441, 449 and 451. Niskus is invoked together with Neptunus in 451, a curse discovered in the estuary in Hamble. 299  As is found in the defixio from Sisak (cf. 529), which was deposited in the River Kupa and whose text evokes the river god Savus in order to send the victims to the riverbed. 300  Specifically, the defixio was discovered, according to Scarfì (1972: 55) ‘a 2 metri di profondità, all’esterno delle palificazioni di sostegno di una grande costruzione in blocchi parallelepipedi di trachite, ubicata ai limiti meridionali dell’area urbana antica’. 301  On these, see above and note 293. 297  298 

See Woodward 1993: 113 and below, section I.6.4. On this, see Audollent 1933b. For the discovery of the site, see Carton 1920: 258–68. 295  296 

43

Sylloge of  Defixiones from the Roman West 6.4.1. Geographical and temporal data about deposits in sacred spaces

an abandoned pool in the so-called ‘Grotto of Tiberius’ (from Fundi/Fondi, cf. 57) dates to the fourth century CE and was written on a reused piece of opus sectile. This curse, which contains various references to biblical passages, seems to have been intentionally deposited in an abandoned space for several reasons. First, the belief that such areas were inhabited by all sorts of spirits and ghosts made it an ideal place to deposit a curse, since these supernatural beings could assist the practitioner. Second, the persuasive analogies established in the text between the archaeological context (an area in ruins) and the physical and economic ruin of the tablet’s victims could also explain the choice of deposition.302 Finally, in the examples from Marsi Marruvium/San Benedetto and Emporion (cf. 83 and 137), the lack of information in the texts (just lists of personal names), as well as the scarce and contradictory accounts about exactly where they were found prevents us from knowing the reasons why the curses were deposited in apparently aquatic contexts.

In the Roman West, sacred spaces were beyond a measure of doubt the most common archaeological context: indeed, over a half of the tablets studied in this volume were deposited in sacred spaces. As Table 6.3 suggests, a large portion of these tablets have been found in the northernmost regions of the Roman West. The most represented provincia is Britannia, where over three quarters of the defixiones belonging to this category were discovered, mostly at Aquae Sulis/Bath and Uley. In Germania, most of the lead tablets have been found at Mogontiacum/Mainz. In Italy, the fountain of Anna Perenna in Rome deserves special attention. Besides this fountain, hardly any texts deposited in sacred contexts have been found in the Italian peninsula, Gaul or the Iberian peninsula.304 To date, there is not a single example from North Africa. Regarding chronology, the first tablets from the Roman West discovered in a temple are the Greek defixiones deposited in the Sicilian sanctuaries at Selinunte and Morgantina.305 Interestingly enough, there is a Oscan curse (cf. 76) dated to the fourth century BCE that came from the sanctuary of Roccagloriosa, which was found with the remains of another 13 lead tablets (today lost). If these tablets were, as it seems, curse tablets, this findspot would be the oldest from the Italian peninsula, dating to just after the sanctuary of Demeter Malophoros in Selinunte. As Poccetti has suggested, this should be understood as a reflection of Greek practices.306 That said, from the fourth century BCE to the first century CE, there are no known examples of defixiones deposited in sanctuaries, since (so far) in the Republican era the use of curse tablets seems to

With only these few rare exceptions, the so-called aquatic sites from the Roman West are, at heart, better understood as different types of sacred spaces. For this reason, it seems more fruitful to study them as such. 6.4. Sacred contexts Within this category, we ought to include all spaces firmly associated with a divinity, whether or not marked with a monument. Accordingly, even if temples and sanctuaries are the most common contexts within this category, we must also remember that there is a much larger range of sacred places for depositing a curse. Among these contexts, there are fountains, ritual wells and pits, amphitheatres and other ‘natural sites’ that were not built upon. As alluded to above, ‘natural sites’ are central to ‘Celtic’ religions in which cultic rites were celebrated in places like rivers and lakes. Again, this only goes to show the difficulty and danger in drawing a stark line between ‘different’ contexts.

In Italy this amounts to roughly a quarter of the total. Something similar can be observed in Gallia, where the tablets from sacred spaces amount to nearly a quarter of the total number of curses from the province. To these we would add the two tablets from Arlon and Dalheim (Gallia Belgica, cf. 203–04): the first was found during the survey of the area near a Gallo-Roman fanum, while the second was discovered by a metal detectorist near the sanctuary and the residential zone of the vicus, which means that it should be attributed to one of these two contexts. In Hispania, the curses found in sanctuaries only amount of 6 per cent of the total number from the province; we could perhaps add a further five tablets from Saguntum that treasure hunters found among the ruins of a building that some scholars have identified as a sanctuary (cf. Corell 2002: 67). Nevertheless, the lack of secure information about the archaeological context means that we ought to exclude these tablets in the interest of being cautious. From Abusina/Eining (Raetia), there is one defixio to which we ought to add four more curses that were recently published by Blänsdorf and have not been included in the sylloge (cf. Blänsdorf 2019). 305  For these tablets (dated to the fifth and first centuries BCE, respectively), see Curbera 1999a: nos 22–33 and 56–64. Faraone 1991a: 22, n. 7 provides a list of Greek defixiones deposited in chthonic sanctuaries. 306  According to Poccetti, the transmission of cursing technology from the Greek to Roman world went through the Oscan populations (cf. Poccetti 1993d: 80, 1999: 555, 2015: 383–86; on this, see section I.9.2). As different scholars have pointed out, Oscan curses contain elements taken from the Greeks (from Sicily and abroad; cf. McDonald 2015: 133–66) and also share certain elements with the early Latin texts (cf. Poccetti 2015). On the patterns of influence and the spread of curses throughout the Italian peninsula, see Vitellozzi 2019. 304 

The general popularity of sacred contexts for deposition can be attributed to the desire to communicate with the deities invoked, whom traditional scholarship has generally categorized as chthonic.303 As Ogden stated, ‘the progression from grave to chthonic sanctuary as a deposition site is easy and explicable: the chthonic gods, like the dead, dwelt under the earth, but could be expected to be much more reliable and powerful’ (1999: 23). Surprisingly, the PGM do not discuss this context save for a single possible exception in the Great Magical Papyrus from Paris: in a recipe for an agōge in which Hekate is evoked it defigens is told to deposit the tablet at a crossroad (PGM IV, 2955), a space that was suitable for this goddess.

302  303 

Alfayé 2019: 268–70 and also section I.3.3.2. See section I.7.2 below.

44

Deposition Contexts Table 6.3. Defixiones discovered in sacred spaces, organised by provenance (It.= Italia, Afr.= Africa, Hisp.= Hispania, Gall.= Gallia, Brit.= Britannia, Ger.= Germania, Raet.= Raetia, Nor.= Noricum, Pan.= Pannonia) and the site of discovery; there is also reference to the archaeological context, date and the total number of tablets from each province Prov.

Context

Date

Total

It.

19–47 76 92

Fountain of Anna Perenna. Sanctuary. Fountain.

4 –5 CE. 2nd CE. 4th BCE.

31/119

Afr.

Carthage (Audollent 1933)

Fountain ‘aux milles amphores’.

–

2/88

Hisp.

121 128

Sanctuary. Sanctuary of Isis.

1 CE. 2nd CE.

2/30

Gall.

150–155 157 159 162 163 164 171 182–183 and 189

Fountain ‘Le Gros Escaldador’. Well. Well. Sanctuary. Fountain ‘Source des Roches’. Fountain ‘Chaude’. Pit. Amphitheatre.

– 3rd–4th CE. 50–60 CE. – 1st CE. 4th–5th CE. – 4th–5th CE.

Brit.

205 206–335 340 341 353 354–440 441 447 449 451 455 458

Sanctuary of Nodens. Sanctuary of Sulis-Minerva. Riverbank (Thames River). Pit. Pit or well (?). Sanctuary of Mercury. Riverbank (Tas River). Sanctuary (metal detector). Little Ouse river (metal detector). Hamble’s estuary (metal detector). Sanctuary. Sanctuary (?).

4th–5th CE. 175–400 CE. – 4th CE. 3rd–4th CE. 75–400 CE. – 4th CE. 4th CE. 4th CE. 3rd CE. –

Germ.

487–517 518–519

Sanctuary of Isis and Mater Magna. Sanctuary of Mater Magna.

1st–2nd CE. 2nd–3rd CE

Raet.

525

Sanctuary.

–

1/6

Nor.

–

–

–

0/1

Pann.

529 531

Kupa River. Sanctuary of Isis.

1st–2nd CE. 1st–2nd CE.

2/9

SD/site

th

th

st

15/55

226/256

36/59

6.4.2. Spaces for communicating with the divine

be restricted to the Mediterranean basin, where funerary contexts were favoured (see section I.6.2.1 and Table 6.1).

As pointed out earlier, the archaeological record attests to an impressive heterogeneity of sacred spaces used for depositing defixiones. Still, we must note that the knowledge that we possess about these sacred spaces has often been limited due to the far from ideal techniques used during some excavations. Many nineteenth-century excavations, not to mention secret or illegal ones, offer little or no information about an object’s exact context with the result that it is more difficult to put together an accurate large-scale analysis. Despite these difficulties and relying on the known characteristics of these sites, it is possible to distinguish two main categories: developed sites and natural ones.

From the first century CE onwards, however, this absence is brought to a close with examples from Gallia (cf. 157 and 163), Britannia (cf. 337), Pannonia (cf. 529 and 531), Germania (cf. 487–517) and Hispania (cf. 121). During the second and third centuries, important sanctuaries emerged, such as that for Sulis Minerva at Aquae Sulis/ Bath (cf. 206, 233, etc.) and to Mercury at Uley (cf. 355, 359, etc.). In addition to these sanctuaries (which were heavily frequented until the beginning of the fifth century), we also find scattered finds across the island (cf. 343–44, 205, etc.) and Gallia (cf. 159 and 204). And of course we must add the fountain of Anna Perenna in Rome (cf. 19– 47), which was most actively used for magico-religious activities during the fourth and fifth centuries CE.

In this first category, sanctuaries and temples predominate, given that they were privileged spaces for communicating 45

Sylloge of  Defixiones from the Roman West sanctuary at Salacia/Alcácer do Sal (cf. 121), where one defixio was found inside a small tank in the smaller of the temple’s two cellae, was dedicated to a deity whose identity is today unknown.

with divinities. Accordingly, these two contexts count for the vast majority of tablets found in sacred spaces. As alluded to, in the Roman West, the sanctuaries of Sulis Minerva and Mercury have yielded the largest number of tablets, with over 200 defixiones from these two places alone (i.e., almost two-thirds of the total). Among the other sites where defixiones have been discovered, we should highlight the importance of the sanctuary dedicated to Isis and Magna Mater in Mogontiacum/Mainz which has given us 34 tablets dating to the first and second centuries CE. Within this same chronological span, there are another two texts coming from the isea of Baelo Claudia/Bolonia and Savaria/Szombathely (cf. 128 and 531). The texts from Centum Prata/Kempraten (cf. 518f.) have been dated between the second and the third centuries and were found in a sanctuary dedicated to Magna Mater; this is also the case for the ones from Abusina/Eining (cf. 525 and Blänsdorf 2019). Of later date is the curse tablet discovered at Lydney (cf. 205), a site that was dedicated to the Roman-Celtic deity Nodens. The remaining sanctuaries from the Roman West, which were dedicated to deities whose identity we can no longer determine are found in Hispania (cf. 121), Gallia (cf. 162) and Britannia (cf. 447, 455 and 458) and have only yielded few isolated items.

Pits, which were structures similar to the Roman favisae, are also generally associated with sanctuaries. In this contest, votive offerings were deposited.308 So far, in the Roman West, we know of at least two defixiones thrown into these pits, which have been found in Gallia (cf. 171) and Britannia (cf. 341). The Gallic pit was discovered in the Cité Judiciaire of Le Mans during an emergency excavation about which we have hardly any information (cf. RIG II.2, p. 296). The pit, dated to the first century CE, contained a reused lead tablet (cf. 171), which, because of the material and context, seems to have been a defixio. Turning to Britannia, we know of at least one pit that held a defixio: excavated in 1988 in Southwark (London), it held a single curse that only bore its victim’s name (cf. 341). The tablet was found alongside other artefacts that have been dated to the fourth century CE: a coin from the reign of Constantine II (cf. Hassall 1992: 310) and some household rubbish, as well as ‘the remains of a partially articulated dog’ (cf. Seeley and Wardle 2009: 153), an animal that can possibly be connected to the underworld (cf. Green 1992: 111–13).309

In these contexts, we can observe different trends and patterns in the ways that curses were deposited (although it is not always clear the reasons why these particular locations were selected). The resulting diversity of practice is seen in the particular places chosen, which include the area around the temple (cf. 162, 447, etc.), within the tenemos (cf. 128 and 455), in the portico (cf. 76) as well as near or even behind the podium (cf. 531). In the case of Mogontiacum/Mainz, 24 curse tablets were found in the rear part of the temple and placed inside of a rectangular well above which two altars were later built (Witteyer 2005: 116 and 2013; DTM, pp. 1–6, 39f.). On these altars offerings, among which we must count several curses, were thrown directly into the sacred fire.

Within this same category of built-up sacred spaces there are also fountains, from which we have some 20 texts. Whether clearly marked as sacred or not, fontes possessed a strong religious character from the Greek Archaic period onwards, because of the power associated with the flowing water and the belief that underground water provided a way to communicate with the underworld.310 In the Roman West, five such fountains were used as a place for depositing curses: Arverni/Chamalières, Amélie-les-Bains and Aquae Tarbelliacae/Dax (all in Gallia) as well as at Arretium/Arezzo and Rome (in Italia). Among these different archaeological sites, we ought to highlight the hot water spring at Arverni/Chamalières, known as ‘Source de Roches’, where 3,000 votive offerings311 were found inside of the principal basin

In other instances, tablets have been uncovered within sanctuaries, as is the case at Aquae Sulis/Bath and Uley. Both of these complexes fit the mould of a Roman-Celtic sanctuary, where the sacred space is marked by a cella and is surrounded by an ambulatory upon which worshipers would walk. At Aquae Sulis/Bath, curses were thrown directly into the sacred spring, which was located in the centre of the complex (cf. 206–335). At Uley, on the other hand, there was a small pool placed in the centre of the cella, which likely served as the place to deposit curses (cf. 354–440). Although this cella seems to have served as the original deposit for these curses, where they were actually found spread throughout all the archaeological strata of the sacred complex (mostly among the secondary votive deposits).307 Finally, we should add to the list the

On this, see Green 1997: 170–71, s.v. ‘pit’. To this British pit, we should perhaps add that from Old Harlow, although it is not clear whether it ought to be classified as a pit or a well. The excavation (carried out in 1970) determined that the structure was 3 m deep and had a curse (cf. 353) as well as ceramic sherds dated to the third and fourth centuries CE (see Conlon 1973: 34 and 40). Furthermore, we should note that the structure where the tablet was found was located near two others that contained animal bones, shells, pieces of glass and pottery (dated to the second and third centuries CE), as well as fibulae and bronze votive offerings. Furthermore, near the top of one of these structures, archaeologists found 90 coins dating between the third and fourth centuries CE. All of this suggests that this was a sacred context. 310  See Romizzi 2005: 242, s.v. ‘Fons (mondo romano)’, in ThesCRA IV, 242–44. 311  Romeuf 2007: 88, where it is specified that 1,500 complete votives along with some 8,500 additional fragments were found (best estimates suggest that there was a total of some 3,000 items). The majority of these were anthropo- and zoomorphic sculptures and wooden plaques (which were inscribed or painted), as well as mallets, coins, fibulae, pottery and other offerings. For a further discussion, see Romeuf 2000 and 2007: 88–93. 308  309 

307  Of the 140 curses, 62 were found within secondary votive deposits dated to phase 5; we do not have precise details about the strata of 47 tablets; the remaining 31 were found among strata 4, 6, 6b, 7a and 7b. For a discussion, see Woodward 1993: 113 and Woodward and Leach 1993: 330–31, fig. 225.

46

Deposition Contexts (Romeuf 2007: 87). Among these votives and deposited in what appears to be a haphazard manner, a Gaulish defixio was discovered that invokes Maponos, the god to whom the sanctuary may have been dedicated (cf. 163).312 Less is known about the six defixiones from Amélie-lesBains (cf. 150–55), which were discovered in 1845 in the main fountain (‘Le Gros Escaldador’) of a bath complex: they were uncovered along with coins dated to the first century CE, which attest to the site’s religious function.313 The final example from Gallia is the so-called ‘Fontaine de la Nèhe’ or ‘Fontaine Chaude’ in Aquae Tarbelliacae/ Dax, which was of Gallo-Roman origins and was built up during the reign of the Severans (Watier and Gauthier 1977: 323). An excavation in 1976 brought to light a curse tablet written after a theft (cf. 164) together with ceramic sherds and a group of coins that had been deposited between the end of the fourth century and beginning of the fifth century CE.314

le puits’ (1975: 22; there were five dogs, one cat or lynx, a deer and an ox).317 Among these remains, archaeologists uncovered a defixio (cf. 157) dating to the mid-first century CE, which was deposited in the well when it was still in use. We should note that this text includes an intriguingly persuasive analogy that compares the falling of the lead tablet to the bottom of the well to the downfall of its victims (cf. 157, A, ll. 1–4). The well at Rauranum/ Rom was discovered and excavated in 1887 in a Roman villa. Measuring 1.8 m in diameter and 20 m in depth, this well contained surplus building materials, animal bones and bronze coins. At a depth of 10 to 12 m, 40 lead tablets were unearthed. Some of these were rolled up, others were pierced with a nail. None of these seems to have been inscribed, though it is believed that they were nevertheless curses, like other uninscribed lead tablets deposited in sanctuaries (e.g., at Uley; cf. Jullian 1897a: 132, n. 1). Below this, archaeologists found a collection of tablets, a sickle, a pick and a curse tablet (cf. 159; note that this is the only published curse from this site), which has been dated to either the third or fourth century CE and has proven quite difficult to decipher. Though we are short on reliable details concerning this well and its contents, the discovery of this important collection of tablets nevertheless provides evidence that the well served as a site for communication with divine powers.

In Italy two fountains have also been discovered. Little is known about the first, which is in Arretium/Arezzo and surveyed in 1869. In this fountain, archaeologists found a defixio together with a group of coins dating to the reign of Antoninus Pius; it has been speculated that the fountain was dedicated to the nymphs who are invoked in the curse (cf. 92). The fountain of Anna Perenna, which was dedicated to the goddess and her nymphs and which was discovered in 1999, deserves special attention.315 The fountain was supplied by a krene (i.e., a spring) in which there were three altars dedicated to this deity (cf. Friggeri 2002: 26–33). Its cistern contained a rich votive deposit made up of more than 500 coins, some 70 lamps, a caccabus, organic remains and, most important for our purposes, 32 defixiones (cf. 19–47).

Finally, we can turn to natural sacred spaces, in which five tablets have been uncovered. Four of the five were found in Britain (cf. 340, 441, 449 and 451), while the fifth hails from Pannonia Superior (cf. 529). Among the British examples we must note that the curses from Brandon and Hamble, which date to the fourth century CE, were found by metal detectorists (cf. 449 and 451). We do not know the date nor the exact circumstances in which a curse from London was found along the banks of the River Thames (cf. 340), where the curse from Venta Icenorum/Caistor St Edmund (cf. 441; of unknown date) was discovered on the riverbank of the River Tas. Only the curse from Siscia/ Sisak (cf. 529), which is dated to either the first or second century CE, was found during an actual archaeological excavation, in this case those carried out on the shores of the River Kupa in 1913.

Ritual wells are also a type of sacred space that was used for depositing defixiones. Two such cases have been recorded in Gaul (cf. 157 and 159). Generally, it is thought that these wells, which are found in Celtic areas, probably served as Roman mundi (gates to the underworld) or as openings to the inner parts of the earth. These had a chthonic quality and were in use before the Roman conquest of the regions in question.316 The first Gallic well was discovered in 1975 in the oppidum Montfo, which was inhabited from the sixth century BCE until the first century CE. This well, which is some 13.5 m deep, contained wooden objects (tablets, combs, pegs), ceramics (including nearly complete jugs) and animal remains, which, according to Bacou and Bacou, ‘sont tombés ou ont été envoyés dans

Though our knowledge about the discovery and context of these tablets is rather hazy, it seems likely that they were originally placed in a flowing river and then were later deposited in a river bank, mud or beach. Such a narrative would certainly fit with the deities invoked in these texts, who are all associated with water, such as Neptune,318

According to Lambert apud RIG II.2, p. 274, ‘Maponos est ici invoqué en tant que divinité de la source de Chamalières’. But see the objections proffered by Romeuf 2007: 94–95. 313  On the archaeological context, see the brief notice from Puiggari apud Henry 1847: 410. 314  See Nony and Tobie 1977: 324–25, who maintain that, even if the hoard of coins dates between 364 and 375 CE, ‘ces émissions sont pratiquement les seules à circuleur au Ve siècle’, a date which would match that of the curse tablet (contra Alfayé 2019: 264). 315  We are still awaiting the full publication of the fountain. For the present time, see Piranomonte 2002, 2005, 2009, 2010a, 2012a, 2013 and 2015. 316  See Green 1997: 224, s.v. ‘well.’ 312 

For the well’s contents, see Bacou and Bacou 1975: 17–22. Given the presence of dogs deposited in wells, Green has argued that these animals were connected to the underworld (1992: 111–13, esp. 112). 318  Neptune is evoked in all the British curses, though in 340 the theonym appears in the Vulgar form Metunus and in the defixiones from Brandon and Hamble (449 and 451) he is alluded to as dominus. In the case from Brandon, the phrase cor˹u˺lo pare(n)tator (l. 7) is used to offer the victim to Neptune and could be linked to the punishment mentioned by Tacitus (Germ. XII, 1) for ignauos et inbellis et corpore infamis caeno ac palude, iniecta insuper crate, mergunt. On this, see Marco Simón 2020. 317 

47

Sylloge of  Defixiones from the Roman West Niskus319 and Savus.320 As is well known, in ‘Celtic’ religion, water was thought to be one of the privileged sites of communication with the divine.321 Indeed, water was viewed as a sacred liminal space that not only was the fount of life but also provided an entrance to the world beyond. Accordingly, such aquatic spaces often served as the site for ritual practice. Among these cultic activities, the deposition of metal is one of the best known, thanks to both ancient textual testimony and also to the archaeological record322 that attests to a flexible practice which evolved to include the deposition of defixiones during and subsequent to the period of ‘Romanization.’

In this instance, the walls of Germanicus’ house served as the place where the curses were deposited, together with other human remains. We are left to wonder if this final detail could be an embellishment of Tacitus’ part, who was ‘imagining’ the magical praxis rather than recounting what actually happened (though note that Dio Cassius tells the same story).324 We learn of a similar event from Libanius, who recounts suffering from a protracted illness that no physician could cure, until ‘a chameleon, of uncertain origin, was discovered in the classroom. It had been there a long time and had been dead for many months. We saw that its head had been placed between its hind feet. Of its forefeet, one was nowhere to be seen, and the other was closing its mouth to keep it silent’.325 Once the animal was discovered, Libanius managed to recover from the mysterious illness.

6.5. Spaces associated with the victim As the name implies, this depositional context was chosen due to its proximity to the target. This purpose could be accomplished by hiding tablets in the victim’s house or place of work (as happened in cases when the victim was a gladiator or a charioteer and the tablets were placed in the amphitheatre or the circus). It is hard to discern the main motive behind this type of deposit, since not only do we have a lack of literary references, but also (and more generally) because of the lack of good information about the tablets’ archaeological contexts.

Likewise, thresholds, which are the liminal spaces par excellence, were also an ideal place to deposit a curse, as Saint Jerome describes. In his Life of Saint Hilarius the Hermit, the author tells the story of a deranged young virgin against whom a man had written ‘some verbal monstrosities and monstrous forms on plates of copper and buried them under the threshold of the girl’s house’.326 Furthermore, we can also add the testimony of of Sophronius who mentions in his story about the doctor Theodorus of Cyprus who was ill and had a revelatory dream that uncovered the cause of this sickness and the cure to his suffering:

Classical literature, however, does provide us with some relevant information about aggressive magical practices that relied on proximity to the victim. One of the bestknown sources comes from Tacitus’ description of the death of Germanicus. As the historian puts it:

‘send one of your servants to Lapithos and tell him to dig in front of your bedroom next to the doorway. There he will find the wicked instrument of the sorcerer. Once it is uncovered, its maker will disappear immediately.’ Theodoros sent (him) to the place, as commanded, and he found the cause of the disability.327

Germanicus’ conviction that he had been poisoned/put under a spell [veneni] by Piso aggravated the disease. They dug up the floor and the walls and found remains of human bodies in them, spells and binding curses [devotiones], and the name of Germanicus inscribed on lead tablets, ashes half-burned and smeared with gore and the other evil devices by which it is believed that souls are devoted to the infernal powers.323

Apparently, this was some sort of magical artefact (the author does not give further details) whose powers were lost upon being discovered.

Dio Cassius, LVII, 18, 9 (translated by E. Cary, 1968): ‘His death occurred at Antioch as the result of a plot formed by Piso and Plancina. For bones of men that had been buried in the house where he dwelt and sheets of lead containing curses together with his name were found while he was yet alive’ (ἀπέθανε δὲ ἐν ᾿Αντιοχείᾳ, ὑπὸ τε τοῦ Πίσωνος καί ὑπό τῆς Πλάγκίνης ἐπιβουλευθείς· ὀστᾶ τε γὰρ ἀνθρώπων ἐν τῇ οἰκίᾳ ἐν ᾗ ᾤκει κατορωρυγμένα καὶ ἐλασμοὶ μολίβδινοι ἀράς τινας μετὰ τοῦ ὀνόματός αὐτοῦ ἔχοντες ζῶντος ἔθ᾿ εὑρέθη). 325  Libanius, Orat. I, 249 (translated by Ogden 2002: 260). On this, see Bonner 1932b; Cracco Ruggini 1996 (esp. p. 162 ff., for an analysis of the metaphorical value of the chameleon which represents the orator’s own political ambiguity); Maltomini 2004 (for a different analysis that focuses on three magical recipes in which chameleons are used to silence an enemy). 326  Jerome, vit. Hilar. XI (translated by Ogden 2002: 230 and slightly modified): …et subter limen domus puellae portenta quaedam verborum et portentosas figuras sculptas in aeris Cyprii lamina defodit. On the media for writing the curse, see section I.3.2, n. 58. 327  Sophronius, Narratio, LV (Migne, P.G., LXXXVII, 3, col. 3625, apud Audollent 1904: CXXII. English translation by Gager 1992: no. 166): τινα τῶν σοι διακονουμένων εἰς Λάπιθον, καὶ πρὸ τοῦ κοιτῶνος τοῦ σοῦ παρὰ τὸ πρόθυρον σκάψαι διάταξον · ἐκεῖσε γὰρ εὐρήσει τὸ τοῦ δεδρακότος κακούργημα, ὑπὸ τὸ πρόθυρον ἀφανῶς καλυπτόμενον · ὅ καὶ φανερούμενον, εὐθὺς ἀναιρεῖ τὸν ποιήσαντα. Πέμψας δέ Θεόδωρος ὡς κεκέλευστο κατὰ τὸν λεχθέντα τόπον, τὸ τῆς ἀρρωστίας εὕρισκεν αἴτιον... 324 

319  Niskus is invoked together with domino Neptuno in the curse from Hamble (cf. 451). Determining the identity of this Niskus has been somewhat difficult given the absence of other parallels. Scholars have associated him as the masculine counterpart of Niskas from 150–51 and 154, or as a ‘Celticized’ version of Neptunus (see Tomlin 1999a: 562, n. 52). 320  The deity of the river in which the curse from Sisak was deposited and upon whom the defigens calls to drown the curse’s victims (cf. 529 B, ll. 1–6). 321  For the symbolic value of water in Celtic religion, see Cunliffe 1988: 359–62; Webster 1995: 449–50; Green 1997: 223–24 (s. v. ‘water’). 322  When it comes to classical sources, one of the most well-known passages is from Strabo (IV, 13), where the geographer mentions the lacustrine deposit of precious objects made by the Volcae Tectosages in the vicinity of modern-day Toulouse. Archaeologically, we should mention the discoveries at La Tène and Llyn Cerrig Bach (see Webster 1995: 450, with further bibliography). 323  Tacitus, Ann. II, 69 (translated by Ogden 2002: 217): ‘…et reperiebantur solo ac parietibus erutae humanorum corporum reliquiae, carmina et devotiones et nomen Germanici plumbeis tabulis insculptum, semusti cineres ac tabo obliti aliaque malefica quis creditur animas numinibus infernis sacrari’. On this episode, see D’Erce 1969; Tupet 1980; Dickie 2010.

48

Deposition Contexts Table 6.4. Defixiones discovered in contexts closely associated with the victim, organized by provenance (It.= Italia, Afr.= Africa, Hisp.= Hispania, Gall.= Gallia, Brit.= Britannia, Ger.= Germania, Raet.= Raetia, Nor.= Noricum, Pan.= Pannonia). Reference is also made to the archaeological context, date and the total number of curses found in each province. Prov.

Context

SD/Site

Date

Total

It.

17 113 (?)

Domestic (‘house of Livia’) Domestic (monumental area)

1st CE 3rd CE

2/119

Afr.

DT 247–251 DT 253–254 Jordan 1988: no. 3

Amphitheatre Amphitheatre Circus

1st CE 1st CE 3rd CE

8/88

Hisp.

127 139 145 148

Domestic (on a floor) Domestic Domestic (entrance of the settlement) Domestic (wall, chryptoporticus)

2nd CE 1st BCE–1st CE 1st BCE–1st CE –

4/30

Gall.

172 175–202

Domestic (ditch) Amphitheatre

1st CE 4th–5th CE

29/55

Brit.

336 337 343 344 346 456–457 446

Domestic (Roman fort) Amphitheatre Amphitheatre Amphitheatre Domestic (oven) Domestic (patio of a mansio) Domestic (on a floor)

– 1st–2nd CE 2nd–3rd CE – 3rd–4th CE 2nd–3rd CE 4th CE

8/256

Ger.

463 482

Domestic (Roman fort) Domestic (ditch)

1st–2nd CE 1st–2nd CE

2/59

Raet.

522 524

Domestic (wall) Domestic (foundations)

1st CE 2nd–3rd CE

2/6

Nor.

–

–

–

0/1

Pann.

528 530

Domestic (entrance of a dwelling) Amphitheatre

1st–2nd CE 2nd–3rd CE

2/9

6.5.1. Geographical and temporal data about deposits closely associated with the victim

Finally, we should turn to the PGM. It is noteworthy that only a single recipe calls for the deposition of a curse tablet in a space closely associated with its victim.328 More specifically, we find this instance in a spell desiring the separation of lovers (PGM XII, 365–75), where the practitioner is asked to deposit the spell ‘where they [i.e., your victims] are, where they usually return’.329 So far, this is the only known case about curse tablets from the papyri. Nevertheless, such a strategy is widely recommended in a distinct context: the PGM tells its readers to place amulets within the walls of houses or businesses that they want to see thrive.330

Despite these dramatic stories, we are hard pressed to find good archaeological cases of this same practice, as Table 6.4 makes clear.331 As the above data demonstrates, the use of curses in this context is almost anecdotal in the Roman West, accounting for less than 10 per cent of the total. As explained in previous pages, the vast majority of defixiones were deposited in funerary contexts (especially in Africa, Italia and Hispania; see section I.6.2 and Table 6.1) or sacred contexts (especially Britannia and Germania; see section I.6.4.1 and Table 6.3). Chronologically speaking, this kind of context first appears around the end of the first century BCE, around a fifth of the attested cases date from the

Wilburn 2013: 45, n. 130 also mentions several recipes from PDM xii. Nevertheless, their contents are too fragmentary to be discussed here. 329  Translation by R.F. Hock (apud Betz 19922). 330  For phylakteria deposited in the house or workplace of the interested party, see the following two examples: PGM IV, 2360–440 (meant to make a house or business thrive); PGM VIII, 53–63 (a request to Hermes to bring success). For amulets, see PGM VII, 390–94 (meant to bring victory in a horse race and hence written on a horse’s hoof) and PGM VII 919–24 (to secure victory, it should be carried in one’s sandals), among others. 328 

To this evidence, we should add an extraordinary curse recently published by A. Varone (2019), which was inscribed on the walls of room 28 of Villa Arianna (Stabiae/Castellammare di Stabia, Regio I). The room was part of the villa’s termae.

331 

49

Sylloge of  Defixiones from the Roman West High Empire and the majority date from Late Antiquity, the period to which more than half of the known cases date. Within this context, we can distinguish two main subcategories: domestic and agonistic spaces (e.g., the circus or amphitheatre).

originally been placed inside a wall (Egger 1957: 72). The tablet from Peiting (Raetia, cf. 524), was found among the foundations of a Roman villa. Generally, scholars have interpreted these spaces as those in which victims either lived or worked. Though this seems likely in certain cases (when the target is a charioteer or a gladiator and the curse was deposited in the circus or the amphitheatre), at times it is impossible to confirm whether the victims lived or worked where a curse were deposited, simply because the texts do not provide any specifics. Just to give one example, the curse tablet from Barchín del Hoyo (Hispania, cf. 145), dated between the first century BCE and the first centuries CE, was deposited in the Iberian settlement Fuente de la Mota, which was abandoned around 210 BCE. Although there is some controversy about the exact findspot, it seems that the tablet was deliberately deposited in the south-eastern entrance of the settlement, among ruins, since ‘these were considered a place in contact with the nether world’ (Curbera et al. 1999: 280).332 For the editors, it could also be explained by the fact that the victims of the text were workers on one of the active mines located near the Iberian settlement. Such a hypothesis is simply impossible to prove. In addition to the Hispanic text, a Gaulish text from Autricum/Chartres (Gallia Lugdunensis, cf. 172) also deserves special mention: dated to the first century CE, it contains three lists of personal names. Given that the curse tablet was discovered in a ditch running along a path, near a ceramic workshop, the excavators have supposed that the curse had intentionally been deposited in a space close to their victims, who could have worked at the workshop. Again, such a hypothesis is purely speculatory.

6.5.2. Domestic spaces In this first group, we can include the tablets found in military and civilian settlements. Among the former, we can count the defixio from Chesterton-on-Fosse (Britannia, cf. 336), which was discovered during the excavations of a Roman fort (though the exact context is unknown). Another similar case is that from Bodegraven (Germania Inferior, cf. 463), which was probably discovered within a Roman fort, ‘between the barracks of the praetentura sinistra’ (L. Swinkels, per litt.). As far as civilian settlements are concerned, we can point towards the defixio from Neapolis/ Terralba (cf. 113), which was written on an ostrakon and can be dated to the third century CE: this curse was found in the monumental area in the southern portion of the city. In this kind of settlement, the archaeological record attests to a diversity in the actual means of depositing curses in this context: sometimes tablets were simply placed in the vicinity of a house, as it happens with 17 (Rome), which was found near the ‘house of Livia’, in an exact context that is hard to define, seeing that, as Panciera (2006: 146) explains, ‘il terreno appariva profondamente sconvolto ed il materiale trovato con la tabella è risultato estremamente eterogeneo’. Tablets could also be placed in a nearby ditch, as is the case with 482 (Groß Gerau, Germania Superior), where the tablet was found inside the ditch of a dwelling together with stucco fragments, a tegula and a coin dating to the reign of Vespasian.

When the victim was indeed living in a domestic space, it is unknown whether—as Tacitus and Dio Cassius claim was the case for Germanicus—these curses were accompanied by the remains of human corpses or other items, which were thought to increase a spell’s potency (Ogden 1999: 19). If such practices did indeed take place, we would again find ourselves in the situation where apparently separate contexts—in this case the domestic and the funerary—overlapped and were blurred. Such a mixed or polysemic context would certainly be thought to have wielded extraordinary power.

On certain occasions, curse tablets were deposited inside of the house itself. Thus, some curses were hidden in the courtyards of large homes. This is the case with 456–57, which were found in the patio of a mansio from the northeastern part of Ratae Corieltauvorum/Leicester. The curse from Canonium/Kelvedon (Britannia, cf. 346) was found in an oven where it had been ‘folded up in a third- and fourth-century stratum’ (Wright 1958: 150). This context, which is unparalleled, does not seem to be linked to the kind of curse (it seems to be a defixio in fures) nor to the deities invoked (Mercury and Virtus). I am left to wonder if this context could somehow be explained in terms of sympathetic magic. In parallel to the stories found in Saint Jerome and Sophronius, curses have been discovered in the entrances of the houses, such as the tablet from Emona/Ljubljana (Pannonia, cf. 528), that was found in the entrance of a dwelling (between rooms 52 and 53 to be precise), a liminal space through which the curse’s victims surely passed. Finally, we know of three curse tablets that were deposited in the walls of houses. One is from Cabrera de Mar (Hispania Citerior, cf. 148) and was found among the material used to fill the walls of a cryptoporticus (see Bonamusa et al., 2000: 169). Upon its discovery, the curse from Cambodunum/Kempten (Raetia, cf. 522) contained traces of grout, which led the editor to suppose that it had

6.5.3. The agonistic sphere The second subcontext that comprises places closely associated with the victim is the agonistic sphere, which is well represented by a considerable group of defixiones unearthed in circuses and, above all, amphitheatres.333 A good example from the circus comes from Carthage, which was excavated between 1982–90 by a team from In addition, we could also suggest the deliberate placement of the tablet in this landscape if the practitioner sought to establish a persuasive analogy between the victims and the depositional context: thus, ruins would bring ruin upon the targets of the curse (see Alfayé 2019). 333  See Sánchez Natalías (2020c: 67–68 and 70–73). 332 

50

Deposition Contexts the University of Georgia who found 13 curse tablets.334 We must stress that the fact that these curses were found within the circus itself makes this cache into a unique find among the defixiones from the Roman West. To date, all other circus tablets have been found in tombs.335 Accordingly, we are left to ask for what reason these 13 defixiones were deposited in this space. The tablets were discovered opposite the carceres and near the spina,336 spaces of a liminal nature where two of the crucial moments of the race occurred (Heintz 1998: 340f.). Thus, the carceres were the place where the daemones invoked were expected to attack first, frightening the horses and —at least— hindering the charioteers’ start.337 In addition, the risky turn around the metae of the spina constituted the perfect spot for teams to fall and get injured, just as the texts request. Some ask for an accident ‘so that tomorrow morning they are not able to run in the hippodrome (…) or circle around the turning post; but may they fall with their drivers...’.338 That said, of the 13 curses found in the Carthaginian circus, only 3 have been published (cf. Jordan 1988). While two of them seem to deal directly with chariot races, the other makes no sense.339

ancient grave (Homer, Il. XXIII, 331), while in the Roman world the cone shape of the metae also evoked a funerary monument, since this shape was a recurrent motif on tombs (Humphrey 1988: 256). Nevertheless, and as mentioned above, so far the use of the circus as a deposition context for curse tablets should be seen as an exception to the rule: most practitioners from the Roman West preferred the use of necropoleis and graves for depositing curses related to games and spectacles. Within this agonistic sphere, we must also examine amphitheatres which provide a hefty number of curses, some of which were directed towards gladiators. It must be noted that the reasons for using the amphitheatre as a place for the deposition of curse tablets are not univocal, given that this enclosure could be conceived of as both a religious space and as one connected to the restless dead and underworld. In fact, it all came down to how the practitioner understood this space, which could serve his/ her various needs in different ways. As shown in the previous table, we know of curses from the amphitheatres of Carthage, Isca Silurum/Caerleon, London, Augusta Treverorum/Trier and Carnuntum/ Petronell. From the first of these, we have 55 curses341 that were discovered by Delattre in 1895 ‘dans l’arène un souterrain aboutissant à une sorte de cul-de-sac carré, ouvert à la parte supérieure, au niveau même de l’arène’ (1897: 318). According to Audollent’s analysis, the presence of so many curses in this space means that it was used for funerary purposes (1901: 303), which has led to it being identified as the spoliarium. This area, where the corpses of gladiators and venatores were stored, had obvious links to the aoroi and biaiothanatoi (Audollent 1904: CX). Of the 55 curses found in this space, the 9 published defixiones target gladiators and venatores who participated in the games celebrated in the same amphitheatre. Accordingly, this context can be seen as having a double valence, mixing the funerary with closeness to the victims.

In any case, the decision to deposit these curses in the circus should not be seen as random: we must remember that it was a building contaminated with miasma, given that numerous accidents (at times fatal) took place there.340 Consequently, these areas were inhabited by aoroi and biaiothanatoi, whose ability to activate defixiones was generally assumed. Accordingly, we can glimpse the attractiveness of this location for depositing curse tablets, above all when the curses were targeting chariot races. In addition to the contamination of miasma, this building was related to the infernal world from its very origin. Thus, Homer already described a race where the cut timber used for the turning posts was thought to be the marker of an 334  On the excavations, see Pintozzi and Norman 1992; for the lead of the tablets and their stratigraphy, see Pintozzi 1991. 335  For example, the ‘Sethian’ defixiones, which were discovered in the columbarium on the Via Appia (cf. Wünsch 1898), or the curse found in the columbarium of the Villa Doria Pamphili (Rome; see Bevilacqua 1998) as well as another from Astigi/Écija (cf. 132). From Hadrumetum there is an important series of circus curses that were found inside of tombs (see DT 265 and 272–95; Grenier 1905: nos 1–2; Audollent 1906: nos 1–2; perhaps Audollent 1910: no. 2, though the exact provenance is unknown). 336  According to Pintozzi (1991: 51–61), of these 13 curses, 1 was found next to the podium ‘but it may have been disturbed from an original location at the arena’s edge’ (1991: 54; US 82-3-948= Jordan 1988, no. 3), 2 in the arena (tablets US 83.1B and US 81-13-2= Jordan 1988, nos 1 and 2) and 4 near the spina (curses US 90.3-206, US 90.3-588, US 90.3260 and US 90.3-279). The exact provenance of the other six defixiones is unknown. 337  Cf. DT 187 (l. 59) and DT 234 (ll. 21–22)= Heintz 1998: 340, no. 22 and 23, on the ancient belief that horses were able to perceive demonic presences. See also Gordon 2012: 48, no. 111 and Faraone 2019: 183. 338  As put in DT 237, l. 35–42 (English translation by Gager 1992, no. 9). There are other examples, such as DT 237 (l. 48–73) and DT 239 (l. 42–48). 339  Although, as the editor has pointed out, one of the tablets ‘does not yield any continuous sense, but its purpose is no doubt related to its findspot’ (Jordan 1988: 119). 340  Regarding the death of charioteer Scorpus, see Martial Epigrammata, X.50 and 53.

Concerning the amphitheatres from Britannia, so far we know of three curse tablets from Isca Silurum/Caerleon (cf. 337)342 and London (cf. 343–44).343 The first was found in the arena, while those from London were discovered among the material used to fill a drain in the arena. Although one of the texts (cf. 344) just contains a list of names, the other two can be confidently identified as curses against thieves. These curses, whose victims’ identity was unknown, evoked Diana and Nemesis. This led us to think that, at least in the case of these two texts, it seems as if the practitioners sought out a context connected Of the 55, 7 are in Latin and only 9 have been published (DT 246–54). For the amphitheatre in Carthage, see Golvin 1988: no. 95 and Rossiter 2016. 342  Found in the amphitheatre’s arena, the curse contains the following invocation: Dom(i)na Nemesis do tibi palleum et galliculas... 343  It reads: [d]eae Dea[na]e dono c ̣apitûlarem... This curse was found in the drainpipe of the arena in London along with two other tablets that were quite poorly preserved. The first (344) simply contains a list of names, while the second, which is illegible, was still folded up (see Tomlin 2003: 362–64). 341 

51

Sylloge of  Defixiones from the Roman West to the deities rather than to their victims (even though they could possibly be aficionados of the games). Following this hypothesis, the practitioners who wrote these tablets understood the amphitheatre as a religious space rather than a space connected with the restless dead and the underworld. In this regard, we ought to remember that within amphitheatres there were cultic spaces dedicated to divinities associated with the games, such as Diana and Nemesis, who were invoked in the British texts. 344

Evidently, given that none of the defixiones are aimed against gladiators or venatores, proximity with the victim was almost certainly not the key issue, as was the case, for instance, with the tablet from the amphitheatre of Carthage. Although the funerary connotations of the amphitheatre could be one key for understanding the deposition of the tablets from Augusta Treverorum/Trier (Wünsch 1910: 3), we must remember that this location was also a religious site. This, in my opinion, is the more compelling explanation (see Grenier 1958: 709, n. 2), particularly if we bear in mind that two of the defixiones invoke Mars and Diana (the latter being a deity who was worshipped in that precise location, as shown by an inscription discovered in the amphitheatre, cf. CIL XIII, 11311, and 182–83), 349 while a third summons a female divinity with the formula bona san(c)ta nomen p̂ iạ (cf. 189, l. 1).350 These texts from Augusta Treverorum/Trier further underline the difficulty and problematic nature of trying to neatly fit texts into one context over another. Nevertheless, our lack of information about the findspot as well as the fragmentary nature of these curses (which today are largely lost and therefore cannot be examined)351 ultimately prevent us from offering a satisfactory solution to the question of the tablets from the amphitheatre in Augusta Treverorum/Trier.

Something similar is at play with at least 3 of the 28 Latin curse tablets discovered in the basement of the amphitheatre at Augusta Treverorum/Trier.345 Unfortunately, little is known about the original deposit of this ensemble from Trier (cf. 175–202), which was found in 1908 in the socalled ‘östlichen Kammer’.346 While some of the defixiones were found 1 m below the surface, others were found on the ground level, while still other tablets were discovered 1.5 m deep. Wünsch (1910: 2f.), who was surprised by the different strata of these finds and the way that the tablets were spread out, offered forward two hypotheses. The first one suggested that the defixiones could have originally been buried in the arena and later fallen down to the basement when it collapsed sometime around the fifth century CE. His second hypothesis proposed that the curses could have been nailed to the wooden posts of the basement where they were eventually buried by the debris after the entire structure collapsed. Yet, only three curses from this collection seem to have holes from nails. Therefore, the first hypothesis—which defended the burial of the defixiones in the arena—seems more plausible. In any case, the lack of data regarding the context of the tablets prevents us from tracing the exact location of the original deposit, rendering any hypothesis far from certain, as Wünsch himself admitted (1910: 2). In addition, the contents of the curses do not further our understanding of why they were deposited in the amphitheatre: 10 of the 30 curses remain unedited (being illegible or not inscribed), while 13 were written for unknown reasons.347 Only three of the tablets could possibly belong—with reservations, given the fragmented condition of the texts—to the group of judicial defixiones.348 Therefore, we are left to wonder why these practitioners chose to deposit their curse tablets in this context.

We find a similar case with the curse from Carnuntum/ Petronell (Pannonia, cf. 530), where during the archaeological excavations carried out between 1923 and 1925 at the amphitheatre in the civilian settlement a defixio was discovered. Little is known about the curse’s archaeological context, which was found on the ground next to the enclosure’s southern door and where a tank was built in Late Antiquity (cf. Egger 1926: col. 137). The curse targets a certain Eudemus and should be classified as a defixio in fures. The tablet evokes infernal deities, such as Dis Pater, Veracura and Cerberus.352 As was the case with the tablets from Augusta Treverorum/Trier, it is difficult to determine why this curse was deposited in this particular context: the practitioner may have been attracted to the chthonic associations of the place or perhaps (s)he chose the area near the gate through which the victim of the curse may very well have passed, thus coming into close contact with the defixio.

344  According to Marcattilli (2005: 185), ‘i sacelli erano allestiti in uno dei carceres dell’anfiteatro, ad una delle estremità dell’asse minore dell’edificio, per lo più al di sotto del pulpitum editoris, comunque in stretto rapporto con il percorso delle pompai’ (ThesCRA IV, 2005, s. v. ‘Arena’: 184–86). For these cultic spaces with amphitheatres, see Golvin 1988: 337–40. In addition, the possible existence of a nemeseum in the amphitheatre in Isca Silurum/Caerleon would reinforce the religious aspect of the depositional context (cf. Golvin 1988: 129, n. 344). 345  Where, according to Golvin (1988: 338), ‘une chapelle devait aussi exister à l’une des extrémités du petit axe de l’arène’, from where the base of a statue dedicated to Diana (found in 1909) may have come (cf. CIL XIII, 11311). 346  Wünsch 1910: 2. For the amphitheatre of Trier, see Golvin 1988: no. 41 and Kuhnen 2017. 347  194, 197 and 201 are not edited; 193 and 202 are illegible, while 198–200 are anepigraphic. 175, 177, 179–81, 184–85 and 187–92 were written for unknown reasons. 348  176 and 182 could be judicial defixiones, since they contain terms such as adṿọ[ca]tụṣ (176, ll. 2–3) and inimicum (no 182, l. 1).

349  Specifically, 182 l. 7 reads: Marti et Diane, while 183, ll. 5–7 reads: [Di]anam et Martem vinculares ut me vindicetis de Ququma... The presence of Mars in this context can be explained by the connection with the fighters in the amphitheatre, as attested by Martial, Spect. XXII, 3 and Statius, Silv., I, 6, 62 (cf. Golvin 1988: 337: n. 169). 350  In 189, where this same deity is evoked (maybe Diana?) as santne dia (l. 4) and domina (l. 7). 351  L. Schwinden (Rheinischen Landesmuseum Trier) has said (per litt.) ‘Die defixiones von R. Wünsch, Bonner Jahrbücher 119, 1910, sind zum Teil bei uns ausgestellt, nicht mehr alle Bleitäfelchen aus dem Amphitheater 1908/1909 sind erhalten’. 352  Together with the ministeria infernorum deu[m] and the larvae (ll. 18–20).

52

7 Categorization of the Defixiones 7.1. Introduction

our understanding,360 we cannot forget that even more often we are left in the dark simply due to the abbreviated nature of the texts. Obviously, it is difficult to identify the exact reason why these texts can be so terse. However, it does seem plausible that in many cases the practitioners orally gave more details and pertinent information. Another option would be that they took the omniscience of the deities invoked for granted: these powerful forces would surely be able to understand the message and take appropriate action.

Based on an analysis of the curses’ contents, it continues to be possible to categorize defixiones following the typology that Audollent first proposed back in 1904. He distinguished between the following: ‘tabellae iudiciariae et in inimicos conscriptae’; ‘in fures calumniatores et maledicos conversae’; ‘amatoriae’ and ‘in agitatores et venatores immissae’.353 While it is true that this taxonomy has generally stood the test of time and has been employed in various subsequent studies,354 it is equally true that scholars have proposed various modifications and/or additions. In 1991, for instance, H. Versnel suggested that the defixiones in fures ought to be reconceived of as ‘prayers for justice’ (and hence questioned their status as curses per se).355 Faraone, for his part, has proposed adding a new genre of defixiones: ‘commercial curses’, which are mostly attested in the Greek world.356 This same author has maintained that the earliest Greek defixiones all arose in agonistic contexts (to which the Latin curses also belong). In such contexts, the defigens and his/her victim compete with one another and use curses as a tool to increase their odds of winning that competition.357

On occasions, this ‘tacit agreement’ between the defigens and deity results in the practitioner simply listing the victims’ names. Providing this apparently minimal information was extremely important, because in magical thinking the very nature of an individual was intimately linked to his or her name.361 Simply inscribing personal names (whether in the nominative or accusative,362 whether alone or as part of a list) proved to be a long-lasting technique, which is attested from the fourth century BCE to the fourth century CE. Examples of listing names come from all corners of the Roman West, though this practice was most common in Italia, Britannia and Germania. Sometimes this frugal form was enhanced through the addition of other details, such as the naming of a victim’s profession or parents, that could be used to help identify the correct victim.363 As Gordon has noted, the proliferation of such lists should be explained by the influence that the everyday use of writing and public epigraphy had on magical practices.364

7.2. A tacit agreement Audollent had already recognized that ‘obscurae nimis sunt tabellarum magna pars quam ut diiudicare valeamus quem ad finem exaratae sint’.358 And indeed this could undoubtedly be applied to many curses from the Roman West. As Table 7.1 shows, in more than half of the known defixiones, only the defigens and (probably) the deities with which (s)he invokes knew the reason that led to the writing of the curse.359 That is, the practitioner did not think it was necessary to spell out in writing the circumstance that led him/her to compose the curse. While the poor state of conservation of a defixio often impairs

Within the larger group of curse tablets written for reasons that are today lost to us, there is an interesting sub-group of 40 texts written by somewhat loquacious defigentes. In certain texts, the author uses verbs, such as mando,365 commendo366 and devoveo,367 to refer to the handing over of These are generally quite fragmentary, as is the case with 8, 32, 139, 167–70, 224–27, 513, and the North African tablets DT 214, 255, among many others. 361  On this, see section I.4.1 and note 220. 362  There is no scholarly consensus about why both nominatives and accusatives are used in the same list of names. For some, this alternation is just a mistake, while for others it would denote practitioners (in nom.) versus victims (in ac.). For the Oscan texts, it has been argued that both nominative and accusative could be a way to distinguish different types of victims, which seems unlikely for the Latin curse tablets. 363  A matronymic qualifies the victims in 20, 235, 461 as well as the North African curses AE 1996, 1717–18 and DT 263, just to mention several examples. For the use of the matronymic, see López Jimeno 1991–92; Curbera 1999b (who stresses the Egyptian origin of this practice); Poccetti 2002: 36. The victims’ professions are named in 52, 101–02, 117, 495 and 502, among others. 364  Cf. Gordon 1999 and Centrone 2010. 365  Cf. 67 (ll. 6–7), 3 (ll. 1–2: Helenus suom nomen eimfereis mandat…). 366  Cf. DT 228 (A, ll. 2–4 and B, ll. 2–3: …commendo tibi Iulia Faustilla…). 367  Cf. 92 (A, ll. 9–10 and B, ll. 1–4: …aput vostrum numen demando devoveo desacrifico…) and 145 (B). 360 

DT, p. XC (for his taxonomy, see pp. LXXXVIII–XCI). Cf. Cesano 1961²: 1564f. and Graf 1995: 117. On the different types of curse tablets, in general, see Edmonds 2019: 65–74; Eidinow 2019: 351–52 and 356; Watson 2019: 62–64. 355  Cf. Versnel 1991 and 2010 (with previous references). His approach has been taken to the extreme by Urbanová 2018: 18–33, who neatly distinguishes between both categories in her book. 356  Cf. Faraone 1991a: 10–11. The category of ‘commercial curses’, which are mostly attested in a Greek context, are not considered in the present chapter, since there are only two known examples from North Africa (see Audollent 1933b). 357  See Faraone 1991a: 11ff. and Graf 1995: 153–54. 358  DT, p. LXXXVIII. 359  The tablets written for unknown reason constitute 390 of 623 or over 60 per cent. In Greek contexts we find something similar: according to Faraone, ‘more than three quarters of the published Greek κατάδεσμοι are inscribed with names only or are so laconic that they give us no hint whatsoever of their specific purpose’ (1991a: 10). 353  354 

53

Sylloge of  Defixiones from the Roman West Table 7.1. Defixiones from the Roman West, according to provenance (It.= Italia, Afr.= Africa, Hisp.= Hispania, Gall.= Gallia, Brit.= Britannia, Ger.= Germania, Raet.= Raetia, Nor.= Noricum, Pan.= Pannonia) and typology (E= erotic; A= agonistic; J= juridical; IF= in fures; U= unknown) Provenance Purpose

E

A

J

Uncertain/other categories

IF

U

It. (119)

3

1

9

2

E?: 48, 73. J?: 65, 105, 113. IF?: 112. Personal revenge: 4

Afr. (88)

14

37

12

1

A?: AE 1907, 165 A?: DT 290-291 Commercial: AE 1933, 234 and 235.

19

13

97

Hisp. (30)

1

1

7

5

E?: 129 J?: 148 IF?: 142

Gall. (55)

1

–

4

1

J?: 161, 182

47

E?: 353. J?: 339. IF?: 212, 243, 248, 264, 285-286, 336, 339, 417, 437. Against perjury: 299.

169

Brit. (256)

–

–

–

74

Ger. (59)

2

–

10

7

Raet. (6)

1

–

2

1

–

2

Nor. (1)

–

–

–

–

E?: 526.

–

Pan. (9)

1

–

4

1

3

Total (623)

23

39

47

92

390

40

the victims to the deities invoked. Indeed, we can even find phrases like ut hu(n)c (h)ostiam acceptum (h)abiatis, with which the practitioner turns his enemy into a sacrificial victim.368 In such instances, defigentes were plainly requesting that their enemies be killed (Watson 2019: 75f.).

(febris),370 madness (amentia) or anxiety (angustiae).371 More commonly, however, defigentes directly demand capital punishment. The target’s death is usually conceived of as the culmination of a slow and painful process during which the victim ought to suffer from other torments, such as fainting,372 torture,373 dismemberment,374 madness,375

But simply ending a victim’s life was not always the desired outcome. In fact, there is also a more common and imaginative group of curses in which deities are invoked so that they can punish victims in different ways. Such punishments are generally directed against a victim’s body but can also extend to other areas such as the victim’s financial ruin.369 Some practitioners ask that their enemies suffer from various physical and/or psychological illnesses, such as leprosy (lepra), cancer (cancer), fevers

For fevers, cf. 5 (l. 9: patiatur febris); for leprosy, cf. 57 (ll. 1–2: qui s[ubbertis]ti libra puerum Elissei ita subber[te] domum Ber[-c.2-]atis...’); for cancer, cf. 483 (ll. 9–10: ut illius manus caput pedes vermes cancer...). 371  For this topic and madness in particular, see 484, ll. 6–7: amentita surgat amentita suas res agat; for anxiety, see 16 (col. I, ll. 2–3 and 5–6: ... quomodo (ha)ec anima intus inc ̣ḷusa tenetur et angustiatur (…) ṣịc ̣ ụṭ anima ṃẹṇṭẹṣ c ̣ọrpos Collecticii quem pepereṭ Agneḷḷạ). 372  96 (A, l. 5 and B, ll. 1–2: Qụịntus Agrippini s(ervus) ụṭẹr ṣạḷṭụẹṇṣịṣ languiat aigrotêt ex omologi(s) fẹ̣ rị(s) igni(s) (…) devincit (…) et sic moriatur) and 115 (l. 3: ...[persequa]ris eum ut male contabescat usque dum morị[et]ur). For the death of victims in general, see López Jimeno 1997: 34 and Marco Simón 2009. 373  Cf. DT 292, B, ll. 2–4: ...demando tibi ex [hoc die] ex ac ora ex oc momento ut crucietur. 374  Like in 117 (ll. 16–18: corpus omnịṣ menbra (...) bisc[e]d[a] Porcelli qui (...) [cada]t languạt et ru[at?]) and 118 (ll. 17–19: disulvit̂e ̣ omnị[a] menbra omnis viscida ipsius). 375  Cf. the defixio from Cirta/Constantine (CIL VIII, 19525, A, ll. 9–11: ut facia[s] il[l]um sine sensum sine memoria) and 486 (A, l. 6: des ei malam mentem). 370 

368  According to 490 (A, ll. 5–6). Likewise, cf. the phrases found in 1 (ll. 2–4: …`h´anc (h)ostiam acceptam habeas et consumas…) and 121 (A, ll. 3–5: ...accipias corpus eius qui meas sarcinias supstulit...). 369  Like in 157 (A, ll. 1–4: quomodo hoc plumbu(m) non paret et decadet sic decadat aetas membra vita bos grano(m) mer(x)…) and 498 (A, ll. 2–5: qu[omo]di hoc liquescet se[-c.3- sic co]llum membra me[du]lla peculiụm d[e]l[i]ques[ca]nt).

54

Categorization of the Defixiones the liquefaction or combustion of the victim’s body376 or the paralysis of the body’s vital functions.377 Indeed, the extreme cruelty of the some defigentes can be shocking: one from Mogontiacum/Mainz asks that the victim be the witness of his own death,378 while another defigens asks to listen as her victim perishes (this, of course, would also offer irrefutable proof that defixio was successfully carried out).379

Figure 7.1. Erotic defixio of attraction, in which charaktêres and various individuals can be seen (two of which appear to be kissing; from Martin 1928: 59, fig. 2).

7.3. Erotic defixiones written by professional practitioners. The remaining curses have been found scattered throughout the western Roman provinces. Most of these tablets date to the Empire, though there are two erotic curses that date to the Republic.

Although there is literary evidence for erotic magic already in Archaic Greece,380 it was not until the fourth century BCE that we have the earliest examples of erotic curses.381 Traditionally, scholars have distinguished between two types of erotic curses382 depending on the desired outcome: the first category consists of the so-called diakopoi that seek to break up or at least interrupt an amorous relationship (this type is attested from the fourth century BCE to the third century CE).383 The second group is made of the so-called agōgai, which are meant to attract the amorous attention of the beloved (these curses are attested between the second and fourth centuries CE).384 The PGM provide evidence for both types of curses (though more attention is given to the agōgai), to the extent that the vast majority of spells and defixiones described in these handbooks are concerned with erotic magic.385

If we analyse the desired outcomes of these erotic curses, seven of the tablets can be classified as diakopoi: the earliest two examples (2 and 71) date to the second and first centuries BCE, while the other five date the High Empire.387 As is later recommended in the PGM,388 these texts frequently request that hatred arise between the two lovers, which would then lead to the end of their romantic relationship.389 Occasionally, such a request is accompanied by references to the victims’ physical appearance390 or sex life, which the curse seeks to interrupt (see Figure 7.1).391 When compared to the diakopoi, the group of Latin agōgai is much more substantial: there are 16 tablets, which date between the first and third centuries CE. Besides the two curses from Germania and Raetia (472 and 524), the rest of these erotic curses of attraction come from Hadrumetum, Carthage or Thysdrus/El Jem and were written by professional practitioners. Hence the similarities between the North African curses and the recipes found in the PGM: in both corpora, falling in love is treated as a sort of disease,392 which causes the victim (usually a woman)393 to suffer from various conditions, including insomnia,394 the paralysis of the body’s vital functions395 or the inability to

In marked contrast to the Greek world, where some authors have argued that as much as one quarter of all defixiones are erotic,386 only a small percentage of the curse tablets from the Roman West (23 of 623) belong to this category. Over half of these curses hail from North Africa (especially from Hadrumetum) and were certainly The liquefaction of the victim’s limbs ...quatmodum hoc plumbum liquescet (488 B, ll. 4–5) is a common request in the collection from Mogontiacum/Mainz. Accordingly, see 493 (ll. 14– 15: ...a[d qu]em modum sal in [aqua liques]cet sic et illi membra m[ed]ullae extabescant), 498 (cf. n. 369) and probably 499 (B, ll. 4–5: ...d[i]liquescant quat{m} modi hoc liquescet). On occasions, the way of disintegrating the victim’s limbs is through combustion, which appears to be alluded to in 16 (col. II, l. 1, where the practitioner asks that the body and soul of Collecticius ardẹạṭ) and perhaps also in 30 (l. 2: ...in foco (…) ut p(er)eat). 377  Cf. 68, ll. 6–8. The phrase neip putiiad, which Mancini (2006: 78) has identified as analogous to Latin ne possit, precedes a list of the vital functions that are to be paralysed. 378  Concretely, 486 (A, ll. 6–10): ...des ei malam mentem malum exitum quandius vita vixerit ut omni corpore videat se emori praeter oculos. 379  See 499, B, ll. 2–3: ...ut eorum ixsitum audiam. 380  In general, see Gager 1992: 78–83; Faraone 1991a: 13–15 and 2001 (where he looks back to work published in 1990, 1992, 1993 and 1994); Ogden 1999: 35–37; Kropp 2008: 184–86; Martin 2010: 113–32. 381  Faraone 1991a: 13, n. 59. 382  For other further classifications, see Gager (1992: 80, n. 8 and 9) and Dickie (2000: 565). Likewise, Faraone (1999 and 2001: VIII) proposes a new distinction: ‘rituals used mainly by men to instill erotic passion in women (êros) and those used primarily by women to maintain or increase affection (philia) in men’ (on this differentiation, also see Dickie 2000: 582–83). 383  Cf. Homer, Il. XIV, 216ff. In this passage, Hera asks Aphrodite to lend her a girdle ‘in which are fashioned all manner of allurements; in it is love, in it desire, in it dalliance―persuasion that steals the senses even of the wise’ (translated by A.T. Murray). 384  For the dates, see Faraone 1991a: 15 with n. 66. 385  Cf. Petropoulos 1988 and Dickie 2000: 565–73. 386  As Gager 1992: 78 and Ogden 1999: 35 assert. 376 

140 dates to 70 CE, while 15 and 173 have been dated to the second century CE. 388  Cf. PGM XII, 370–75; LXI, 39–47, 60ff. and O2, 25ff. 389  Cf. 71 (IA, l. 6: ut il(l)ic il(l)a(n)c odiat) and 2 (A, ll. 1–6: ...quomodo mortuos qui istic sepultus est nec loqui nec sermonare potest seic Rhodine apud M(arcum) Licinium Faustum mortua sit nec loqui nec sermonare possit). 390  Cf. 71, IA, ll. 2–3: ...facia(m) capi(l)u(m). 391  Thus, in a curse from Rome (cf. 15), a persuasive analogy is established between the victims and the medium in which it was inscribed (a small sculptural group depicting two adults flanking a child) so that the victims ...non qumbere inter s[e] (A, l. 14); that is, so they could not maintain sexual relations. Likewise, the mention of a penis in the curse from Aquinum/Aquino (cf. 73, l. 10; although the reading seems problematic) and of a fututor (‘copulator’) in the curse from Maar (cf. 173, ll. 1–2) could be connected to a possible request for a man’s impotence, like that that Ovid describes at Am., III, 7.27–36 and 73–84). 392  Cf. Winkler 1991: 222–24. 393  Among the tablets studied here, the only exceptions are 472 and a curse from Hadrumetum (DT 270). Likewise, of the 35 erotic recipes found in the PGM only 5 have male victims. 394  Insomnia plagues the victims of DT 230 (A, l. 2), 265 (A, ll. 8–9), 266 (l. 7), etc. 395  This includes being unable to eat (DT 266, ll. 7–8), sit down (DT 270, l. 7) or speak (DT 270, l. 8). For a discussion, see Martinez 1995. 387 

55

Sylloge of  Defixiones from the Roman West remember her loved ones.396 These conditions are supposed to continue until a burning397 and uncontrollable398 passion drives the beloved into the defigens’ arms. The aggressive and profoundly symbolic399 language used in the PGM and the North African curses contrasts with that found in the two curses from Europe, which are milder and only ask that their victims love the respective defigentes.

is based on Pelops’ famous prayer to Poseidon in Pindar’s Olympian 1.403 That said, the earliest archaeological evidence from the Greek world dates to the third century BCE and consists of curses directed against actors and athletes. In contrast, agonistic curses from the Roman West concern the games from amphitheatres and circuses, spectacles whose popularity soared during the High Empire. Despite differences in exact context, Gager is certainly right to stress, ‘what Greek and Roman performances shared was a keen sense of competition, copious rewards and enormous popularity’ (1992: 43).

Finally, we must mention other curse tablets that perhaps should be grouped together with the erotic defixiones (in Table 7.1 these curses are marked with ‘E?’), though their texts are not as explicit as those just discussed. Amongst these, a curse from Faviana/Mautern (cf. 526), which should perhaps be classified as a diakopos, targets a woman named Silvia and asks that inversu(m) maritu(m) c{e}ernis (526 B, ll. 1–2; perhaps an attempt to break up a couple?). A curse from Rome (cf. 48), which for its part could be possibly included among the agōgai, targets Caecilia Prima and asks (among other torments) that various parts of her body burst into flames.400 Furthermore, the text repeats the word turpidines (cf. 48 B, ll. 3 and 5, for turpitudines), a term that is synonymous with obscenitas, that is ‘obscenity, lewdness’ (Lewis and Short, s.v. II). In addition, the defigens, probably a man (cf. 48 B, l. 33), invokes Aurora, who was considered to be a ‘hunter of young lovers and a destroyer of romantic relationships’ (Bevilacqua 2006–07: 327). Another curse, which comes from Old Harlow (cf. 353), may have been written because of an amorous rivalry, since there seems to be reference to an affair (‘negotium Ettern(a)e’, 353 A, ll. 2–3). Finally, it is worth noting that scholars have generally interpreted the anatomical curses, in which a victim’s body and vital organs are directly cursed, as erotic defixiones, even if there is no direct evidence for this in the texts themselves.401

For their part, the PGM contain two recipes for agonistic curses: the first (PGM III, 1–164) calls for the drowning of a cat in whose body the practitioner is told to place three tablets and then wrap the cat’s body in a fourth defixio written on a sheet of papyrus on which chariots, charioters, chariot boards and racehorses have been drawn.404 Next, the defigens is told to place the animal in a grave or in the stadium405 and then sprinkle the area with the water used to drown the cat. The second PGM recipe (IV 2210–17), which is less elaborate, urges the practitioner to burn garlic and a snake’s slough before inscribing a cursing formula on a tin tablet, which should then be deposited in the grave of someone who had died prematurely. Although the archaeological record has not turned up physical evidence for the rituals described in the PGM, there are 39 agonistic curses written in Latin, almost all of which hail from North Africa. Among the exceptions that come from elsewhere, the earliest is a curse from Astigi/Écija (cf. 132) dated to the first century CE) that targets the grex Antoniani, 2 factiones, 8 agitatores and 13 quadrigae. So far, this is the only agonistic tablet from Hispania, and the only one of this genre from the whole Roman West that has been written by an amateur practitioner. In support of this hypothesis, we note its very simple structure: a list of names both in the nominative and the accusative. In addition to this defixio, we must also add a curse from Rome (cf. 5) belonging to the ‘Sethian’ collection and dated to either the fourth or fifth century CE. This curse, which was written by a professional practitioner, only tangentially belongs to the group of agonistic curses, since it targets Praeseticius, the miller who worked for one of the stables belonging to rival factiones.406

7.4. Agonistic defixiones Provoked by the extreme intensity that characterized theatrical and athletic competitions in Antiquity, agonistic defixiones first arose in the fifth century BCE.402 This date Cf. DT 230 (A, l. 9–11), 266 (ll. 15–17: ...ut obliviscatur patris et matris et [propinquor]um suorum et amicorum omnium [et aliorum] virorum), 268 (l. 3), etc. 397  The ‘heat’ of passion is clearly communicated through the use of verbs like uro, comburo and aduro, which are found in DT 227 (ll. 1–3), 270 (ll. 6, 9–11, 18 and 19) and in those curse tablets published by Audollent in 1908 (frag. 1, l. 7 and frag. 2, l. 5) and 1930 (l. 4). 398  See DT 266 (l. 19: ...solum me in mente habeat), 270 (ll. 8–9) and Audollent 1930, l. 9. 399  Cf. Gager 1992: 81f. and Versnel 1998: 252–58. 400  With verbs like uro, peruro and aduro. Cf. 48 A, ll. 17–19 and B, l. 17. 401  This is the case with 56, 129 and 535, all of which target women. To these, we could add 51, which, according to Audollent (1906: 367) was written by ‘un amoureux dedaigné [qui] voue au malheur la femme qu’il aime et son rival’. In this text, the author, who curses his victims limb by limb, asks that their bodies be dissected; after being inscribed, the defigens pierced the tablet so that the hole coincided with the word pectus (between A, ll. 4–5 and B, ll. 3–4), which hardly seems to be accidental. For the piercing of the tablet coinciding with specific words or formulae, see section I.5.4. 402  Faraone 1991a: 11–13 and Ogden 1999: 32–33. For agonistic curse tablets in general, see Gager 1992: 42–49; Kropp 2008: 181–84; Martin 2010: 93–98; Tremel 2004 on curses and sporting events; Gordon 2005 and 2012b. For an extended version of this section, which includes Greek and Latin defixiones in the Roman West, see Sánchez Natalías 2020c: 72–74. 396 

Of the remaining curses belonging to this group, most were deposited in the necropoleis of Hadrumetum and Carthage, though there are two series of curses that were discovered in Carthage’s circus and amphitheatre. The 403  Cf. Pind., Ol. I, 75–78, where the hero urges the god to stay his rival’s spear. 404  PGM III, 19–20: τὰ ἅρμα καὶ τοὺ[ς] ἡνιόχους καί [τοὺς δί]φρους καί τοὺς μονάτορας. 405  This is following the proposal of Heintz 1998: 342, n. 34, where the reconstruction [τῷ σ]ταδ[ίῳ] is offered. 406  According to the interpretation of Solin (1976: 90 and 1995: 118). Twenty Greek tablets from the same collection belong to the group of agonistic defixiones (see Tremel 2004: 70–90).

56

Categorization of the Defixiones spatial and temporal proximity of these curses (all date to the High Empire) help explain the various shared formulae, which originated in a small number of models. Furthermore, tablets can occasionally be grouped together based on the use of the same charaktêres, onómata or even iconography.407 Direct analysis of some of the series has revealed that the curses were in fact written by the same hand.408 Within the group of North African agonistic defixiones, the different objectives that defigentes pursued allow us to distinguish between two subcategories: curses pertaining to the amphitheatre and those related to the circus. The former, discovered at the close of the nineteenth century in what could very well be the spoliarium from Carthage’s amphitheatre, consists of 55 tablets (only 9 of which have been published).409 Most of these texts target venatores, also known as bestiarii, who were the stars of the venationes, a type of spectacle dedicated to hunting of wild animals. These games were first held in the circus and then later in the amphitheatres between the second century BCE and the sixth century CE.410 Generally, Carthaginian practitioners hoped that the venatores would be publicly vanquished in the amphitheatre.411 At times, the defigens asks that a venator suffer from insomnia the night before entering the arena.412 As a result, he would be too weak and sluggish to catch413 or kill414 his prey and, finally, the hunted would end up killing the hunter (see Figure 7.2).415

Figure 7.2. DT 246: a servant characterized as Mercury Psychopomp checks the death of another gladiator in the arena (From Peyras 1996: 131).

importance given to the factiones (for Procopius, these rivalries could divide a city)417 and aurigae. According to Brown (1970), charioteers, as clients of the local aristocracy, represented certain political and social interests, which made them into ‘undefined mediators’ between these aristocrats and the lower classes. Given the extraordinary nature of the rivalries between different factiones and aurigae, daily life was spiced up with fistfights, gambling and, of course, defixiones, which were considered an indispensable tool for determining the outcome of a given race.

The majority of agonistic curses from North Africa, however, are oriented toward the circus, a fact that attests to the great popularity of these games and the emotions that they provoked in Imperial Carthage and Hadrumetum. Indeed, the passion and excitement that is reflected in the defixiones is also attested in the literary record. Thus, Ammianus Marcellinus (Historia XXVIII, 4) states that for the die-hard fans of chariot races, the Circus Maximus was ‘their temple, their dwelling, their assembly and the height of all their hopes’.416 Other authors lament the

These tablets were mostly deposited in the city necropolis418 and date to the High Empire. Written in both Greek and Latin, the bulk of Latin texts hail from Hadrumetum, whereas the majority of Greek texts come from Carthage (where only four such Latin defixiones have been found).419 Whatever language used for writing these circus curses, the objective was always the same: destroy the other factiones. With this aim in mind, practitioners would write the names of aurigae as well as their horses

See Gordon 2005: 69–76 and Németh 2011. As I was able to confirm during the autopsies of the tablets DT 276– 84. These are part of my current research project, ‘The Latin defixiones from North Africa Revisited’ (DeLAR). 409  Specifically, there are DT 246–54. Of these tablets, only DT 252 and 246 were written in Greek. For the archaeological context, see section I.6.5.3. 410  According to Mancioli 1987: 66–68. This spectacle was brought to Rome by M. Fulvius Nobilior to celebrate his triumph over Aetolia and Cephallenia (187 BCE; cf. Livy XXXIX 22, 2). The practice was banished from Italy by Totila in the sixth century CE. 411  Cf. DT 247, ll. 1–6: ...[occi]dite exterminate vulnerate Gallicu quen peperit Prima in ịsta ọrạ in ampiteatri; DT 248 B, ll. 1–4; DT 250 B, l. 6, among others. 412  Cf. DT 250 A, ll. 4–5: ...Ιεκρι auferas somnum, non dormiat Maurussus quem peperit F[e]licitas. 413  Cf. DT 247, l. 10: ...non liget ur[su]; DT 250, B, l. 10: ...nec lac ̣[ueos] possit super ursum mittere and DT 253, ll. 12–13 and 19–20. 414  Cf. DT 247, ll. 15–18: ...ụṭ neque ursu neque tauru singulis plagis ọc ̣cida[t n]eque binis plagis occidt neque ternis plagịṣ oc[ci]dat tauru ursu; DT 253, l. 20. 415  Cf. DT 247, l. 20: ...allidat illu ursus et vulneret illu; DT 250 B, l. 18: ẹt remiṣe ferarum morsus. 416  Translated by J.C. Rolfe. Cf. also Tac., Dial. 29. 407  408 

417  Cf. Procop., Wars I, 24. And so it happened in 532, when the greens and the blues united against Justinian in the Nika riots. The uprising was brought to an end after six days of unrest thanks to the decisive intervention of Theodora, who reminded the emperor ‘purple makes for a lovely shroud’. Thus, she interrupted the plan to flee that was being organized. 418  Except for a group of 13 tablets discovered inside the circus of Carthage, which were specifically found across from the carceres and near the spina. For a discussion, see section I.6.5.3 n. 336. 419  Ogden, for his part, maintains, ‘it is curious that even in the Latin west circus curses tend to have been written in Greek: perhaps because most charioteers came from the Greek east’ (1999: 33).

57

Sylloge of  Defixiones from the Roman West circle around one of his eyes (the exact one depended on whether he was representing the plaintiff or the defendant) to protect himself from the evil eye (Pliny, Ep. VI, 2.2). Pliny the Elder, for his part, explains that carrying part of a hyena’s intestinal tube during a trial could lead to success in law-courts and help a speaker win certain types of cases (among other risky situations), by it (cf. HN XXVIII, 106).

and ask that the animals’ feet be tied together,420 that they would be unable to move421 and hence fall to the ground422 and lose the race.423 Defigentes also sought to impair the aurigae, asking that they be unable to steer their chariots and control their horses,424 or to see their opponents;425 all of this would prevent them from ultimately winning the race. For some defigentes, however, such requests did not go far enough, and in several tablets we find much more drastic demands: the famous series of curses invoking Baitmo Arbitto, for example, ask that the aurigae and their horses be dismembered426 and tortured427 before dying.428

The PGM, for their part, contain two references to the use of amulets in courtrooms433 as well as a recipe for a juridical curse tablet.434 In the latter, the defigens is instructed to use the lead from a yoke to make a tablet on which he or she would write the following formula: ‘restrain the wrath of him, (...) and the anger and tongues of everybody, in order that they might not be able to speak to him’.435 Next, the practitioner is to fumigate the tablet with incense and then put it under the sole of his or her left foot.436

7.5. Juridical defixiones While the previous section addressed agonistic curses centred on the sporting events held in the amphitheatre and the circus, we can also identify strong agonistic elements in curses dealing with the law courts. In these spaces, rhetorical and legal competitions took place and would be settled only after the (im)partial decision of the iudex was delivered.429 In this setting, defixiones proved to be an important resource that was undoubtedly thought to improve one’s chances in court, since they sought to silence the accusations and testimony that could implicate or harm a defigens. These texts, which were written before a verdict was reached,430 reveal a complex range of emotions, such as fear, guilt and shame, provoked by the prospect of public humiliation.431 The potential use and influence of defixiones in the legal realm was well known in Antiquity and could even be bandied about as a justification for a lawyer losing train of thought while speaking. In this regard, Cicero mentions Curio, ‘who all of a sudden forgot his entire case and said that this had been caused by Titinia’s spells and incantations’.432 Also well known is the case of the orator Marcus Aquillius Regulus, who used to paint a

The use of juridical defixiones is first attested in Selinunte in the sixth century BCE.437 Indeed, such curses played an important role in the Greek world, from which we have 67 identified items.438 Interestingly enough, the formulae found in these Greek curses were later taken up and employed in the Roman world. So far, we know of 47 juridical defixiones from the Roman West, which mostly hail from Italia, Africa, Hispania and Germania.439 The earliest examples, which date to the Republic and were found in Italia and Hispania, were clearly influenced by Greek curses.440 That said, we must remember that most of the non-Greek juridical defixiones from the Roman West date to the High Empire. In this group of defixiones, it is commonplace to brand the victim as an inimicus441 or adversarius,442 though occasionally more precise lexemes are used to specify someone’s particular role in a legal proceeding. Such

Cf. Jordan 1988: no. 3, col. C. For one of many examples, see DT 233, ll. 29–32: ...trado tibi os equos ut deteneas illos et inplice[ntur] [n]ec se movere posse[nt]’ and DT 275, ll. 29–30: ...ne currere possint nec frenis audire possint nec se moere possint. 422  A good example would be DT 272 A, ll. 11–12: ...cadan frangan disiungantur male guren, among many others. 423  Cf. DT 272, A ll. 12–13: ...palma vincere [n]on possin; DT 280, l. 14; DT 281, ll. 13–14, etc. 424  A good example is DT 275, ll. 32–33: ...nec lora teneant nec retinere equos possint, among many others. 425  Cf. DT 275, ll. 33–34: ...nec adversarios suos videant, among others. 426  Such as in DT 288 B, ll. 5–7: ...auferas ab eis nervia vires medullas impetos victorias; DT 289 B, ll. 6–7. 427  Cf. DT 292 B, l. 4; DT 293, ll. 5–6 and DT 294, ll. 9–10. 428  Cf. DT 286 B, ll. 4–10; DT 287 A, l. 12 and DT 295, l. 25. 429  For a similar assessment, see Bablitz 2007: 199, who affirms, ‘the Roman courtroom was an arena where parties entered into combat’. For juridical defixiones generally, see Faraone 1991a: 15–16; Gager 1992 116–24; Kropp 2008: 180–81; Martin 2010: 47–62; Scholz 2011a. For an extended version of this section, which includes Greek and Latin defixiones in the Roman West, see Sánchez Natalías 2020c: 70–72. 430  See Faraone 1991a: 15 and especially n. 67. 431  See Gager 1992: 116–17. 432  Cicero, Brutus, 217: subito totam causam oblitus est idque veneficiis et cantionibus Titiniae factum esse dicebat (translated by D. Ogden 2001: no. 172). In addition, we should mention Aristophanes Vespae 946–48, where we read of an orator whose jaw froze up during a speech before the jury (see Faraone 1989). 420  421 

See PGM XXXVI, 65–68 and IV, 2145–240. See PGM VII, 925–39. 435  PGM VII, 934–39: κάτεχε τὴν ὀργὴν τοῦ δεῖνα καὶ πάντων τὸν θυμὸν καὶ τὰς γλώσσας ἵνα μὴ δυνηθῶσιν λαλεῖν τῷ δεῖνα, translated by R.R. Hock (apud Betz 19922). 436  This final performance has been related to the representations of captives in royal sandals, through which the Pharaoh constantly commemorates past victories and encourages future ones. Cf. Faraone 1991b: 173–74. 437  That is SGD 95. 438  According to Faraone 1991a: 16, whose data has also been used by Gager (1992: 117) and Ogden (1999: 31). 439  To a lesser degree, these tablets are also found in Gallia, Pannonia and Raetia. See Table 7.1. 440  This influence is especially clear in the tablet from Marsala (cf. 109) which dates to the third century BCE and was written in a mixture of Latin and Greek. The curse targets a group of individuals whom the defigens would like to silence. The remaining tablets from the Republic (dated to the first century BCE) are 63–64, 66, 103, 49–50, 125–26 and 137. To these, we should probably add a curse tablet from Corduba (on this, see García Dils de la Vega and Rubio Valverde 2018: 278–81). 441  Cf. 103 (Col. II, ll. 5–6: ...si quis [i]nimicus inimi[ca] adve[r]sariu hostis), 17 (B, l. 1), 137 (l. 10), 182 (l. 1), 160 (Tab. I, l. 10), 468 (A and B, l. 1), 470 (ll. 1, 2 and 9), etc. 442  Cf. 103 (Col. II, l. 6; see the previous note), 49 (A, ll. 5–6: ...si qui arṿosarius auṭ arvosaria), 17 (B, l. 6), 135 (B, l. 4–5), 478 (A, ll. 3–4), 528 (l. 9), 529 (A, l. 1 and B, l. 5), DT 221 (l. 10), among others. 433  434 

58

Categorization of the Defixiones 7.6.1. A different approach: prayers for justice?

terms include advocatus,443 testes,444 iudices445 as well as ‘et omnes qui illi ass[unt] et doc[e]n[t illu]m…’.446 Since they were well aware of the danger that damning testimony provided, defigentes are pre-eminently concerned with silencing opponents and witnesses:447 if a defigens could immobilize their tongues448 and impair their mental faculties,449 their opponents would be unable to respond to questions in court,450 testify and, ultimately, win a suit.451

As alluded to above, Versnel has proffered a new definition of this category of inscriptions and has questioned whether or not they should truly be classified as defixiones.454 Instead, he has defended a new category that he calls ‘prayers for justice’ and defined as ‘pleas addressed to a god or gods to punish a (mostly unknown) person who has wronged the author (by theft, slander, false accusations or magical action), often with the additional request to redress the harm suffered by the author’.455

7.6. Defixiones against thieves

Some of the defining characteristics of the ‘prayers for justice’ include the use of epithets and honorific titles to name the deities and the humble and respectful tone used by the authors. They normally justify their action and, on certain occasions, give their own names to the deity invoked. While these characteristics have been used by Versnel to distinguish ‘pure’ prayers for justice from curse tablets, the scholar is also aware of the vast area left in between these two categories. In fact, he labels an important number of texts as ‘borderarea cases’, in other words, texts whose characteristics combine aspects of curse tablets and ‘prayers for justice’. The acknowledgment of such a notable grey zone demonstrates just how difficult it is to establish hard lines between categories.

At the time when Audollent put forth his taxonomy of curse tablets in 1904, the category of defixiones in fures, calumniatores et maledicos conversae was the smallest, with only five examples from the Latin West. Today, this is no longer the case: the important archaeological discoveries of the final third of the twentieth century have turned this once small group into the most fully attested type of curse from the Roman West.452 Indeed, this type of curse tablet now constitutes almost a sixth of the entire corpus.453 This increase in the number of instances has piqued scholars’ interest in this type of text so much so that certain authors have argued that the category does not really exist.

443  Cf. 63 (frag. C, ll. 10–11), 134 (A, l. 5), 135 (A, ll. 6–7), 160 (Tab. I, ll. 8–9), etc. 444  Cf. 63 (frag. C, ll. 10–11). 445  Cf. 22, ll. 13–14: ...arbitri Surae. 446  Cf. 471, ll. 3–5. Furthermore, in 520 A, ll. 4–5, any overlooked adversary is included with the expression ...et quisquis adversus il(l)am loqut. 447  The desire to render a victim mute is expressed with the verb obmutesco (Cf. 160, Tab. II, ll. 6–8: ...quomodi in hoc m[o]nimont animalia ommutuerun(t) nec surgere possun nec illi; 529, B, l. 6, etc.) or more exceptionally, taceo (Cf., perhaps, 277, l. 3: …tacituri...). Many curse tablets employ phrases with mutus (e.g., 131, ll. 1 and 4: ...mutus tacitus siet; 522, A, ll. 1–2: ...mutus sit Quartus and l. 4: ...ut e[i]us os mutu(m) sit Mutae; DT 219, A, ll. 2 and 8–9: ...facias ilos muttos). Reference is made to robbing a victim of the power of speech with the lexeme λαλια, found in 109 (A, l. 7); also see the phrases in 478 (A, ll. 5–6 and B, ll. 3–5: ...neque loqui pos[s]it); 529 (B, ll. 6–7: ...ne contra nos lucuia(nt)); DT 217 (A, ll. 2–3: ...nec dicere nec facere); etc. Finally, note the Oscan curse 66, which reads ...nep fatíum nep deíkum pútían`s´ (ll. 6 and 8), which Vetter (1953: no. 4) has interpreted as equivalent to Latin nec fari nec dicere possint. 448  Such as the Oscan 64 (ll. 6–7: ...olu solu fancua recta sint, interpreted by Vetter (1953: no. 7) as ...ut illorum lingua rigidae sint), the Latin 532 (ll. 7–9: ...eo modo hoc ego averso graphio scribo sic linguas illorum aversas ne pos(s)int facere contra (h)os…) and DT 218 (ll. 6–7: ...adligate linguas); among many other cases. Additionally, perhaps we should add the Oscan 65, an anatomical curse that targets the tongue (fanguam, l. 3). 449  Cf. 109 (A, l. 6: ...καὶ ψυχὴν αὐτοῦ ὡς μὴ δύνα[ται], repeated on A, l. 8 and B, l. 2), 522 (A, l. 7: [Qu]a[rt]us ut insaniat), etc. 450  Cf. 467 (B, ll. 1–3: ...sic non possit respo[nde]re qua(e)s[tionibus]), 470 (ll. 8–9: ...nusquam contra nos [inve]ṇissẹ respoṇs[io]nis cum loquantur inimici); DT 218 (ll. 8–9: ...ne adversus nos respondere), etc. 451  Cf. 160, Tab. I, ll. 5–6: ...nec illi hanc litem vincere possint. 452  Mostly in Britannia, where the sanctuaries dedicated to Sulis Minerva (Aquae Sulis, Bath) and Mercury (Uley), together with scattered findings, have provided an impressive collection of evidence. 453  Authors like Gager (1992: 177) and Ogden (1999: 38), nevertheless, include in this category all of the defixiones found in Aquae Sulis/Bath and Uley. This is suspect since many of the texts are illegible or too fragmentary to be securely classified as curses against thieves. For this type of curse in general, see Kropp 2008: 186–89; Martin 2010: 67–79; Scholz 2011b; Urbanová 2018: 24–30 and 180–203. An extended version of this section has appeared in Sánchez Natalías 2022.

The debate over and interest in Versnel’s proposal has nearly divided the scholarly community into those who embrace the category of prayer for justice (e.g., Faraone, Jordan, Tomlin and Urbanová), those who are sceptical of it (e.g., Gager, Ogden, Kropp, Martin and Gordon) and those who have rejected it outright. Among the last group, M. Dreher has been the adamant in the recent years, arguing that these texts would be better defined as defixiones criminales. For this scholar, this new typology has nothing new when compared with the old Audollent’s label defixiones in fures. In fact, the high number of ‘border-area cases’ should preclude us from making any neat distinction between curses and prayers for justice (Dreher 2012: 30), just as the authors of these texts themselves probably did not think in such terms.456 But, as I said, other experts in the field have only rejected certain aspects of the so-called ‘prayers for justice’. Scholars like Ogden (1999: 38–39) have pushed back against Versnel’s idea according to which ‘gods other than the usual chthonic deities are often invoked’ (2010: Versnel 1987 and especially 1991, 2010 and 2012. Versnel 2010: 279, where he adds, ‘The other typical features are: 1. the principal states his or her name; 2. some grounds for the appeal are offered; this statement may be reduced to a single word, or may be enlarged upon; 3. the principal requests that the act be excused or that he be spared the possible adverse effects; 4. gods other than the usual chthonic deities are often invoked; 5. these gods (…) may be awarded a flattering epithet (…) or a superior title (…); 6. words expressing supplication (…) are employed as well as direct, personal invocations of the deity; 7. use of terms and names referring to (in)justice and punishment’ (also see Versnel 1991: 68). For a contrary view, see Gager 1992: 179, n. 2. 456  Cf. Gordon 2013: 267f.; Faraone and Gordon 2019: 322f. 454  455 

59

Sylloge of  Defixiones from the Roman West 279). Ogden has rightly pointed out that among the deities mentioned in these curses there still seems to be a preference for those who have chthonic connections, like Mercury,457 Demeter458 and certain infernal demons.459 Indeed, we could easily expand this list to include Proserpina (cf. 120), Dis Pater, Cerberus and Veracura (cf. 530), Hades (cf. 121), Nemesis (cf. 337) and Diana (cf. 343). Finally, we ought to add that even if available evidence allows us to form general ideas about deities like Sulis, Metunus and Niskus, the fact remains that little is known about these figures and questions about whether they are chthonic or not still have to be resolved.460

her own.464 Generally, the practitioner gives his/her name in the nominative, sometimes followed by a patronym and, exceptionally, by a matronym. A curse tablet from Aquae Sulis/Bath, 303, provides a good example of this: in reporting the theft of six silver coins, the practitioner not only identifies himself as Annianus, but also provides his mother’s name (Mantutina). The addition of this final piece of information is certainly extraordinary and reflects the author’s desire to be clearly identified by the deity (as usually happens with erotic curse tablets). Even if Versnel has moved the debate forward more than any other scholar to date, no consensus has been reached and the validity of the category ‘prayer for justices’ remains an open and important question within the larger field.465 As a scholar, I am aware of how useful taxonomies can be for understanding any phenomenon from Antiquity. That said, the debate in this case has gone too far, and the rigidity of these taxonomies is no longer helpful. Given that curse tablets worked as a magico-religious technology that was adapted by the different societies that employed them, it is much more useful to understand these artefacts from a more emic perspective (contra Versnel 2010: 326). To establish a hard border between ‘prayers for justice’ and curse tablets seems not only artificial but also unnecessary. Doing so just brings back to the great dichotomy between ‘magic’ and ‘religion’ that scholarship has sought to abolish for the last decades. As Gordon has pointed out, curse tablets are better understood in terms of instrumental religion, that is, as a tool that served individuals to solve their specific, personal crises. As is always the case, the way we categorize something largely depends on our own point of view: in Gordon’s words (2013: 269), ‘the instrumental religion of one party is the malign magic of the other party. It is entirely a matter of perspective’.466 The fact that a certain practice was not institutionalized or sanctioned by the Roman authorities does not mean that it was not normalized within specific social contexts and communities.467

Scholars have also been wary of Versnel’s claims about the humble language and respectful tone found in these tablets (e.g., the use of epithets and honorific titles). Thus, Kropp has analysed these features as a ‘protective strategy’ through which the author (i.e., the defigens) finds a way to communicate with a supernatural audience, which in the author’s eyes is at once mighty and possibly dangerous.461 In Kropp’s own words, ‘the defigens articulates his awareness of the status difference between him and the god invoked, insofar as he gives his request a non-peremptory character and, at the same time, qualifies his own position as dependent and inferior’ (2010: 366). Finally, let’s address the question of an author identifying him- or herself in the text. While it is true that in the bulk of known curse tablets defigentes do not normally name themselves (though, for obvious reasons, we find exceptions in certain erotic curses),462 in the present group of texts some writers share their own names with the deity invoked. This has been taken as evidence for the absence of fear before the deity, since they are asking for justice.463 But, at the same time, such reasoning would imply that all the authors who did not specify their names were, in turn, afraid of the deity. In my opinion, this conscious decision can be understood in a different manner: as the result of the varying perceptions that practitioners had about divine omniscience. For some of them, gods were omniscient and able to find the stolen property with almost no information at all. For others, giving the gods more information about a case would be helpful and increase the chance of a successful resolution. In this scenario, where (obviously) the authors of the texts did not know the identity of the wrongdoers, the only name that could possibly be provided was his or

To conclude, Versnel has identified important differences among the so-called ‘prayers for justice’ and the larger corpus of defixiones. Despite these differences, all these tablets are fundamentally a private and direct means of communicating with the divine. After considering the supposed differences between the two groups, there remains, in my view, more fundamental commonalities, Nevertheless, within Britannia, the practitioner identifies him- or herself in less than half of the defixiones in fures (for some examples, see 205, 210, 237, 346, 350, 356, 451, etc.). Outside of Britannia, this sort of self-identification is securely documented in four further curses (479, 483, 492, where the defigens identifies herself with her husband’s name, and 523). For a cognitive approach to the self-identification of practitioners in British curse tablets against thieves, see Sánchez Natalías (forthcoming). 465  See Gager 1992: 175–78; Graf 1995: 155f.; Ogden 1999: 37–39; Martin 2010: 67–68; Dreher 2010 and 2012; Gordon 2013: 266–69 and 2014: 783–84; Faraone and Gordon 2019: 321–23. Versnel (2010: 277– 78, 324–27 and 2012) has answered some of these critiques. 466  On this, see also McKie and Parker 2018: 2f. and McKie 2019: 445. 467  Such as the sanctuary of Isis and Mater Magna in Mainz, for which see Veale 2017. 464 

Cf. 346, 355–56, 358, 361, 459, 523, etc. 458  Cf. DT 2–3, 6, 11–12 and SGD no. 60. 459  Cf. DT 74–75, SGD no. 21. 460  Ogden maintains, ‘while Minerva had no chthonic associations, Sulis’ sacred spring was a body of underground water’ (1999: 38–39). 461  Cf. Kropp 2010: 366, where she notes that the ‘interaction between human-beings and supernatural powers, especially infernal ones, invites such strategies–we may recall that prayers to the gods of the underworld were not spoken aloud.’ 462  Except in the case of erotic defixiones, where the PGM recommend that the practitioner include his own name, a practice which, as is well know, is found in the curses themselves. 463  As Tomlin puts it, ‘the writer is asking for justice, and is not afraid to give his name’ (2002: 16). 457 

60

Categorization of the Defixiones including formal characteristics (types of materials used, deposit contexts, manipulation, etc.) and intent. Indeed, these tablets reveal the same emotions (e.g., anger, wrath and desperation) that underlie the larger group of defixiones.468 As should be clear at this point, I prefer Audollent’s denomination ‘defixiones in fures’, which I use throughout the rest of this volume.

I hand over the thief whom I see’. At this point the papyrus adds: ‘If one of them does not swallow what was given to him, he is the thief.’474 The PGM recipes as well as the texts from Asia Minor share a common feature: the public confession of the guilty party. Such a confession, however, is not attested among the curses against thieves from the Roman West (regardless of the language in which they were written). Thus, it is worth wondering whether these types of confessions were local solutions to a similar problem: how to catch a thief.

7.6.2. Defixiones against thieves (or in fures) Written in a response to a theft, curses against thieves seek to recover a stolen object and/or punish the wrongdoer. So far, the earliest examples of such curses come from the Greek world, and more specifically from Cnidus in Asia Minor. There, excavations carried in the nineteenth century by C.T. Newton uncovered the temple of Demeter, in whose temenos they ‘found in several places portions of thin sheets of lead, broken and doubled up. On being unrolled, these sheets proved to be tablets inscribed with imprecations’ (Newton: 1863: 382).469 In six of these texts,470 dated between the second and the first centuries BCE,471 the defigentes denounce the theft that they have suffered and ask that the culprit publicly admit to his or her wrongdoing while being wracked with fevers and other torments.

When it comes to the Roman West, and besides 20 tablets mostly hailing from Hispania and Germania,475 the overwhelming majority of the defixiones in fures come from Britannia and date between the second and fourth centuries CE. There, most of the evidence has been found in the sanctuaries of Sulis Minerva (Aquae Sulis, Bath) and Mercury (Uley). That said, it is worth stressing that there are also over 20 tablets that come from different parts of the island. This would mean that, so far, and without counting the dense concentrations of the main sanctuaries,476 the number of curses against thieves is still higher in Britannia than in any other province from the Roman West. In addition, there is also another important point to consider: of the whole corpus of British curses, the only tablets that explicitly record the reason for which they were written are the ones in response to a theft. The rest of the texts are either too fragmentary or too sparse for us to determine the causa defigendi. Was the technology of curse tablets being adapted in Britannia for an acutely felt need?

The desired confession attested in the curses from Cnidus has often been compared to certain Lydian and Phrygian stelae,472 dated in between the second and the third centuries CE, in which the guilty party publicly confess their crime while asking for reconciliation with the gods. But, in addition to these stelae, there are also interesting parallels in PGM V, where we find two recipes for catching a thief. The first recipe (PGM V, 70–96) gives the following instructions: first paint an eye on the wall and make a hammer from a piece of wood taken from the gallows; next, strike the painted eye with the hammer while reciting the formula: ‘Hand over the thief who stole it. As long as I strike the eye with this hammer, let the eye of the thief be struck and let it swell up until it betrays him’.473 The second recipe (PGM V, 182–212) is preceded by an invocation to Hermes (which makes perfect sense, given his own experience as a thief). It recommends gathering up the usual suspects and offering them some wheat and cheese, while saying ‘Master IAO, light-bearer,

Whatever the case may be, it is worth stressing that the close temporal and geographical proximity of these British curses have allowed us to define them as a compact group (as is the case with the agonistic curses from North Africa). While this collection does use phrases and expressions found in some ‘continental’ curse tablets, they also contain an important series of idiosyncratic formulae of great interest. A fundamental feature of many of these curses is the transfer of part or all of the stolen property to the deities invoked. According to Roman law, transferring something that you no longer possessed remained possible, since a stolen object still legally belonged to the owner, even if (s)he did not physically have the item. This transference of ownership has the effect of turning a normal theft into a sacrilegious act and hence branding the thief as impious.477 This type of transference could consist of a full donation of the stolen property to the deities invoked (such as in

On the negative emotions at display in the Greek examples (which served as a socially accepted outlet), see Salvo 2012. 469  Contra Versnel 2010: 280–81 and Urbanová 2018: 25–26, n. 59, who think that the tablets were on display. Newton himself proposed that hypothesis, arguing that the curses ‘were probably suspended on walls, as they are pierced with holes at the corners’ (1863: 724, emphasis added). Nevertheless, of the known curses, only DT 2 has a hole in the middle of the upper edge. Therefore, and even if this tablet was displayed at one point, it was eventually folded or rolled up (as the rest were so) before being deposited in the sanctuary. 470  See DT 2–4, 6 and 11–12. 471  According to Versnel (1991: 72). Gager 1992: no. 89 dates them to the first centuries BCE. 472  Known as ‘confession stelae’. For a discussion, see Versnel 1991: 75ff. and, more recently, Gordon 2016. 473  PGM V, 91–95: παράδος τὸν κλέπτην τὸν κλέψαντά τι. ὅσον κρούω τὸ οὐάτιον σφύρῃ ταύτῃ, ὁ τοῦ κλέπτου ὀφθαλμὸς κρουέσθω καὶ φλεγμαινέσθω, ἄχρι οὗ αὑτὸν μηνύσῃ (translated by W.C. Greese apud Betz 19922). 468 

PGM V, 211–12: ἐὰν δέ τις αὐτῶν μὴ καταπίῃ τὸ δοθὲν αὐτῷ, αὐτός ἐστιν ὁ κλέψας (translated by W.C. Greese apud Betz 19922). 475  There are other isolated examples from Italia, Gallia, Raetia and Pannonia. 476  On the problems derived from dense concentrations of data and subsequent interpretation, see Faraone 2012. 477  Faraone et al. claim, ‘by ceding the stolen goods to the god, the curse retroactively turns a common thief into a blasphemous temple robber’ (2005: 170). 474 

61

Sylloge of  Defixiones from the Roman West 127, 237, 265, 337, etc.). Nevertheless, this is not always the case: at times, only a half of the value of a stolen object is given to the gods (cf. 205 and 443); in other cases, the amount is a third (cf. 356) or a tenth of the stolen object (cf. 349). In these occasions, which so far have only been attested in the British corpus, the deity would be paid back with a fraction of the stolen property’s value only after its recovery. In such situations, the author of the text establishes a sort of agreement with the gods in a way that is almost contractual.478

could be, at times, somewhat impatient, vehemently asking the deities invoked483 to act as soon as possible (quantocius: cf. 259 and 367), or before seven (ante dies septem: cf. 457) or nine days (ante dies novem: cf. 267, 340, etc).484 While the simplest scenario in this kind of text is handing the thief over to the deities invoked (cf. 220, 359, 451, etc.), the most common punishments include the paralysis of the wrongdoer’s vital functions485 or the repayment of the stolen object with the thief’s own blood.486 Sometimes the practitioner wants a punishment to last until the culprit returns the stolen goods to the temple,487 while at other times a defigens will ask that the curse hounds the thief for the rest of his or her life.488

Traditionally, defining such an agreement has proven quite controversial. For some scholars, it can be labelled as a sort of votum.479 Nevertheless, and as Versnel has rightly pointed out, this does not quite work, since the object is ‘ceded, not vowed’ (2010: 342–434) to the god or goddess. Even if we cannot define this agreement exactly as a votum, it remains likely that the practitioners who wrote these tablets were influenced by this concept, given its presence in the sanctuaries where the curses were deposited (Tomlin 1988a: 70 and McKie 2019: 444f.). As mentioned above before, Britannia is the only province where there is evidence for giving the deity a fraction of the total value of the stolen object. Thus, it seems appropriate to wonder whether we are dealing with a local and somewhat idiosyncratic adaptation of an existing magical-religious technology. Perhaps the practitioners were adjusting the traditional votum to address a new problem. In these texts, the authors present a unilateral agreement with the deities, which has a contractual nature. Once they have handed over the property of the stolen object to the gods, the practitioners had also passed over a sense of obligatio, which is inherent to the votum. In other words, the gods became the ones who —if they so wished— would be responsible for finding the object. At this point, the issue would be solved only when both the deity and the practitioner received their portion of the stolen object, as stipulated in the text itself. Through this mechanism, practitioners were not only encouraging divine action, but also establishing a mechanism to see whether the gods followed through on their side of the bargain, if you will.

femina is found instead or the phrase is completed with the addition of puer/puella and/or baro-mascel/mulier, like in 249 (A, ll. 3–4 and B, ll. 1–3), 364 (l. 3, where we find vel instead of si), 451 (ll. 6–8), among many others. These formulae were so popular that they could even appear in abbreviated form, like in 386 (‘sbsmspsp’= si baro si mulier si puer si puella) and 351 (ll. 3–4, ‘simausib’ =si m(ulier) au[t] si b(aro)). To these we must add the Hispanic curse from Italica 127, which reads ...si quis puel(l)a si mulier sive [ho]mo involavit (ll. 7–9). Besides the pair servus/ liber, there are others that refer to legal status, like libertinus (cf. 236, ll. 10–11), si ancilla (cf. 257, l. 7) and si ingenuus (cf. 459 l. 5). An isolated, but extremely interesting, pair is found in 303 A, ll. 1–2: ...seu gen(tili)s sẹu Ch(r)istianus; also note ...ṣi paga[n]us si mil[e]s (cf. 450, B, ll. 2–4). ̣ Finally, we ought to mention the formulae which refer to anyone who may have important information about the theft, like in 302 (A, ll. 3–4: ...qui medius fuerit) or 342 (l. 4: ...et ṣị qui ạfuerẹ), among others. 483  Cf. 215 (A, ll. 10–12: ...dea Sulis ṃạximo letum [a]digat; compare with 361, ll. 7–8); 259 (B, l. 9: ...c ̣ụm quantiocius coṇṣumas); 350 (l. 8: ...quicumque illam involasit a devo moriotur; 451 (ll. 16– 19: ...qui hoc involavit sanguem eiius consumas et decipias doṃiṇ[e] Nẹ[p]tune). 484  For this magical deadline, see Marco Simón 2010b; for magical time frames in general, see Sánchez Natalías 2019a. 485  This punishment is usually found in a non permittas formula, which can be followed by the phrase nec somnum and/or nec sanitatem (cf. 205, 237, 361, etc.). There are multiple versions of this formula, which keeps the victim from eating, drinking, talking, procreating, sitting, lying down, defecating, etc. For a few examples, see 358 (ll. 4–9 and 12–14: ... ne meiat ne cacet ne loqụaṭụr ne dormiat nẹ vigịlet nec sạḷutem nẹc sanitatem… ne co(n)scientiam de perferat); 365 (ll. 8–10: ...sanitatem ei non permittas nec iacere nec sedere nec bibere nec manducare); 443 (ll. 7–9: ...sanit[atem] nec bibere nec ma[n]d[u]care nec dormi[re] [nec nat]ọṣ sanos habẹ[a]ṇṭ). 486  E.g., 249 (A, ll. 5–7: ...sangu(in)em suum in ips˹um˺ aen˹um˺ fundat); 299 (ll. 8–9: ...facias illum ṣạṇguine suo illud satisfacere); 304 (ll. 5–6: ... ̣ sanguine et vitae suae illud redemạt); 337 (ll. 5–8: ...non redimat ni vita sanguine suo); 346 (l. 3: ...sangu(i)no suo solvat); 451 (ll. 10–11: ... sanguem eiụṣ qui conscius fueris and ll. 16–18: ...sanguem eiius consumas et decipias). 487  E.g., 205 (ll. 9–11: ...donec perferạ(t) usque templum [No]dentis); 215 (B, ll. 2–6: ...do[ne]c ̣ caracallam meam ad templum sui numinis per[t] ulerit); 443 (ll. 9–10: ...nessi hanc rem [meam] ad fanum tuum [at-] tulerint); 356 (A, ll. 6–8: ...nissi quand[o] res s(upra) dictas ad fanum s(upra) d[ic]tum attulẹrit). 488  See 250 (B, l. 4: ...quoad vixerit); 350 (ll. 3–4: ...usque die`m´ qụọ moriatur); 459 (ll. 6–7: ...diem mortịs); perhaps 257 (ll. 3–5: ...tandiu (…) quandiu hoc [ill]ud, which in Tomlin’s words ‘may have the sense of “so long as he shall live”… or “so long as he shall retrain the stolen property”’ (apud Tab. Sulis 52); 305 ([B, ll. 2–5] where the victim is asked to do something impossible: ...nisi ut Ẹuṭic ̣ia modium nebulae modium ṿeniạṭ fumi).

The culprit, who for obvious reasons is rarely named in the defixiones in fures,480 can be referred to with phrases like ‘qui involaverit’.481 Most frequently, however, the thief is referred to through an all inclusive formula comprised of opposing pairs, such as ‘si vir si femina, si servus si liber...’.482 When it comes to punishment, the practitioners Cf. Sánchez Natalías forthcoming 1. Lambert 2004. Kiernan 2004, for his part, has connected the formulae attested in these curse tablets with the ones attested in other votives. 480  Except for 259, 355, 441, 443, 456 and 457, where the name of the thief (or a suspect) is given. 481  E.g., 220 (ll. 2–4: ...qui destrale involaverit); 349 (A, l. 9 and B, l. 1: ...quiuis involavit); 451 (l. 2: ...dno (h)ominẹm qụị ld˹um˺ invọḷạṿ[it]); 454 (l. 5: ...eu(m) qui invalaverit). 482  One of the most commonly repeated is si vir si femina si servus si liber, which is found (in whole or in part) in 215 (A, ll. 7–9), 268 (l. 4), 310 (A, ll. 3–4), just to mention a few examples. Sometimes the pair vir/ 478  479 

62

8 The Pantheon of Deities Invoked 8.1. Introduction

Graeco-Roman, indigenous or eastern origins. Many of these supernatural beings are only attested once in the corpus of defixiones from the Roman West. Table 8.2 only includes those that are mentioned on more than one occasion, many of whom are attested in various provinces.

Undoubtedly defixiones provided one of the most direct and private ways to communicate with the divine in Antiquity. As a result, these texts provide a valuable source of information about various divinities and supernatural beings that are underrepresented in —if not totally absent from— other types of sources. It is only in curse tablets that we can learn about such deities and see the kinds of schemes in which practitioners sought their assistance.489

8.2. Graeco-Roman gods and supernatural entities Given the so-called process of ‘Romanization’,493 it is hardly surprising that Graeco-Roman numina appear in tablets from every province from the Roman West, though at a lesser frequency in North Africa, where a very different sort of divine being predominates.

As can be deduced from Table 8.1, only 180 of the 623 defixiones from the Roman West contain theonyms or explicit references to any gods who were tasked with enforcing the contents of a curse.490 However, we ought to add an additional 228 curses to this number which were deposited in sanctuaries or other sacred spaces: even though these texts do not mention the sanctuary’s titular divinity explicitly, it is only logical to deduce that such curses were addressed to that god.491 The same logic applies with graves and necropolis: even though 112 defixiones do not explicitly mention the Di Inferi or the Di Manes to which these spaces were consecrated, these curse tablets were likely conceived of as messages to these gods (contra Cubrera 1999a: 160). Accordingly, the number of curse tablets that mentioned or can be linked to a deity rises to 520 curses or 84 per cent of the entire corpus. Furthermore, of the remaining tablets many are damaged or fragmentary; hence, they could certainly have invoked a god. In other cases, however, it seems likely that a practitioner would have orally invoked a deity while reciting specific formulae like those preserved in certain recipes from the PGM.492

In this collection of divinities from the Graeco-Roman pantheon, there is a fairly compact and consistent group made up of chthonic deities, which are occasionally referred to with the fixed phrase Dis Inferis.494 That said, we also frequently find Pluto and Proserpina, the lords of the underworld,495 who are sometimes accompanied by their monstrous dogs.496 Within this group, special attention ought to be paid to a Roman curse tablet dated to the first century CE (cf. 48, Bevilacqua 2009 and 2010b), in which we encounter a large group of mythical beings in conjunction with the king and queen of the underworld. In addition to well-known beings (e.g., Aurora, the Sphinx, the Giants, the Furiae, the Maniae or the Chimera), we also find a series of predatory birds (aves Nocturnae, aves Harpyiae), the Ustores inferi (body burners) and the Ossifragae (the bone breakers). Leaving aside this tablet, we should also include Ceres,497 Mercury,498 the Manes,499 the monstrous Lamia500 as well as people who died prematurely among the underworld beings invoked within the curse tablets from the Roman West.501

Among the 167 curses that contain a theonym or direct references to a god, we find a wide range of deities, demigods, demons and other supernatural beings of

On this controversial term, see Beltrán Lloris 2017 with further bibliography. 494  For just several examples, see 56, 70, 105, 122 and 129. 495  Cf. 10–14, 103, 160, 526 (where they are called Iupiter Infernus and Iuno Inferna), 533 (where he is called Dito Pater), among others. Pluto is invoked as Dominus Megarus in the tablet from Salacia/Alcácer do Sal (cf. 121; for a discussion, see Marco Simón 2004: 86). Persephone/ Proserpina appears in 109 and 120 (where she is called Ataecina Turibrigensis Proserpina; see Marco Simón 2011b: 49–55) as well as the North African tablets DT 268 and (perhaps) DT 213, among others. Furthermore, the references to the regina tenebrarum in DT 288–89 could also allude to the same goddess. 496  Both Tricipites et bicipites, cf. 103 and 48. 497  Cf. the Oscan curse 68, where she is invoked as Keres Arentika. 498  Cf. 81, 258, 346 (invoked alongside Virtus), 353, 355–58, 459 and 523 (alongside Moltinus) as well as the North African curses DT 251 (called sangtus deus Mercurius infẹ[rnu]ṣ) and DT 253, among others. 499  Cf. 17, 469 (where the larvae are also mentioned), 526 (together with Pluto and Aeracura), 530 (where they are called the ministeria infernorum), DT 222, etc. 500  Cf. 82. So far, this is the only known instance of such an invocation. 501  E.g., the atélestoi found in the tablet from Lilybaeum/Marsala (109) and perhaps the biaothánatoi in DT 295 (ll. 27–28). 493 

489  For this issue, cf. Ogden 1999: 44–46; Kropp 2008: 94–103 and Abb. 3; Bailliot 2010: 83–93 and Marco Simón 2010d. For the Greek defixiones, see Kagarow 1929: 59–75. This section is an updated version of a Spanish publication (cf. Sánchez Natalías 2013b). 490  The 180 curse tablets comprise almost a third of the total. In this group, we find generic terms to refer to gods, such as deus (cf. 157, 162, 185, etc.), domina (cf. 128, DT 269, etc.) and daemon (cf. 114, DT 229, 233, etc.). 491  There are, however, some quite interesting exceptions: in the sanctuary of Sulis Minerva, for example, there is a curse that addresses Mercury (cf. 258) and others that evoke Mars (cf. 238 and 302). The latter deity is also named in several tablets from the temple of Mercury in Uley: cf. 357 (where he is mentioned alongside Mercury) and 356 (where the author wrote Mercury’s name over an original invocation of Mars-Silvanus, which may suggest that there was a mistake in the original invocation), 373 and 431. 492  For the importance of orality in the actio magica, see section I.4.1 above. For the recipes in the papyri, see PGM III, 29–40, 43–53, 71–94 and 98–161; IV, 335–406; IV 1748–812; IV, 2235–36; V, 335–43; VII, 431–32; etc.

63

Sylloge of  Defixiones from the Roman West Table 8.1. Defixiones invoking a deity according to provenance (It.= Italia, Afr.= Africa, Hisp.= Hispania, Gall.= Gallia, Brit.= Britannia, Ger.= Germania, Raet.= Raetia, Nor.= Noricum, Pan.= Pannonia) Explicit mention of a deity

No explicit mention of a deity, but deposited in a consecrated space

Total

Remaining curses

It.

32

22 (sanctuaries of Rome and Roccagloriosa) 41 (necropolis vel sim.)

95/119

24

Afr.

39

2 (other sacred spaces) 40 (necropolis vel sim.)

81/88

7

Hisp.

9

7 (necropolis vel sim.)

16/30

14

15

10 (sacred spaces) 8 (necropolis vel sim.)

33/55

22

Brit.

66

103 (Aquae Sulis/Bath, Sulis Minerva) 61 (Uley, Mercury) 1 (other sacred spaces) 1 (necropolis vel sim.)

232/256

24

Germ.

13

27 (Mogontiacum/Mainz, Attis and Magna Mater) 13 (necropolis vel sim.)

53/56

3

Raet.

3

1 (necropolis vel sim.)

4/6

2

Nor.

1

1/1

–

Pan.

2

6/9

3

Gall.

1 (Savaria/Szombathely, Iseum) 1 (other sacred spaces) 2 (necropolis vel sim.)

Table 8.2. Table containing the supernatural beings that are invoked on more than one occasion in the curses from the Roman West, classified by provenance (It.= Italia, Afr.= Africa, Hisp.= Hispania, Gall.= Gallia, Brit.= Britannia, Ger.= Germania, Raet.= Raetia, Nor.= Noricum, Pan.= Pannonia). For the supernatural beings invoked only once, see index. Supernatural beings

It.

Ablanathanalba

5

Abraxas

1 2

Angels

Afr.

Hisp.

Christ Demons/Daemones

1

8 6 1 5

Dis Pater/Pluto/Hades/Orcus

7 1

Iao

Nor.

Pan.

2

19 2

2

3 2 6

1

1 30

1

1

3 2

1 (?) 1

Jupiter

1

Magna Mater/Cybele Manes

Raet.

3

2

Di Inferi

Ger.

1

Diana/Nemesis

Hermes/Mercury

Brit.

1

Attis Cerberus

Gall.

8

1

1

2

2

Mars

1

9

Muta Tacita

1

Neptune (/Metunus)

2 1

4

Nymphs (Anna Perenna) (Niskas/Niskus)

3

Proserpina/Persephone Aeracura/Veracura

8

3 2

1

Sulis Minerva

1

1

1 21

64

3

The Pantheon of Deities Invoked 8.3. Indigenous divinities

As a result of Greek influence, the first curses to invoke chthonic divinities are Oscan and appeared in southern Italy, specifically in Petelia/Strongoli (cf. 81) and Capua (cf. 68). Dated to the fourth or third century BCE, these texts invoke Hermes and Keres Arentika (together with the supernatural beings that are members of her ‘crew’), respectively. Other early examples, dated to the third century BCE, come from Carthage and Lilybaeum/ Marsala and summon Persephone (who appears together with the Titans and atélestoi in 109).502 Beginning in the first century BCE, we can note a diversification of the gods mentioned in the corpus: Pluto, Cerberus, the Manes and the Dei Inferi all enter the record. Furthermore, beginning in the first century the geographical regions in which chthonic deities are mentioned in defixiones began to expand, until including all corners of the Roman West.

As is well known, the process of ‘Romanization’ was not simply a one-way street: in the religious context not only were certain Roman cults taken up by provincial societies, but indigenous religion was also recognized and protected under the Roman Imperial rule.511 In addition to the numerous official examples of this bidirectional process and maintenance of local practices, we must also look towards the illicit and subaltern ways in which local traditions came to influence the practice of writing defixiones. Although these tablets began as foreign technology in indigenous provincial communities, they nevertheless were taken up and used as a means to communicate with indigenous deities, who appear to have been thought of as more ‘reliable’ in the eyes of certain practitioners. The consequences of this act of adapting a Roman practice to the indigenous pantheon are far from trivial: as Marco Simón has pointed out, in some cases the mention of these deities in defixiones provides us with the only evidence for the existence of their cult (2010d: 293–97). To date, we know of examples of this phenomenon from Hispania, Gallia, Britannia, Raetia and Pannonia, which all date to the High Empire. Among the attested invocations, we can distinguish two types: interpretationes and those directed to indigenous divinities (either alone or in conjunction with other Graeco-Roman divinities).

With the group of Graeco-Roman deities found in the curse tablets, we must also mention those that are chthonic and who tend to appear in texts dating to the High Empire. Such gods include Jupiter,503 Gea,504 Diana/Nemesis505 and Mars.506 In some curses, it is clear that the gods (or other supernatural beings) are invoked since they were particularly well suited to carry out the defigens’ requests. This is certainly the case with Muta Tacita, who was summoned to silence those who may have testified against a defigens in the law court (cf. 522, 529, etc.), and Cacus, who was invoked together with Mercury and the Celtic Moltinus in order to catch a thief (cf. 523). Graeco-Roman deities presided over local cults, as we see in Neapolis/ Terralba (cf. 113), where Marsyas was worshiped or in Italica/Santiponce (cf. 127), where we find Domina Fons, who has been the subject of some debate.507 Besides Domina Fons, there are other aquatic deities, like Neptune (invoked in four British curses found near rivers)508 or the Nymphs,509 among whom the best attested is the ancient goddess Anna Perenna, in whose sanctuary there is evidence for an intriguing religious syncretism during Late Antiquity.510

To start with the former case, there are rare but interesting instances in which Graeco-Roman deities are assimilated to indigenous gods, which is the result of an interpretatio carried out by the local community. A good example of this is seen in the goddess Ataecina-Proserpina, who is mentioned in a curse from Emerita Augusta/Merida inscribed on a marble plaque (cf. 120). The indigenous theonym Ataecina has been linked to Proto-Celtic *adaki (>adaig, ‘night’), which would certainly suit a chthonic goddess.512 Another nice example of interpretatio is found in the British goddess Sulis Minerva, who was worshiped at the sanctuary in Aquae Sulis/Bath. In nearly half of the curses from this sanctuary which contain an invocation of the goddess, we find the indigenous theonym Sulis, which may reflect a local preference that endured for centuries after Roman conquest.513

502  For the Carthaginian curse, I follow the interpretation of Amadasi Guzzo 2003 who proposes that Persephone-Elat is invoked (esp. 28, n. 17 and 30). For a different point of view, see Faraone et al. 2005. 503  Securely invoked in 349 and perhaps in 168. 504  Referred to as Domina Terra in DT 220. 505  Diana/Nemesis is invoked in two British curses against thieves that were deposited in the amphitheatres of London and Isca Silurum/ Caerleon (cf. 343 and 337) as well as in another two curses from Augusta Treverorum/Trier, where she appears alongside Mars (cf. 182–83). 506  Cf. 238, 302, 452, 356 (together with Silvanus), 357 (together with Mercury), etc. 507  The editors of the defixio (Gil and Luzón 1975: 126–27) have proposed Fonsfor or Fons Fori[nae] as the reconstruction of the theonym, whereas Canto (1985: 161) has proposed Fons Fovens and Tomlin (2010: 254) has suggested Fons Font[i], which seems more convincing given the palaeographical features of the text. 508  Cf. 441, 449 and 451. 509  Cf. 92, where they are invoked with the formula uti vos Aqu`a´e ferventes siv[e] ṿọs Nimfas [si]ve quo alio nomine voltis adpe[l]lari. On this formula, which is typical of Roman priests, see the interesting remarks by Gordon 2019b: 429. 510  The nymphs are named directly in 23. A reflection of this syncretism is evidenced by the simultaneous invocation of Christ and the deas vest[ra]s found on 31.

There is only a small group of seven defixiones in which an indigenous deity is exclusively invoked. Such texts have, to date, only been found in Gallia and Britannia. The defixiones from Gaul contain three tantalizing theonyms: Niskas, Maponos and Adsagsona.514 The first Cf. Marco Simón 1996: 236–38. Cf. Marco Simón 2011b: 49–55 (esp. n. 48 on the theonym’s etymology). 513  The goddess is called Sulis in 16 tablets (213, 215, 224–26, 249–50, 254, 259, 267–68, 270–71, 274, 299 and 311), Sulis Minerva in 5 curses (237, 239-240, 251 and 265) and Minerva in one (275). Note that Sulis Minerva is the most common way of naming the goddess in the seven inscribed paterae that were deposited in the sacred spring. 514  If we accept Schwinden’s reconstruction for the curse from Arlon (cf. 204, Schwinden 1988: 29–30), we can add the theonym Cam[ulus]. 511 

512 

65

Sylloge of  Defixiones from the Roman West name is attested in three tablets that were discovered at the spring in Amélie-les-Bains in 1845 and were written in a language that has defied certain identification (cf. 150–51 and 154). The theonym Niskas, which Lizop has connected to the Basque word neskas (‘young women’), may refer to the nymphs who watched over the spring. Niskas, we will see, may even have a homologue in Britannia. With respect to Maponos, Fleuriot (1977: 178f.) has analysed the theonym as *ma(k)kwo- (‘young man’) with the suffix -ono-; he appears to be a god who was linked to Apollo. Although Maponos is attested in the Gaulish defixio from Arverni/Chamalières (cf. 163, deposited in a sanctuary of which he may have been the patron deity), this god was also worshiped in Britannia, where several intriguing inscriptions mentioning his cult have been found.515 Finally, Adsagsona is invoked in the Gaulish curse from L’Hospitalet-du-Larzac (cf. 158), which was discovered in a grave and targets a group of women (we do not know why the defigens wrote the curse). According to the text’s editors, Adsagsona could possibly be the theonym of a chthonic deity, on the grounds that the curse contains the Gaulish word autumnos (‘underworld’).516

while the second, who is the personification of the river in which the tablet was deposited, was meant to drown them. Lastly, the two remaining curses come from Raetia and contain the Celtic theonyms Ogmius and Moltinus. The first deity is found in a defixio from Brigantium/Bregenz (cf. 521) alongside Dis Pater519 and is analogous to the Graeco-Roman Hercules. Moltinus, for his part, is a Celtic god of cattle. He is invoked in the curse from Veldidena/ Wilten (cf. 523) together with Mercury and Cacus in order to catch a thief and recover the defigens’ lost property, that is, a sum of money and several steers. Given the defigens’ purposes, the invocation of Moltinus and Cacus seems apt.

The British examples, which come from Ratae Corieltauvorum/Leicester and Lydney, mention Nodens and Maglus, respectively. The first, Nodens (‘cloudmaker’), is a well-attested indigenous divinity of healing, whose sanctuary was built in Lydney Park in the third century CE.517 Maglus, for his part, is invoked as a god (deo) in 456; according to Tomlin (2008a: 208), this could be a theonym only attested here in the British Isles, which is derived from the Celtic root *maglos (‘prince’).

Cybele, the first of the oriental deities to arrive to Rome around the year 205 BCE, is often found paired with Attis in the corpus of defixiones. She is, in fact, one of the most invoked deities, as can be seen in curses from Salacia/ Alcácer do Sal (cf. 121), Groß Gerau (cf. 482), Centum Prata/Kempraten (cf. 518–19) and Mogontiacum/Mainz, where 2 of the 34 defixiones explicitly mention the pair of Cybele and Attis (cf. 490 and 494). As Gordon (2012a) has pointed out, these gods are invoked as avengers not so much because of their mythological background but rather due to their respective iconographies, which must have had a powerful impact on practitioners. Thus, the representations of Cybele in her chariot drawn by lions and of Attis with his Phrygian cap surrounded by stars were highly evocative. In the case of Magna Mater, her iconography was thought to represent her untiring ability to hunt down a curse’s victims; in Attis’ case, his cap was taken as a reference to his extraordinarily effective cosmic powers.

8.4. Oriental powers The third group of numina that are invoked in the defixiones consists of oriental deities like Cybele, Attis, Isis, Christ and the demon Abraxas (among others).520 Such eastern cults began to arrive to Rome at the end of the third century BCE. The spread of these cults is well documented in a conspicuous group of curse tablets from all over the Roman West, which mostly date to the Imperial period.

Finally, we must turn to the four cases in which there is a joint invocation of indigenous and Graeco-Roman gods who are not connected through an interpretatio. So far, these instances hail from Britannia, Raetia and Pannonia.518 The curse from Hamble (Britannia, cf. 451), which was discovered in an estuary, invokes Neptune and Niskus. The latter appears to be an aquatic deity that has been connected to the Niskas mentioned in the curses from Amélie-les-Bains (cf. 150–51 and 154). Unfortunately, so far there is no further evidence that has shed more light on this tantalizing parallel.

In contrast to Cybele and Attis, the rest of the abovementioned oriental deities appear with much lesser frequency in the corpus of defixiones from the Roman West. Of these deities, the Egyptian Isis enjoyed a flourishing cult in Hispania during the High Empire. Nevertheless, there is only a single curse tablet from her sanctuary in Baelo Claudia/Bolonia (cf. 128), in which the goddess is invoked with the epithet muromem.521

The curse from Siscia/Sisak (Pannonia, cf. 529), which was found near a river, is a juridical defixio. The curse invokes the Roman goddess Muta Tacita together with the indigenous Savus. This was, no doubt, an effective pairing, since the first was meant to silence the victims, Cf. Green 1997, s.v. ‘Maponos’. For a concise discussion, cf. Lambert apud RIG II.2: L-98, p. 266. 517  For an overview of this deity, see Green 1997, s.v. ‘Nodens’. On certain occasions, Nodens is linked to Mars and Silvanus. 518  In the Spanish version of this section (cf. Sánchez Natalías 2013b), I mistakenly added Hispania to this list, assuming that Salpina was a deity invoked in a curse tablet from Cordoba (cf. 122). This had always been assumed since the editio princeps of the text (for the latest instance of this assumption, see Marco Simón 2013, with previous references). Nevertheless, as I realized shortly after, Salpina ought to be taken as the personal name of the curse’s victim (cf. 122 and Sánchez Natalías 2014). 515 

Following the reconstruction proposed by Egger 1943. In the Spanish version of this section (cf. Sánchez Natalías 2013b), I wrongly added the Egyptian Seth, who was supposedly invoked in 24, a container from the sanctuary of Anna Perenna in Rome. However, a reexamination of this short text (Sánchez Natalías 2020b) led to a stronger interpretation of the curse, in which the victims’ name (and not the deities’) can be read (cf. 24). 521  Which would be for muronem, derived from the epithet Myronyma (for a discussion, see Bonneville et al. 1988: 22, n. 4). In addition to the curse from Baelo Claudia/Bolonia, we must add the one found in the

516 

519  520 

66

The Pantheon of Deities Invoked In addition to Isis, we also find Abraxas, the cock-headed anguipede whose name (equivalent to the number 365) evoked cosmic totality. The earliest invocation to this deity comes in a second-century CE curse tablet from Autun (cf. 166, Gallia Lugdunensis). The text, written in a mixture of Greek and Latin, contains a list of names together with a series of voces magicae and the name of Abraxas between two charaktêres. This curse’s sophistication strongly suggests that it was written by a professional practitioner. This is also the case with a Late-Antique curse tablet from the sanctuary of Anna Perenna in Rome (cf. 30), where Abraxas is invoked as a mediator between the practitioner and the deity who was supposed to harm the victim. In this same fountain, a series of containers have also been thought to contain a representation of Abraxas.522 Nevertheless, and as argued above (cf. section I.4.2.2.1), this identification ought to be revised, given the absence of some characteristic iconographic features of Abraxas and the presence (on his abdomen) of the abbreviated Christian formula Ἰησοῦς Χριστός Ναζωοραῖος, Ἰησοῦς Χριστός Ναζωοραῖος, καὶ Θεός, Θεός, Θεός (cf. Németh 2012b and 2015). This formula, attested with slight variations on several containers (cf. 25–27 and 35–36), is quite plausibly a way to identify or ‘name’ the figure, as happens in other instances (cf. Baitmo Arbitto, Figure 4.7). In other words, I think this could be a new and (doubtlessly surprising) representation of Christ. Whatever the case may be, Christ also seems to be invoked in a different curse tablet from this same findspot (cf. 31, ll. 4–5), whose text reads: deas vest[ra]s (...) et Christum nostrum. Although the reading has proven difficult to decipher (see the different versions published by the editor) and I have been unable to confirm it from the published drawings, the sentence would point towards two different (but related) religious spheres: your (pagan) deities and our Christian ones. Therefore, whatever practitioners were active at the Fountain of Anna Perenna identify themselves as Christians, a characteristic which would certainly job with some other aspects of the artefacts that they produced.523

Among the most frequent invocations to demons, we find ‘adiuro te demon’524 and ‘demon quicumque est’,525 which often appear alongside the formulae ‘adiuro te demon quicumque est te demando’.526 Through these invocations, the practitioners directly addressed the deceased who occupied the spaces which would be used to deposit curses, either in the necropolis or in what might be the spoliarium of Carthage’s amphitheatre.527 In both of these contexts, we find examples of tablets in which the origin of these powers is linked to the underworld, which is hardly surprising.528 However, in two Carthaginian curses, the texts specify that those demons who ought to execute the curse came from Egypt-Campania and Hispania-Africa, perhaps in an attempt to imbue the demons invoked with greater ‘exoticism’ (and power).529 In an attempt to encourage the demons to act, these professionals could name them alongside certain divinities, such as deus pelagicus (‘the god of the sea’), who is invoked in the well-known Baitmo Arbitto series. 530 Such a choice makes perfect sense, given that Baitmo is represented on top of a boat (cf. section I.4.2.2.1, Figure 4.7). In addition to the deus pelagicus, some tablets also invoke a deus vivus omnipotens (which could well be an early reference to Christianity)531 or even the Jewish god Ιαω Σαβαωθ.532

Last, but certainly not least, we must mention the North African curses from the High Empire, where there is a marked preference for invoking demons rather than deities. This tendency could be explained by the fact that the texts were mostly written by professional magical practitioners, who were trained (or at least influenced) by Graeco-Egyptian magical practices. In this magical tradition, such supernatural beings played a fundamental role, as is clearly attested in the PGM.

Cf. DT 248 (A, ll. 1–2), 250 (A, ll. 27–29), 290 (B, l. 1), etc. A variation on this formula is found in DT 233 (ll. 27–29): ...excito [t]e demon qui ic conversans. 525  Cf. DT 265 (B, ll. 5–7). 526  Cf. DT 286 (B, ll. 1–2), 291 (A, ll. 3–7 and B, ll. 1–2), Audollent 1910 (B, ll. 1–2), etc. 527  Cf. DT 251, col. I, l. 4, where the practitioner invokes the anim(a)e (h)uius loci (for this context, see section I. 6.5.3). 528  Cf. DT 250 (A, ll. 28–29), 266 (ll. 2–3) and 295 (l. 11). 529  Specifically, cf. DT 230 (an erotic curse found in a grave and dated to the second or third century CE, in which Καταξιν, Νοχθιριφ, Τραβαξιαν, Βιβιριξι and Ρικουπιθ are invoked) and DT 250 (an agonistic curse from the amphitheatre dated to the fourth century CE, in which the demons Βαχαχυχ, Ιεκρι, Νοκτουκιτ, Βυτυβαχκ as well as the deus omnipotens Παρπαξιν are invoked). 530  Cf. DT 286, 290–91, 293 and Audollent 1910. 531  Cf. DT 247, ll. 18-19, DT 248 (A, ll. 2–3, where we find mention of deum vivum) and 268, l. 4 (where Persefina, for Persefona, is invoked together with de[um] me[um] vivum). 532  Cf. DT 286, 291, 293 and Audollent 1910. Ιαω is invoked on his own in DT 264. If we accept Corell’s far from certain reading, this deity may also be attested in 143. 524 

sanctuary of Isis in Savaria/Szombathely (cf. 531); although there is no written invocation in the curse, it seems probable that the goddess was invoked orally during its deposition. 522  See section I.4.2.2.1 and n. 175. For his relationship with Anna Perenna, see Piranomonte and Marco Simón 2010: 11–13. 523  To the already mentioned invocation of Christ on the containers (cf. 25–27 and 35–36), we could also add the suggestive analogy between light and life pointed by Mastrocinque 2007 for the lamps whose wick was made of a curse tablet (cf. 19–20, 30–31).

67

9 Distribution of Defixiones in the Roman West533 9.1. Introduction

In an attempt to depict the diffusion of this cultural practice throughout the Roman West, this section provides a geographical analysis of this cultural phenomenon.538 With this purpose in mind, I rely on a series of maps which respect the provincial division and the limes from the High Empire.539 Obviously, such divisions did not have the same importance or relevance in earlier historical periods. Nevertheless, I have decided to use them as a means of better organizing and guiding the following discussion. The corpus of curses has been divided into three different periods, each of which has its own map. These respect standard periodizations used in Roman History (Republic, High Empire and Late Antiquity).540 In addition, there is a fourth map that contains an overall picture of the phenomenon across all periods. If there is only a single item from a site, the map includes the curse’s number in the following sylloge. For sites that have yielded various tablets, it was often impossible to include all catalogue numbers; therefore, in those cases the map simply contains the site’s name. More information about the number of curses from these sites can be found in the following pages and in the sylloge proper.

In Antiquity, curse tablets were a transferable magicoreligious technology. As such, and together with other cultural innovations from the Graeco-Roman world, they travelled throughout the area that would come to be ‘the Roman West’. Reconstructing exactly how this process unfolded has proven a rather difficult question, since the geographical diffusion and cultural adaptation of cursing rituals not only depend on individuals (both the professional ‘magicians’, whom Plato already mentioned, and individual practitioners who possessed enough knowhow to make curse tablets),534 but also on larger, more complex phenomena (e.g., Greek colonization or what scholars have traditionally called ‘Romanization’).535 To date, the oldest curse tablets from any part of the Mediterranean are the Greek texts from Sicily (largely from Morgantina and Selinunte). Leaving aside the thorny question of whether writing curse tablets was a Sicilian innovation or whether the practice came from the metropolis,536 we say with certainty that the permeable practice of writing curses spread along Greek colonization routes to end up reaching other areas of the Mediterranean basin (such as the Italian and the Iberian peninsulas, and perhaps even to North Africa) and the Black Sea. Once taken up by the local communities, curse tablets were adopted and adapted to better serve the specific idiosyncrasies of each community. In certain cases, it was these very local populations that transmitted this magicalreligious technology to the Romans.537 Once entrenched within Roman social and religious culture, the habit of writing curse tablets slowly continued to be disseminated via the Roman army. In fact, military conquest and the establishment of new trade and communication routes unavoidably gave rise to new intercultural contact between various populations. The practice of writing curse tablets spread alongside other technological and cultural innovations between these communities. Here it is worth stressing how even a legally sanctioned practice like writing curses could spread slowly but relentlessly through areas that were in contact with Rome. Indeed, this is one of the more illicit and surreptitious facets of ‘Romanization’.

In the following pages, the dates used for curses are based on three different factors: the archaeological context, palaeography and linguistic features of each tablet. Given that defixiones were often ‘intrusively’ introduced in a particular site after other artefacts had already been deposited, archaeological context can often only provide a terminus post quem. Therefore, the ‘internal characteristics’ of each text (i.e., palaeography,541 linguistics or, rarely, prosopography)542 can provide a more concrete date, though it must be stressed that there has not always been scholarly consensus of the dating Greek curse tablets have not been included in this analysis. For their spread throughout the Roman West, see Sánchez Natalías 2019c. 539  It must be stressed that the maps for Late Antiquity as well as the global one use provincial divisions from the reign of Diocletian with the Mauritaniae (Tingitana, Caesarense and Tabia), the Numidiae (Cirtensis and Militiana), Proconsularis Zeugitana and Byzacena. 540  This is irrespective of the problems related to standard periodizations, as rightly pointed out by Flower (2008: 5–9). 541  For an overarching analysis of the palaeography of the Latin defixiones, see Bartoletti 1990; additionally, see the studies of corpora from Aquae Sulis/Bath (cf. Tomlin 1988a: 84–94), Mogontiacum/Mainz (cf. Blänsdorf 2012a: 41–47) as well as the so-called ‘Sethian’ defixiones from Rome (cf. Wünsch 1898: 53–56). There are other important studies of smaller groups of defixiones, such as 10–14 (cf. Fox 1912a: 51–55, plate VIII), or even single tablets, such as 16 (cf. Muzzioli 1939: 47f., Tav. II) or 132 (cf. García Dils de la Vega et al. 2013: 247–48), among others. 542  The appearance of specific formulae as well as the (very rare) mention of well-known victims allow us to determine a curse’s date more securely; a small group of curses from Emporion (cf. 134–36), in which the local authorities are cursed, provide the best example of this. For the use of formulae to date texts from the sanctuary of Sulis Minerva, see Tomlin 1988a: 73f.; for further references, see section I.4.3. 538 

For a preliminary version of this section, see Sánchez Natalías 2012c. Urbanová 2018: 398–412 provides a series of maps whose aim is to underscore the differences and similarities between curses and ‘prayer for justice’ and their spread in the Roman West (both chronologically and in the case of curses, according also to content). 534  On this, see section I.2. 535  On the controversial term of ‘Romanization’, see Beltrán Lloris 2017. 536  Cf. Curbera 1999a: 158f. and Vitellozzi 2019: 342. 537  See section I.9.2. 533 

69

Sylloge of  Defixiones from the Roman West

Figure 9.1. Defixiones dating to the Republican period.

of certain palaeographical features. Given the scarce evidence and unavoidable role of (at times subjective) scholarly judgement, in many cases proposed dates fall within a rather wide range that cannot always be limited to a single century (e.g., ‘from the High Empire between the first and third centuries CE’). The limitations concerning what we can say in good faith about a defixio’s date, of course, greatly complicate the present task.

preserved in the archaeological record, since, as the verb incantassit implies, these appear to have been oral curses. In addition, two additional laws from Tablet VIII further suggest that magical-religious practices primarily took the form of oral speech acts in the fifth-century Rome. The first one reads, si quis occentavisset sive carmen condidisset, quod infamiam faceret flagitiumue alteri (‘if anyone had “sung in enmity”, that is, composed a song which would bring infamy or disgrace on someone else’) and alludes to the slandering of a rival.545 The second law reads qui fruges excantassit (‘whoever has bewitched fruits’)546 and alludes to a magical damage to the harvest and hence by extension to the owner’s wealth. In short, the lexemes employed in the law (carmen, incantassit, occentavisset and excantassit) stress orality. Importantly, this contrasts with contemporary Greek tablets discovered in Sicily in which the cursing verbs refer to writing (γράφω and its compounds: καταγράφω, ἐνγράφω, ἐγκαταγράφω) and not speaking, as Poccetti has shown (2005 and 2015: 378–82).547

Finally, curse tablets with an unknown provenance have not been included in the maps, though they are included in the overall numbers used throughout this section.543 In addition, all of the defixiones that cannot be dated have been included in a final map which seeks to give a full picture of the spread and distribution of cursing practices throughout the Roman West. 9.2. Distribution of defixiones during the Republic The oldest evidence for cursing in Latin comes from Rome’s first legal code, the XII Tables, which were written in the mid-fifth century BCE. Tablet VIII threatens to punish qui malum carmen incantassit, that is, ‘whoever casts a magic spell’.544 This legal prohibition undoubtedly provides evidence for cursing practices that have not been

If we turn to the archaeological record, the earliest curses were written in Greek and were found in Sicily, though, as said before, these curses will not be discussed here. The map clearly shows that the first non-Greek curse tablets were Oscan and found on the Italian Peninsula, dating to the fourth century BCE (see Figure 9.1). These artefacts, which must be understood as arising from the

These curses hail from Italy (cf. 16 [all we know is that it was donated to the Museo Nazionale Romano in the 1930s (see Muzzioli 1939: 42)], 112 and 116–19), Hispania (cf. 149), Britannia (cf. 459–60) and Pannonia (cf. 535). 544  XII Tables VIII, 1a. Translation by Crawford 1996. On this (and the two following references), see Rives 2002 and 2003. 543 

XII Tables VIII, 1c. Translation by Crawford 1996. XII Tables VIII, 4. Translation by Crawford 1996. 547  Cf. also Chiarini 2019. 545  546 

70

Distribution of Defixiones in the Roman West cultural interaction between Greeks and Oscans, hail from the sites of Roccagloriosa (cf. 76), Castiglione di Paludi (cf. 77), Crimisa/Cirò Marina (cf. 75) and Tiriolo (cf. 74). To this list of tablets, we must add a curse tablet from Capua (cf. 68), three from Laos/Marcellina (cf. 78–80) and another from Petelia/Strongoli (cf. 81), all of which date to the fourth or third century BCE. Interestingly, all these settlements are located near the sea (either the Ionian or Tyrrhenian), which meant that they were easily accessible to Greek traders. In fact, Eubeans, Phocaeans, Aecheans and Dorians all turned the area stretching from the Gulf of Naples to Calabria into a zone of Greek influence and colonization (Poccetti 2015: 384f.).

justice’, a genre that has its origins in the Near East,550 Amadasi has claimed that the tablet actually betrays Hellenic influences on the grounds that there is a probable invocation of Persephone and the presence of formulae that are paralleled in the Attic corpus of defixiones.551 The latter hypothesis strikes me as the more probable, given that Amadasi’s argument reflects the spreading of curse tablets from Sicily to other territories, such as the Iberian or the Italian peninsulae, mapping onto larger patterns of commerce and colonization. If we return to Italy, the oldest lead curse tablet written in Latin dates to the second century BCE and does not come from Rome, but rather Pompeii (cf. 71). Accordingly, this find spot brings us back to the gulf of Naples, a zone that was obviously under Greek influence. Due to the well attested use of defixiones and the similarities between certain formulae, Poccetti has argued that the early Latinlanguage curses should be closely linked to the Oscan world.552 During the second century BCE, however, we begin to find new Latin curses further north in Ostia (cf. 52) and Caere/Cerveteri (cf. 95). Both of these curses were found in a necropolis, and they share the same structure: a simple list of the victims’ names.553 To these Latin texts we should add three Etruscan curse tablets from Volaterrae/ Volterra (cf. 88–90), which were found at the entrance of a hypogeum. According to Massarelli (2014: 185 and 195, n. 5), the texts were written in the type II alphabet from Maggiani’s typology, which, together with certain prosopographical features, allows us to date the texts in the second century BCE.

Thus, the concentration of curses in the above-mentioned settlements must be understood in light of their deep connections to the Hellenic world: Greek colonization in Magna Graecia left its imprint on magical practices throughout southern Italy.548 Indeed, this imprint is especially discernible in a curse from Petelia/Strongoli (cf. 81) which contains a rather peculiar mix of Greek and Oscan. The text, which was written in both languages, contains a list of names in which the nominative forms reflect Oscan morphology, while the genitive forms use Doric Greek endings. In addition, the cursing formula placed at the end of the text was written in both languages (although it is not a translation sensu stricto). By the third century BCE, at least four new curses can be added to the earliest defixiones from the Italian Peninsula. In fact, we not only find curses in a broader geographic area, but also in new cultural contexts, such as Samnium, Sicily (where, of course, there were already Greek curse tablets), Etruria and North Africa. More specifically, these thirdcentury tablets include an Oscan tablet from Aquilonia/ Monte Vairano (cf. 84), a curse written in both Latin and Greek from Lilybaeum/Marsala (cf. 109), an Etruscan text from Aquae Populoniae/Campiglia Marittima (cf. 94) and two Phoenician curses from Carthage. The Phoenician curses were discovered in the necropoleis of Bir ez-Zeitun and Douïmes (cf. DT 213–14, both found at the end of the nineteenth century). Unfortunately, the poor condition of the first tablet has prevented scholars from offering a reading or a date of the text.549 The second curse, however, is in much better shape and has given way to a stimulating discussion about the origin of cursing rituals in North Africa. Indeed, recent studies have proffered a series of competing theories: while Faraone and other scholars have argued that this tablet is an instance of a ‘prayer for

Together with these examples of rather abbreviated and curt curses, there are eight known defixiones from Rome that date to the first century BCE (cf. 1–3 and 10–14). These texts differ markedly from the earlier Latin curses: not only are the texts longer, but they also contain complex formulae and invoke a different set of deities. A good example of this is found in the famous five curses that are now housed at Johns Hopkins University (cf. 10–14): written, in all likelihood, by a single practitioner, these texts share extremely similar formulae and all invoke Proserpina, Pluto and Cerberus. But Rome is not the only site that evidences an increased use of defixiones: also from Latium, there is a set of three curses that were found Cf. Faraone, Garnand and López-Ruiz 2005: 169 and 180ff. Amadasi Guzzo 2003: 25–31, esp. 31, where the author concludes ‘il formulario di questa tavoletta si discosta dalle maledizioni note in ambito semitico di nord-ovest, per avvicinarsi invece ai formulari greci. L’iscrizione è stata messa in rapporto con le tabelle attiche (…) Se davvero la dea invocata è Persefone, il cui culto è stato introdotto a Cartagine dalla Sicilia, si potrebbe proporre che anche l’uso di deporre scongiuri in piombo nelle tombe (...) sia entrato a Cartagine —come il culto di Demetra e Kore— dal territorio siciliano di cultura greca’. 552  Cf. Poccetti 1993d: 73–96, esp. p. 80; Poccetti 1999: 545–61, esp. p. 555, where he stresses, ‘La diffusione della scrittura di maledizioni su lamine plumbee nel mondo romano non avviene prima della fine dell’età repubblicana e probabilmente per mediazione osca: non casualmente le prime tabellae defixionis latine appaiono in ambienti di stretto rapporto con gli italici (Campania e Delo) e riportano formulari comuni a quelle osche.’ On the earliest Latin curses from the Italian peninsula, see Bevilacqua 2017 and Vitellozzi 2019: 354–56. 553  For a discussion, see section I.4.3. 550  551 

548  On this, see Poccetti 2015: 383–86 (with previous references) and more recently Vitellozzi 2019. Both authors explore the similarities between the Oscan and the Greek texts. In addition to these, see also McDonald (2015: 137–49), who elucidated the influence of Greek speakers from Sicily (and abroad) on the Oscan texts. For an in-depth analysis of the Oscan texts, see Murano 2013. 549  Cf. DT 214. On this tablet, Clermont-Ganneau has claimed (apud DT 214): ‘Le P. Delattre a aussi découvert récemment, à Carthage, une tabella devotionis avec inscription phénicienne; mais elle est dans un tel état de mutilation que l’étude n’en a pas encore été possible, les fragments s’effritant au moindre contact.’

71

Sylloge of  Defixiones from the Roman West within a cinerary urn in Nomentum/Mentana (cf. 49–51); further south in Campania, there is also a collection from Cumae (cf. 62–65), which were written either in Latin or Oscan and are largely judicial in nature. To this list of first century texts, we should also add another tablet from Capua (cf. 66). Finally, further north in Ateste/Este (cf. 103) another Latin-language curse containing an invocation to Proserpina, Pluto, Cerberus and his twoheaded brother Orthus has been found in a necropolis (probably a judicial defixio).

remains of a necropolis at Antium/Anzio (cf. 59), which merely contains the victim’s name. By way of conclusion, and irrespective of the questions about where the practice of writing curse tablets first originated (i.e., probably in Sicily, as the archaeological record has demonstrated so far), it seems clear that the spread of the practice of inscribing curses is best explained as the result Greek cultural influence. In an initial stage, this practice clustered in southern Italy amongst the Oscan settlements and subsequently spread to other Latin and Etruscan communities in the Peninsula. In the second and first centuries BCE, the practice of writing defixiones spread beyond the Italian Peninsula, reaching not only the principal cities of the Iberian Peninsula but also Carthage in North Africa.

During the first century BCE, we also see that the practice of writing curses spread outside Italy. In the Iberian Peninsula, important cities like Emporion, Carmo/Carmona and Corduba/Cordoba have turned up important evidence.554 In ancient Emporion, which was founded by colonists from Phocaea around 600 BCE and served as Rome’s entrance to the peninsula (Gn. Scipio disembarked there in 218 BCE and began his conquest), we find the largest collection of defixiones from Hispania: of the eight curse tablets from this site,555 one dates from the first century BCE (cf. 137). In the Republican period, ancient Carmo was one of the principal cities of Ulterior Baetica, of which Corduba became the capital when the province was created. There is one tablet from the former site (cf. 129) and two from the latter (cf. 125–26). Like the tablet from Emporion, these two are juridical in nature. To these texts, we should also add the one from El Portal (cf. 133), discovered under unknown circumstances and the text of which consists of a list of names. Although the curse’s editor has dated the tablet between the second half of the second century BCE and the beginning of the first century BCE (which would make of this defixio the oldest Latin curse tablet from the Iberian Peninsula), the palaeographical features of the text (similar to 129 or 131) more likely point to a first-century BCE date.

9.3. Defixiones from the High Empire As part of the hidden side of the ‘Roman epigraphic habit’,557 the practice of inscribing curse tablets exploded during the High Empire. The numbers speak for themselves: of the whole corpus, more than a half of known curse tablets have been dated between the first and third centuries CE. Even if curse tablets have nothing to do with the monumentalization of writing or with the growing interests in self-representation, the proliferation of these ‘testi privatissimi’558 provides a window onto practitioners’ innermost worries and concerns. The processes of ‘Romanization’ coupled with the greater diffusion of writing led to a sharp rise in creation of curses. In fact, and as the following map demonstrates, this tendency is attested in all of the empire’s provinces (see Figure 9.2). In the first century CE, we see two main nuclei for the production of defixiones: Italy and North Africa. In the Italian Peninsula, the largest concentration of texts comes from the areas located between Picenum and Bruttium et Lucania: two-thirds of the known curses were found between Urbs Salvia/Urbisaglia (cf. 86) and Copia Thurii/ Sybaris (cf. 82).559 Further north, there is a tablet from Altinum/Altino (cf. 104) and another from Verona (cf. 107), while outside of the peninsula other texts were found on Sardinia (Orosei, cf. 110–11) and Corsica (Mariana/ Lucciana, cf. 115). If we turn to North Africa, Carthage emerges as the focal point with over 30 tablets,560 though there is also a curse from Hadrumetum (DT 304). Turning to the Iberian peninsula, in Hispania Citerior there are four tablets from Emporion (cf. 134–36 and 138) and another

Of the seven curses dated to either to first century BCE or CE, six of them come from Hispania.556 Three of these texts were found in Corduba/Cordoba (cf. 122–24) and one in Emporion (cf. 139). The Corduba/Cordoba curses, all of which were found in the same necropolis, were written for unknown purposes (two of them consists of mere lists of names [cf. 123–24] while the third [cf. 122] consigns its victim to the underworld deities). Another tablet comes from Celti/Peñaflor (cf. 131) and seeks to silence its victim. The sixth curse from Hispania is a unique bilingual lead tablet written in Greek and Latin that was deposited in an abandoned Iberian settlement near Barchín del Hoyo (cf. 145). Finally, from Italy there is a text discovered in the

On this concept, coined by MacMullen (1982), see Beltrán Lloris 2014b. 558  Cf. Palumbo in Bettarini 2005: VI. 559  Here I am referring to the curses from Rome (cf. 17 and 48), Antium/Anzio (cf. 60), Cumae (cf. 61 and 64), Cales/Calvi Risorta (cf. 70), Marsi Marruvium/San Benedetto (cf. 83) and Septempeda/ San Severino (cf. 85). Also add two tablets from loca incerta (cf. 116 and 119). 560  For all African defixiones, I have followed the date given by each curse’s editor. It is worth stressing that there has been debate over the dates and I have not been able to examine all the texts in person. See DT 215–30; 232; 233; 243; 244; 247–51; 253; 254; 258–61; 303. 557 

See Marco Simón 2019 for a discussion of the spread of curse tablets in the Iberian Peninsula. The author examines not only the first Greek and Latin texts, but also a dozen Iberian texts that (despite problems with the language) could be identified as defixiones. 555  Two of the eight tablets are in Greek (dated to the end of the fourth or beginning of the third century BCE), while the other six are in Latin (cf. 134–39). To these we must add two new defixiones still under study (see the notice for Emporion in the introduction to Hispania, cf. section II.2). 556  Although these tablets could be included in either the Republican or High Empire map, I have chosen the former to make the maps more legible. 554 

72

Distribution of Defixiones in the Roman West

Figure 9.2. Distribution of defixiones during the High Empire.

from Saguntum/Sagunto (cf. 144).561 In Hispania Baetica, there is an interesting curse from Salacia/Alcácer do Sal (cf. 121) which was written in response to a theft, and also an agonistic curse tablet from Astigi/Écija (cf. 132).

Chamalières and Autricum/Chartres, cf. 157, 163 and 172 respectively), Britannia (London, cf. 338) and Raetia (Brigantium/Bregenz, cf. 520, Cambodunum/Kempten, cf. 522 and Veldidena/Wilten, cf. 523). The majority of the tablets from this period should be classified as juridical.

In the same century, however, the practice of inscribing curses also moved north to Germania (Colonia/Cologne, cf. 464; Cruciniacum/Bad Kreuznach, cf. 466–70, 473–74, 476 and Groß Gerau, cf. 483), Gallia (Montfo, Arverni/

Between the first and second centuries CE, Germania Superior stands out due to the impressive collection discovered at the sanctuary of Isis and Magna Mater (Mogontiacum/Mainz), where 34 defixiones dating to this period were found (cf. 487–517). In connection with the Mainz collection, we should point again out Cruciniacum/

To which we should add a defixio that was not inscribed (cf. the notice for Saguntum in the introduction to Hispania, section II.2).

561 

73

Sylloge of  Defixiones from the Roman West Bad Kreuznach (cf. 471 and 475), Groß Gerau (cf. 482), Bodegraven (cf. 463) and Waldmössingen (cf. 479), where curses with a similar chronology have been discovered. If we turn to Britannia, there are four tablets that date to the first and second centuries CE that were discovered in London (cf. 342), Isca Silurum/Caerleon (cf. 337), Wanborough (cf. 352) and Uley (cf. 369). The last of these British texts deserves special mention, since it is the earliest representative of what would become the dominant trend in the province: depositing curses in sanctuaries. Finally, within this period, we have three tablets from Italy (from Tibur/Tivoli, cf. 55; Pompeii, cf. 72 and Cremona, cf. 105), five from Hispania (Aratispi/ Villanueva del Cauche, cf. 130; Emporion, cf. 138 and Saguntum/Sagunto, cf. 140, 142–43),562 two Gaulish examples from the Gallia (L’Hospitalet-du-Larzac, cf. 158 and Les Martres-deVeyre, cf. 161), one from Raetia (Brigantium/Bregenz, cf. 521) and three from Pannonia (Emona/Ljubjlana, cf. 528; Sisak, cf. 529 and Savaria/Szombathely, cf. 531).

these sanctuaries and other isolated discoveries, Britannia quantitatively stands out as the most prolific province from the end of the second century through the third. Outside of Britain, the most noteworthy epicentre for the production of curses is found in Hadrumetum,571 with 17 examples. The other North African sites that have yielded curses in this period are Thysdrus/El Jem,572 Carthage573 and Cirta/ Constantine.574 Most of these curses are agonistic or erotic, and in contrast with the ones from Britain, were written by professional practitioners. In the rest of the Roman West during this period, there are only a few isolated defixiones with no discernible common characteristics. There are 5 tablets from Italy575 as well as another 12 spread between the Gallia (Durocortorum/Reims and Wolberg, cf. 174 and 204 respectively), Germania (Aventicum/ Avenches, cf. 481 and Centum Prata/Kempraten, cf. 520), Raetia (Peiting, cf. 524 and Abusina/Eining, cf. 525), Noricum (Faviana/Mautern, cf. 526) and Pannonia (from Carnuntum/Petronell, Aquincum/Budapest and locus incertus, cf. 530, 532–34 and 535 respectively).

In the second century, we see an uptick in number of defixiones found in the provinces bordering the Mediterranean, especially in Hadrumetum,563 the site from which nearly half of the known second-century CE curses come. Furthermore, the manufacture of curses is widely attested in Italy, from Regio X (Venetia et Histria) to Campania.564 In Hispania, the four tablets dated to the second century share two main characteristics: they are all juridical and all hail from the southern half of the peninsula.565 In addition, there are sporadic finds in the Gallia,566 Germania,567 Britannia568 and Pannonia Superior,569 most of which can also be classified as juridical. Dating to the end of the second century through to the third century, there are two large caches of curses from Britannia, one from the sanctuary of Sulis Minerva (Aquae Sulis/Bath) and the other from the sanctuary of Mercury (Uley).570 Between these two sites, there are over a hundred tablets dated to this period, the majority of which are defixiones in fures. So far, between the curses from

By way of conclusion, we can say that during the High Empire the increased production of curse tablets provides evidence for both a ‘subaltern’ facet of the ‘Roman epigraphic habit’ as well as demonstrating the geographical diffusion of this cultural practice through the process of ‘Romanization’. Key factors in this phenomenon include the diffusion of literacy and the spread of basic knowledge about how to write a curse. To understand the extent to which locals took up and actively adopted the technology of cursing, we need look no further than onomastics, which reflects local epigraphy and provides substantial evidence for indigenous deities.576 Even though this cultural practice expanded to every corner of the Roman world, we must acknowledge that five nuclei loom large in the existing evidence. Thus, we have the large collections from the sanctuaries in Aquae Sulis/Bath, Uley and Mogontiacum/ Mainz, where the practitioners wrote their own texts and seem to have been especially preoccupied with matters related to justice. In contrast, the collections from Carthage

To these, we should add a curse of unknown provenance, cf. 149. There are 18 defixiones from Hadrumetum: DT 266, 267, 270, 272– 84; AE, 1907, 68–69. 564  I am referring to the curses from Pola/Pula (cf. 101–02), Verona (cf. 108), Arretium/Arezzo (cf. 92), Ostia (cf. 53), Rome (cf. 15), Minturnae/ Minturno (cf. 56), Salernum/Salerno (cf. 69) and Aquinum/Aquino (cf. 73) which adds up to a total of nine tablets. 565  Cf. Emerita Augusta/Merida (cf. 120), Italica/Santiponce (cf. 127), Baelo Claudia/Bolonia (cf. 128) and Saguntum (cf. 141). 566  This is the case for the tablets from Mediolanum Santonum/ Villepouge-Chagnon (cf. 160), Autun (cf. 166), Mediolanum/Evreux (cf. 167–70) and Maar (cf. 173). 567  Frankfurt (cf. 478), Roßdorf (cf. 480) and Centum Prata/Kempraten (cf. 519). 568  Specifically, a curse from London (cf. 339), three from Aquae Sulis/ Bath (cf. 206, 215 and 235) and three from Uley (cf. 357–58 and 361). 569  Cf. 527, from Poetovio/Ptuj. 570  There are 52 defixiones from Aquae Sulis/Bath (213–14, 233–34, 236–68 and 272–86), while there are 56 from Uley (cf. 354–55, 359, 364–67, 370, 372–73 and 386–431). To these, we must add the curses from Dodford (cf. 453), London (cf. 343), Chesterton-on-Fosse (cf. 336), Ratae Corieltauvorum/Leicester (cf. 456–57), Puckeridge-Braughing (cf. 448) and two items from loca incerta (cf. 459–60). Two more curses, from Ratcliffe-on-Soar (cf. 349) and Pagans Hill (cf. 443), date to the third century.

There are 18 curses from Hadrumetum; for a discussion, see DT 263– 65, 268, 269, 286–95, 297, Audollent 1910: 142–48 and Audollent 1930. 572  See Foucher 2000. 573  There are seven from Carthage (cf. DT 255; Audollent 1933b; Jordan 1996: 115–23 and Jordan 1988: 129ff.). 574  CIL VIII, 19525. The date of this text ought to be reconsidered and placed in the fourth century CE given the palaeographical features of the text, which seems to be written in new Roman cursive. Such a proposal ought to be supported with an autopsy. 575  Specifically, Concordia Sagittaria/Concordia (cf. 99–100), Fontanaccia (cf. 96), Verona (cf. 106), Neapolis/Terralba (cf. 113). 576  There are many examples that are quite varied. Some texts betray linguistic syncretism, such as those in which there is a mixture of Celtic and Latin (cf. 219, 223, 453) or Gaulish and Latin (cf. 159 and 163). As far as onomastics go, in Britannia there are many examples of indigenous personal names as can be seen in tablets such as (cf. 235, 258, 266, etc.; for a discussion, see Mullen 2007). A similar phenomenon can be detected in the Graeco-Oscan curses from south Italy (e.g., 76, 78, 81). The invocation of indigenous deities, whom some practitioners may have considered more ‘reliable’, also merits mention. They include (but are not limited to) Maglus (cf. 456), Nodens (cf. 205), Ataecina Turibrigensis Proserpina (cf. 120), Savus (cf. 529) and Adsagsona (158). In my view, the evidence for such a wide array of local gods attests to local populations’ willingness to take up the magical praxis of inscribing defixiones (for a discussion, see section I.8.3 and Marco Simón 2010d).

562 

571 

563 

74

Distribution of Defixiones in the Roman West 9.4. Defixiones during Late Antiquity

and Hadrumetum are largely agonistic and were written by local experts. Throughout the rest of the empire, there is a substantial number of curses spanning the Roman West from Salacia/Alcácer do Sal (cf. 121) all the way to Agri decumates (cf. 477, 482–83, etc.). The wide geographical distribution of these tablets reflects the importance with which provincial societies invested the practice of writing defixiones. Furthermore, the different ways in which curses were used in particular local contexts attests to the flexibility of this religious-magical technology for addressing specific needs, concerns and anxieties.

As the following map clearly shows (see Figure 9.3), to date, the archaeological record has not provided widespread evidence for the use of defixiones in many provinces during Late Antiquity. This seems to be a continuation of a trend already initiated in the final decades of the third century BCE. In general, the few examples that we do have for this period are either scattered throughout isolated sites

Figure 9.3. Distribution of defixiones during Late Antiquity.

75

Sylloge of  Defixiones from the Roman West Aquae Tarbelliacae/Dax and Lutetia/Paris (cf. 164 and 165, respectively). In addition to these, there are also two curses from Hispania (Bracara Augusta/Braga, cf. 146– 47) and one from Germania (Gelduba/Krefeld-Gellep, cf. 462) that date to the fourth or fifth century. So far, there is only one defixio that has been securely dated to the fifth century, an extremely fragmentary tablet from Giuncalzu (Sardinia, cf. 114). A curse from Fundi/Fondi (cf. 57), which was discovered in the Grotto of Tiberius and was written by a practitioner who knew the Bible well, has been dated to the sixth century.

or concentrated in important urban centres. While the evidence for the use of defixiones in Hispania, Germania, Raetia, Noricum and Pannonia is quite scarce, we do have good evidence for the continued popularity of this magical practice in other sites. From the third to the fourth centuries CE, Britain provides a clear exception to the general trend due to the numerous finds from the sanctuaries of Sulis Minerva (Aquae Sulis/ Bath: cf. 210, 269–71, etc.) and Mercury (Uley: cf. 356, 360, 362, etc.) as well as another 12 tablets in the southern third of Great Britain, which were discovered in the vicinity of the aforementioned sanctuaries and ancient Londinium.577

Despite some notable exceptions, we can generally say that the archaeological record shows the diminishing popularity of writing curses throughout Late Antiquity. Indeed, besides the fervent activity attested at the sanctuaries of Sulis Minerva, Mercury and Anna Perenna, as well as at Porta San Sebastiano and Trier’s amphitheatre, finds from the rest of the Roman West are few and far between. These larger caches can be both multi- or single-authored. At the same time when individual practitioners who had baseline knowledge of cursing practices were writing their own texts (as can be clearly seen in Britannia), there were also professional practitioners offering their services in places like Rome, where tablets betray Graeco-Egyptian innovations (e.g., the use of iconography, charaktêres or voces magicae).

The Gaulish tablet from Rauranum/Rom (cf. 159) also dates to between the end of the third century and the beginning of the fourth century CE. This curse was found in a well together with 40 tablets that apparently were not inscribed. As mentioned elsewhere, the fact that they were found together would suggest that these were also curse tablets and that other (possibly illiterate) practitioners were using the same space for conducting similar magicalreligious practices. One of the most important collections from Rome has been dated to the fourth century: the so-called ‘Sethian’ defixiones, which were discovered at the end of the nineteenth century in a columbarium near Porta San Sebastiano. This cache includes 5 Latin texts (cf. 5–9) together with some 43 Greek tablets (cf. Wünsch 1898: SV 6-48). To this collection, which was written by several professional magical practitioners, we should also add another curse from the urbs: a tablet, which contains a list of charaktêres, was found together with an agonistic Greek curse tablet in a columbarium located at Villa Doria Pamphili (cf. 18).

Although there is an argument to be made that the spread of Christianity and the persecution of magical practices were behind the waning use of defixiones in the Roman West (cf. Dickie 2001: 242–62), we should be careful not to generalize. Indeed, and even if cursing was prohibited by the law since (at least) the time of the XII Tables, yet these legal proscriptions were unable to halt the rising popularity of defixiones across the Roman West and their extreme success in specific cities, settlements and sanctuaries. Given the fragmentary state of our knowledge, it is hard to identify other variables that could have led to this change: perhaps in certain areas of the Roman West, cursing became more of an institutionalized practice (rather than a DIY one)? Maybe practitioners continued cursing but stopped using lead and opted instead for perishable materials that left no trace in the archaeological record? Was there a shift in social perception of how these tablets worked or even if they were actually effective? In addition to these potential factors, ‘chance’ could be also a key factor. As the incredible discoveries of defixiones in Britannia, Germania and Italia have shown over the last decades, with a little ‘luck’ a single excavation can reveal a new trove of tablets from any period of time and radically shift our understanding of the waxing and waning of this cultural practice.578

Another extraordinary collection (probably written by professional practitioners), dating from the beginning of the fourth century to the fifth century CE, has been unearthed in Rome. The cache was discovered in the fountain of Anna Perenna (cf. 19–47), where curses were not only written on lead and copper sheets (sometimes inserted in lamps as if they were wicks) but also on small lead containers. In addition to this exceptional cache, we have to add the so-called ‘Bologna’ curse tablets, whose provenance is unknown and also appear to be the work of a professional magical practitioner (cf. 117–18). Outside of the Italian Peninsula, there is another important collection from Augusta Treverorum/Trier (Gallia Belgica, cf. 175–202) that was discovered in the city’s amphitheatre and contains 30 lead tablets dated to either the fourth or fifth century. From around the same time, Gaul has turned up two other isolated discoveries from

9.5. Overall distribution of defixiones in the Roman West In addition to the period-specific maps found in the previous section, we must take a more general view of

Here I am referring to the curses from Brean Down (cf. 447), Hamble (cf. 451), Farley Heath (cf. 455), Calleva Atrebatum/Silchester (cf. 454), London (cf. 341), Eccles (cf. 446), Old Harlow (cf. 353), Canonium/ Kelvedon (cf. 346), Thetford (cf. 442), Brandon (cf. 449), Ratcliffe-onSoar (cf. 351) and Lydney (cf. 205). 577 

On this, see Tomlin 2010: 270, where he compares the evolution of the discoveries from Britannia and Hispania. 578 

76

Distribution of Defixiones in the Roman West

Figure 9.4. Distribution of defixiones in the Roman West.

the distribution of the corpus throughout the Roman West. With this goal in mind, the following map (see Figure 9.4) brings together the curses discussed in the previous sections as well as the tablets that cannot be dated.579 For a variety of reasons, these tablets constitute nearly 20 per

cent of the whole sylloge. Nevertheless, we can distinguish three clear phases. In the first phase, which stretches from the Republic until the reign of Augustus, we can see that the origin of this praxis is closely linked to the Hellenic communities of Magna Graecia. Their contact with other populations through colonization and trade (first in Italy and later in Hispania and probably in North Africa) planted the seed that allowed for the initial spread of defixiones through the Mediterranean provinces. The influence of Hellenic practices is clearly seen in the use of certain formulae,

579  For Italy, cf. 40, 42–47, 87, 91, 93, 97–98, 112; for Africa Proconsularis, see Hadrumetum (AE 1905, 171; Audollent 1908: 290–96 and AE 1968, 620), Carthage (AE 1907, 165 and AE 1933, 56) and Naro/HammamLif (Audollent 1910: 137–41, Pl. XXVIII); for Hispania, cf. 148; for the Galliae, cf. 150–56, 158, 161–62, 171 and 203; for Britannia, cf. 207–09, 211–12, 216–32, 287–98, 317–35, 340, 344–45, 347–48, 350, 374–88, 440–41, 444–45, 450, 452 and 458; for Germania, cf. 461.

77

Sylloge of  Defixiones from the Roman West layout and even the contexts in which the tablets were deposited. In the Italian Peninsula, the Oscan population seems to be largely responsible for transmitting the habit of writing curses to the Romans.

Finally, let us turn to geographical distribution. In Italy we see that the peninsula is rich in archaeological sites that have turned up defixiones. During the Republic, there was a dense network of sites in the southern portion of the peninsula, above all in settlements located by the sea which were in close contact with Greek colonies. In the High Empire, Rome emerged as the epicentre of a broader pattern of distribution of curse tablets that stretched from Concordia Sagittaria/Concordia all the way to Copia Thurii/ Sybaris, with some examples attested also on Corsica and Sardinia. Turning to North Africa, the earliest known curse dates to the third century BCE. That said, it was not until the High Empire that the praxis truly blossomed, principally in the most important cities, Carthage and Hadrumetum, where several magical practitioners (both Latin and Greek speakers) offered their services. Moving to the north-west, we find a similar pattern in Hispania, though on a much smaller scale. The first known Latin tablets date to the Republic and come from two of the leading settlements of the time: Emporion and Corduba/Cordoba. During the High Empire, the practice continued not only in those communities, but also in Saguntum/Sagunto. The rest of the curse tablets have mostly been found in the southern half of Baetica.

In a next phase, it was the Romans who spread this originally Greek practice. Thus, during the High Empire, the use of curse tablets reached its zenith, both numerically and also in terms of geographical distribution. The diffusion of this cultural practice has been traditionally and logically associated with Roman conquest and the process of ‘Romanization’: as Roman legions arrived at new areas, they brought with them the practice of inscribing defixiones as part of their cultural baggage. Indeed, curses from this period have been found in the majority of provinces. This praxis was then adopted and adapted by provincial societies, as can be reflected in the onomastic aspects of the texts, the invocation of indigenous deities and even certain linguistic syncretisms. That said, the boom in the production of curse tablets also coincides with a different process, that is, the ‘Roman epigraphic habit’. This should hardly be seen as coincidental: the widespread use of defixiones also evidences the diffusion of writing more generally or at least a rudimentary knowledge thereof. Together with the spread of writing, there was also a spread in the basic knowledge of the technology of cursing, a magical-religious savoir faire that was probably, in many instances, transmitted by word of mouth. Such a technology was indeed very useful for attempting to solve all kinds of personal crises and misfortunes, above all for those who could not appeal to Roman authorities. In such instances, ‘necessity created ability’, as Hanson has put it (1991: 181).580 But, at the same time as individual practitioners decided to manufacture their own curse tablets mostly in the north-western provinces of the empire, there were also professional practitioners offering their ‘services’ in cities such as Carthage or Hadrumetum, where the texts clearly betray GraecoEgyptian techniques.

Further north, it was not until the High Empire that the praxis of inscribing defixiones was adopted; as mentioned several times above, this process should be linked to the process of ‘Romanization’ and the arrival of the Roman army. In Gaul, for instance, the earliest curses date to the first century CE and are found mostly from Aquitania and Narbonensis (provinces that have turned up many isolated finds). The practice is attested here and there throughout the Late Empire, the cache from the amphitheatre in Trier (Gallia Belgica) being an outstanding example. The first tablets to appear in Roman Britain date between the first and second centuries CE. By the end of the second and during the third century, the popularity of the practice increased exponentially. While the sanctuaries in Aquae Sulis/Bath and Uley do provide the bulk of the evidence, it is nevertheless true that curses tablets also hail from many other sites located in the South of the island, in conjunction with temples, urban centres and rural communities in civilian areas.581 While acknowledging that the archaeological record is far from complete, we must work with the evidence that we do have. Within this framework, the sheer number of tablets discovered at the British sanctuaries dwarfs the caches that (so far) we have found at any other sanctuary or temple. Is difficult to explain how this technology became so popular for communicating with the gods, especially in such a personal manner. To this end and in contrast with other provinces (in which there are professional practitioners offering their ‘services’), Britannia gives evidence of individual scriptores, a fact

So far, the archaeological record signals a marked decline in the use of defixiones in Late Antiquity. Overall, curses have been found at a much smaller number of sites, both individually and in larger caches, which do point to sustained activity in certain areas, whether on an individual level (such as in the sanctuaries of Sulis Minerva in Aquae Sulis/Bath or Mercury in Uley), or through the contracting of professional practitioners (such as in the Fountain of Anna Perenna in Rome). Despite these limited sites where the practice seems to have been sustained, the overall evidence shows that the writing of defixiones waned during Late Antiquity: barely a sixth part of the known tablets can be dated to this period. To buttress the idea that there was at least a fairly widespread level of basic literacy, we must remember that of the 130 tablets from Bath only 2 were written by the same hand (cf. 300–01).

580 

On their distribution, see Mattingly 2004: 18–21. In general, see Tomlin 2002. 581 

78

Distribution of Defixiones in the Roman West that demonstrates how this practice came to be embraced by local populations.582 In Germania, the bulk of the defixiones hail from Germania Superior. Indeed, the most prolific sites are Mogontiacum/ Mainz (with the sanctuary of Isis and Magna Mater) as well as nearby Cruciniacum/Bad Kreuznach. The curses from the Agri Decumates are also of great interest since they attest to the importance of the Roman army for the diffusion of this praxis. To date, all curses from Germania date to the High Empire. The same chronology applies to the curses from Raetia, Noricum and Pannonia, which have collectively offered 16 tablets scattered across the three provinciae. To conclude, let us briefly return to the content of these curses. Even if most of the texts do not mention the reasons why they were written, when they do so, matters of justice (whether related to the courts or in response to a theft) are the most widely attested motivation for writing a curse tablet. That said, there is certainly a wide range of motivating factors: anything perceived as an injustice, a misfortune or a personal crisis could lead a practitioner to write or commission a defixio. Although for some scholars, this fact provides evidence for an ancient Mediterranean belief,583 I think that this connection can be more broadly defined as part of the human condition: when faced with adversity, it is common to ask for divine intervention.

The large number of individual authors is evident in the cache from Bath (cf. n. 580). For the existence of magical practitioners in North Africa and Rome (sanctuary of Anna Perenna), see section I.2 and nn. 39–41, with further references. 583  As Tomlin has put it, ‘the British texts… express a much older belief which is not western and localised, but is common to the Mediterranean world —the idea that we can ask the gods for justice’ (2010: 247). 582 

79

BAR IN TERNATIONA L SE RIE S 3077 (I)

‘Apart from the sheer value of presenting over 520 texts together in one volume, [this book’s] greatest achievement is perhaps the provision of drawings of all the originals, so as to enable readers not merely to check the texts printed … but also to give a better impression of the materiality of these unprepossessing objects. The author has herself inspected many of the tablets in various countries, and provides a number of better readings. … She can be counted one of the three or four leading experts on these Latin texts in the world.’ Professor Dr Richard Gordon, University of Erfurt ‘A precious and fundamental tool for knowledge and research in a complex branch of studies.’ Dr Gabriella Bevilacqua, Sapienza University of Rome ‘There has been growing interest in “defixiones” in recent years, with other books on the subject, but no “Collection” (Sylloge) as comprehensive as this. It is a corpus with a very full introduction and commentary on each item. … It is very wide-ranging and will become an essential work of reference.’ Dr Roger Tomlin, Wolfson College, University of Oxford ‘The inaccessibility of these tablet texts and the impossibility of conducting systematic overviews has been a major hurdle to advancing the scholarly understanding of these materials. This book represents a large advance in the scholarship. … While [it] will be most useful to specialists in ancient Greek and Roman religion and magic, it will be of value to scholars in a variety of other fields as well. Social historians will be provided with a corpus of data that lies well outside the normal materials, allowing insights into the lives of marginalized peoples who would not otherwise appear in the epigraphic record (much less the literary or historical one).’ Professor Radcliffe G. Edmonds III, Bryn Mawr College

This is volume I of a two-volume set. Defixiones, also known as curse tablets, are one of the most revealing sources for ancient magicoreligious practices. Born of rancour, anger, desire, love, envy, or just out of desperation, curse tablets were a strategy for obtaining ‘individual justice’ for those who could not accomplish their purposes through the available means, due to a lack of knowledge, power or legal/economical resources. In this volume, the reader will find a detailed catalogue that discusses 535 curse tablets written in Latin and a wide range of local languages. The catalogue is preceded by a full introduction in which the main features of these inscriptions are discussed together with leading scholarship. Such a detailed yet global study of these texts sheds light on various aspects of curses that vary on a regional basis, thus showing how this magico-religious technology was not only adopted but also adapted in new and creative ways by the local populations throughout the Roman West. Celia Sánchez Natalías received her doctorate in Ancient History from the Universities of Zaragoza and Verona (2013, in co-tutelle). Her research has focused on magical-religious practices in Antiquity and the curse tablets from the Roman West.

Printed in England

B A R I N T E R NAT I O NA L S E R I E S 3 0 7 7 ( I I )

2022

Sylloge of Defixiones from the Roman West A comprehensive collection of curse tablets from the fourth century BCE to the fifth century CE

Volume II C E L I A S Á N C H E Z N ATA L Í A S

B A R I N T E R NAT I O NA L S E R I E S 3 0 7 7 ( I I )

2022

Sylloge of Defixiones from the Roman West A comprehensive collection of curse tablets from the fourth century BCE to the fifth century CE

Volume II C E L I A S Á N C H E Z N ATA L Í A S

Published in 2022 by BAR Publishing, Oxford BAR International Series 3077 (II) Sylloge of Defixiones from the Roman West, Volume II ISBN  978 ISBN  978 ISBN  978 ISBN  978

1 1 1 1

4073 4073 4073 4073

5931 1 Volume I (paperback) 5932 8 Volume II (paperback) 1532 4 (Set of both volumes) paperback 5382 1 (Set of both volumes) e-format

doi  https://doi.org/10.30861/9781407315324 A catalogue record for this book is available from the British Library. © Celia Sánchez Natalías 2022 Cover image  Numen from SD 118. Courtesy: Bologna, Museo Civico Archeologico, Archivio fotografico. The Author’s moral rights under the 1988 UK Copyright, Designs and Patents Act are hereby expressly asserted. All rights reserved. No part of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any form of digital format or transmitted in any form digitally, without the written permission of the Publisher. Links to third party websites are provided by BAR Publishing in good faith and for information only. BAR Publishing disclaims any responsibility for the materials contained in any third-party website referenced in this work.

BAR titles are available from: Email Phone Fax

BAR Publishing 122 Banbury Rd, Oxford, ox2 7bp, uk [email protected] +44 (0)1865 310431 +44 (0)1865 316916 www.barpublishing.com

Contents VOLUME I List of Figures................................................................................................................................................................... xiii List of Tables...................................................................................................................................................................... xv I. Prolegomena.................................................................................................................................................................... 1 1. A Brief History of Previous Scholarship...................................................................................................................... 3 2. A Definition..................................................................................................................................................................... 5 3. Media............................................................................................................................................................................... 7 3.1 Introduction................................................................................................................................................................ 7 3.2 Specific Media............................................................................................................................................................ 7 3.3 Non-specific media.................................................................................................................................................. 12 3.3.1 Perishable non-specific media......................................................................................................................... 12 3.3.2 Non-perishable non-specific media................................................................................................................. 12 4. Inscribing the Defixiones............................................................................................................................................. 15 4.1 Introduction: the written and spoken word.............................................................................................................. 15 4.2 The act of inscription............................................................................................................................................... 16 4.2.1 The text: layout, language and writing............................................................................................................ 16 4.2.1.1 The layout............................................................................................................................................ 16 4.2.1.2 Language and writing.......................................................................................................................... 19 4.2.2 Other components linked to the text............................................................................................................... 22 4.2.2.1 Imagines magicae................................................................................................................................ 22 4.2.2.2 Magical symbols and charaktêres....................................................................................................... 25 4.2.2.3 Other voces magicae............................................................................................................................ 26 4.3 Formulae defigendi.................................................................................................................................................. 28 5. The Manipulation of Defixiones.................................................................................................................................. 31 5.1 Introduction: technical considerations..................................................................................................................... 31 5.2 Folded tablets........................................................................................................................................................... 31 5.3 Rolled tablets............................................................................................................................................................ 32 5.4 Pierced tablets.......................................................................................................................................................... 33 5.5 Other modi operandi................................................................................................................................................ 35 6. Deposition Contexts: Where the Curses have been Discovered............................................................................... 37 6.1 Introduction: putting ‘context’ into context............................................................................................................. 37 6.2 Funerary contexts..................................................................................................................................................... 37 6.2.1 Geographical and temporal data about funerary deposits............................................................................... 38 6.2.2 The various modi operandi.............................................................................................................................. 39 6.3 Aquatic contexts....................................................................................................................................................... 41 6.3.1 Geographical and temporal data about aquatic deposits................................................................................. 42 6.3.2 Aquatic contexts as sacred spaces................................................................................................................... 42 6.4 Sacred contexts........................................................................................................................................................ 44 6.4.1 Geographical and temporal data about deposits in sacred contexts................................................................ 44 6.4.2 Spaces for communicating with the divine..................................................................................................... 45 6.5 Spaces associated with the victim............................................................................................................................ 48 6.5.1 Geographical and temporal data about deposits closely associated with the victim....................................... 49 6.5.2 Domestic spaces.............................................................................................................................................. 50 6.5.3 The agonistic sphere........................................................................................................................................ 50

iii

Sylloge of  Defixiones from the Roman West 7. Categorization of the Defixiones................................................................................................................................. 53 7.1 Introduction.............................................................................................................................................................. 53 7.2 A tacit agreement...................................................................................................................................................... 53 7.3 Erotic defixiones....................................................................................................................................................... 55 7.4 Agonistic defixiones................................................................................................................................................. 56 7.5 Juridical defixiones................................................................................................................................................... 58 7.6 Defixiones against thieves........................................................................................................................................ 59 7.6.1 A different approach: prayers for justice?....................................................................................................... 59 7.6.2 Defixiones against thieves (or in fures).......................................................................................................... 61 8. The Pantheon of Deities Invoked................................................................................................................................ 63 8.1 Introduction.............................................................................................................................................................. 63 8.2 Graeco-Roman gods and supernatural entities......................................................................................................... 63 8.3 Indigenous divinities................................................................................................................................................ 65 8.4 Oriental powers........................................................................................................................................................ 66 9. Distribution of Defixiones in the Roman West.......................................................................................................... 69 9.1 Introduction.............................................................................................................................................................. 69 9.2 Distribution of defixiones during the Republic........................................................................................................ 70 9.3 Defixiones from the High Empire............................................................................................................................ 72 9.4 Defixiones during Late Antiquity............................................................................................................................. 75 9.5 Overall distribution of defixiones in the Roman West.............................................................................................. 76

VOLUME II II. Sylloge........................................................................................................................................................................... 81 Note to the reader........................................................................................................................................................... 83 Diacritical signs............................................................................................................................................................. 84 Location......................................................................................................................................................................... 85 Sylloge of defixiones from the Roman West (SD)......................................................................................................... 86 1. Italia........................................................................................................................................................................... 89 2. Hispania................................................................................................................................................................... 199 3. Galliae...................................................................................................................................................................... 229 4. Britannia.................................................................................................................................................................. 265 5. Germania.................................................................................................................................................................. 385 6. Raetia....................................................................................................................................................................... 437 7. Noricum................................................................................................................................................................... 445 8. Pannonia.................................................................................................................................................................. 449 III. Bibliography............................................................................................................................................................. 463 IV. Appendices................................................................................................................................................................. 495 1. Table of correspondences......................................................................................................................................... 497 2. Tables Prolegomenon............................................................................................................................................... 509 3. Plates........................................................................................................................................................................ 517 V. Indices.......................................................................................................................................................................... 567 Index of Deities and Supernatural Beings.................................................................................................................... 569 Select Index of Passages.............................................................................................................................................. 571 Index of Places............................................................................................................................................................. 573

iv

List of Figures Figure 2.1. Two daemones shackling a charioteer................................................................................................................ 5 Figure 3.1. Curse tablet made of lead and papyrus............................................................................................................... 8 Figure 3.2. Curse tablet on stone.......................................................................................................................................... 8 Figure 3.3. Different shapes of lead tablets: irregular......................................................................................................... 10 Figure 3.4. Different shapes of lead tablets: tabula cerata................................................................................................. 10 Figure 3.5. Different shapes of lead tablets: disc................................................................................................................ 10 Figure 3.6. Different shapes of lead tablets: tabula cum capitulo...................................................................................... 10 Figure 3.7. Different shapes of lead tablets: tabula ansata................................................................................................. 11 Figure 3.8. Non-specific media: tegula from Thysdrus/El Jem.......................................................................................... 13 Figure 3.9. Non-specific media: curse on a pewter plate from Aquae Sulis/Bath.............................................................. 14 Figure 4.1. Defixio from Lincolnshire................................................................................................................................. 15 Figure 4.2. Different types of right-to-left layouts: Fectio/Vechten.................................................................................... 18 Figure 4.3. Different types of right-to-left layouts: Colonia/Cologne................................................................................ 18 Figure 4.4. Different types of right-to-left layouts: Celti/Peñaflor..................................................................................... 18 Figure 4.5. DT 233: ‘Fence’-style layout............................................................................................................................ 18 Figure 4.6. Detail from the Aratispi/Villanueva del Cauche defixio................................................................................... 22 Figure 4.7. Imagines magicae: daemon Baitmo Arbitto..................................................................................................... 23 Figure 4.8. Imagines magicae: Saturnius............................................................................................................................ 24 Figure 4.9. Imagines magicae: tied charioteer.................................................................................................................... 24 Figure 4.10. Imagines magicae: victim encircled by snakes.............................................................................................. 24 Figure 4.11. Charaktêres attested more than 15 times........................................................................................................ 26 Figure 5.1. Folded defixio................................................................................................................................................... 32 Figure 5.2. Rolled defixio.................................................................................................................................................... 33 Figure 5.3. Defixio rolled around a bird bone..................................................................................................................... 33 Figure 5.4. Pierced defixio.................................................................................................................................................. 34 Figure 5.5. Pierced defixio from Pompei/Pompeii.............................................................................................................. 34 Figure 5.6. Cut defixio from Ateste/Este............................................................................................................................. 35 Figure 5.7. Melted defixio................................................................................................................................................... 35 v

Sylloge of  Defixiones from the Roman West Figure 7.1. Erotic defixio of attraction, in which charaktêres and various individuals can be seen (two of which appear to be kissing).................................................................................................................................... 55 Figure 7.2. DT 246: a servant characterized as Mercury Psychopomp checks the death of another gladiator in the arena......................................................................................................................................... 57 Figure 9.1. Defixiones dating to the Republican period...................................................................................................... 70 Figure 9.2. Distribution of defixiones during the High Empire.......................................................................................... 73 Figure 9.3. Distribution of defixiones during Late Antiquity.............................................................................................. 75 Figure 9.4. Distribution of defixiones in the Roman West.................................................................................................. 77

vi

List of Tables Table 3.1. Materials used to make specific media................................................................................................................ 8 Table 3.2. Non-specific media used for writing defixiones................................................................................................. 12 Table 4.1. Defixiones grouped by provenance and their characteristics............................................................................. 17 Table 5.1. Defixiones classified according to provenance and type of manipulation......................................................... 32 Table 6.1. Defixiones discovered in funerary contexts....................................................................................................... 38 Table 6.2. Defixiones discovered in aquatic contexts......................................................................................................... 43 Table 6.3. Defixiones discovered in sacred spaces.............................................................................................................. 45 Table 6.4. Defixiones discovered in contexts closely associated with the victim............................................................... 49 Table 7.1. Defixiones classified according to provenance and typology............................................................................. 54 Table 8.1. Defixiones invoking a deity according to provenance....................................................................................... 64 Table 8.2. Table containing the supernatural beings that are invoked on more than one occasion in the curses from the Roman West.................................................................................................................................... 64

vii

II. Sylloge

81

Note to the Reader The following sylloge gathers and systematically organizes 535 defixiones from the Roman West that were published by the summer of 2018.1 Throughout the catalogue as well as the prolegomena, these curse tablets are referred to with a number in bold. The following pages discuss curses written in Latin, a mixture of Latin and Greek as well as those written in indigenous languages, such as Celtic, Gaulish, Oscan and Etruscan. These are the general inclusion criteria for the volume. That said, there are two important caveats that must be addressed.

In the vast majority of cases, every curse tablet has its own file, which comprises four main sections: 1. The first one deals with basic information, such as the findspot and the current location of the tablets (together, when possible, with the inventory number; for museums, see the list of abbreviations below), the date, the type of material on which the curse was written and its measurements (in centimetres: height × width × thickness; or diameter [marked with the symbol Ø] × thickness; when measurements are given in parenthesis, this indicates that the tablet is broken). 2. The second section comprises the reading of the tablet, following the diacritical signs of CIL II2 (see below). In most of cases, the text is accompanied by a drawing. 3. The last section of the file contains the following: 3.1. Bibliography. 3.2. Image source and/or reference to a photograph in the plates (Appendix IV.5). 4.  A commentary, which includes the following aspects: –  Date and circumstances of the archaeological context. –  External characteristics of the item. – Textual analysis (number of lines, layout or ordinatio, palaeography [capitals, old Roman cursive or new Roman cursive], content of the text, remarkable lexical or formulaic issues and onomastics). – Manipulation of the tabella (folded, rolled, pierced, etc.). – Translation (either taken from previous editions or my own, whenever I could responsibly translate the texts).2

I. The North African provinces (i.e., Proconsularis and Byzacena) have not been included in the sylloge. This is because there are not enough published images with which I could work. This means that I was unable to examine curses whose readings clearly need to be verified through high-quality images or autopsy. Furthermore, these texts are highly visual, and publishing them without their iconography would be a serious shortcoming (preparing these curse tablets for a separate publication is the goal of an ongoing research project). Nevertheless, the basic characteristics of the North African corpus have been taken into account in the prolegomena of this volume. II. The geographical dispersion and the sheer volume of the corpus have made it difficult to apply the highest epigraphic criteria for every piece in the catalogue: there is not an apparatus criticus in the following entries, and I was only able to carry out autopsies of a portion of the texts (as specified in the commentary). For these reasons, I use the word sylloge instead of corpus.

In cases where the published information about the pieces is extremely scarce (e.g., 42–47, 290–98, etc.), I have collected them in joint files.

The organization of the sylloge follows the geographical scheme set out in CIL, categorizing pieces according to the regio or provincia. The order in which certain sites from each region or province are discussed is determined by the earliest editio princeps. That said, curses from the same site are dealt with sequentially even if the first editions were not published at the same time. For example, all the pieces from Anna Perenna (47 in total) are found in sequence after number 19, which was first published in 2002, even though the latter pieces in the collection were published in 2012. When two pieces from the same regio/ provincia were first published in the same year, they are organized alphabetically by (modern) toponym. 1  Nevertheless, new scholarly discussions of previously published texts have been included in the commentary and bibliography. So far, since summer 2018, 14 new curse tablets from Italy (cf. Varone 2019), Hispania (cf. García Dils and Rubio 2018 and Moret et al. 2019), Galliae (cf. Belanger Sarrazin et al. 2019), Britannia (cf. Tomlin 2018), Germania (cf. Scholz 2019) and Raetia (cf. Blänsdorf 2019) have been published. To these, we ought to add the forthcoming publication of the Uley curse tablets (by R.S.O. Tomlin).

2  This means that I do not translate any of the following: texts that are too fragmentary or that defy reasonably certain interpretation, simple lists of names, or texts whose reading I cannot confirm through the drawings or photographs that have been published to date. Also, texts that are identical or nearly identical have been translated only once (cf. 10–14).

83

Diacritical signs (according to CIL II2) abc clear letters ABC clear letters of uncertain interpretation ạbc ̣ ambiguous letters, identifiable by their context +++ traces of illegible letters abc letters read by previous editors and now lost [-c.1-] lost letters (known quantity/number) [---] lost letters (unknown quantity/number) [------] lost line ------ lost lines (unknown quantity/number) âb ligatures `ab´ letters added in Antiquity to correct or complete a text [abc] lost letters (because of damage), restored by the editor {abc} superfluous letters, struck by the editor omitted letters, restored by the editor (a)bc omitted letters (due to phonetic or grammatical reasons), restored by the editor ˹abc˺ letters corrected by the editor a(bc) abbreviation expanded by the editor [[abc]] erased but legible text «abc» text struck over erasure (crux) image (described in commentary by the editor) (?) uncertain reading · interpunctuation (vacat) blank space / new line

84

Locations AK Bonn: Akademisches Kunstmuseum in Bonn. AM: Archaeological Museum. ASM: Archäologische Staatssammlung München. BHM: Budapest History Museum. CLPC: Centro Logístico de Patrimonio Cultural (Junta de Andalucía). DAD: Dépôt Archéologique Départamental. DAR: Deposito/Antiquarium di Roccagloriosa. DL; GKE: Direktion Landesarchäologie; Generaldirektion Kulturelles Erbe. IMMH; TA: Isis- und Mater Magna-Heiligtum; Taberna Archaeologica. MUF Freiburg: Museum für Ur- und Frühgeschichte, Freiburg. M: Museo/ Musée/ Museum. MA: Museo Arqueológico/ Musée Archéologique. MAC: Museu d’Arqueologia de Catalunya. MACB: Museo Archeologico Civico di Bologna. MAN: Museo Arqueológico Nacional/ Museo Archeologico Nazionale. MANP: Museo Archeologico Numismatico Provinciale. MAP: Museo Arqueológico Provincial. MHM: Museo Histórico Municipal. MN: Museo Nazionale/ Musée National. MNA: Museu Nacional de Arqueologia/ Museo Nazionale Archeologico. MNAR: Museo Nacional de Arte Romano (Merida). MNHA: Musée National d’Histoire et d’Art (Luxembourg). MNM: Magyar Nemzeti Múzeum (Budapest). MNRPMT: Museo Nazionale Romano. Palazzo Massimo alle Terme. MNRTD: Museo Nazionale Romano. Terme di Diocleziano. MOLA: Museum of London Archaeology. NRLM: National Roman Legion Museum (Caerleon). PSM: Prähistorische Staatssammlung München. PUGKW: Provinciaal Utrechts Genootschap van Kunsten en Wetenschappenv. RLT: Rheinisches Landesmuseum Trier. SBAM: Soprintendenza per i Beni Archeologici per il Molise. SBASN: Soprintendenza per i Beni Archeologici di Sassari e Nuoro. SBAVV: Soprintendenza per i Beni Archeologici del Veneto. Nucleo Operativo di Verona. SMBPKA: Staatliche Museen zu Berlin Preussischer Kulturbesitz, Antikensammlung. SSBAR: Soprintendenza Speziale dei Beni Archeologici di Roma. TLMF: Tiroles Landesmuseum Ferdinandeum.

85

Sylloge of defixiones from the Roman West (SD) 1. ITALIA (119)

Sardinia: 110–11 Orosei 112 Locus Incertus 113 Neapolis, Terralba 114 Giuncalzu Corsica: 115 Mariana, Lucciana Locus Incertus 116–19

Roma 1–48 Regio I: [Latium] 49–51 Nomentum, Mentana 52–54 Ostia 55 Tibur, Tivoli [Latium adiectum] 56 Minturnae, Minturno 57 Fundi, Fondi 58 Ardea 59–60 Antium, Anzio [Campania] 61–65 Cumae, Cuma 66–68 Capua, Santa Maria Capua Vetere 69 Salernum, Salerno 70 Cales, Calvi Risorta 71–72 Pompeii, Pompei 73 Aquinum, Aquino Regio III [Bruttium et Lucania]: 74 Tiriolo 75 Crimisa, Cirò Marina 76 Roccagloriosa 77 Castiglione di Paludi 78–80 Laos, Marcellina 81 Petelia, Strongoli 82 Copia Thurii, Sybaris Regio IV [Samnium]: 83 Marsi Marruvium, San Benedetto 84 Aquilonia, Monte Vairano Regio V [Picenum]: 85 Septempeda, San Severino 86 Urbs Salvia, Urbisaglia Regio VII [Etruria]: 87–91 Volaterrae, Volterra 92 Arretium, Arezzo 93 Perusia, Perugia 94 Aquae Populoniae, Campiglia Marittima 95 Caere, Cerveteri 96 Fontanaccia Regio VIII [Aemilia]: 97–98 Classis, Classe Regio X [Venetia et Histria]: 99–100 Concordia Sagittaria, Concordia 101–02 Pola, Pula 103 Ateste, Este 104 Altinum, Altino 105 Cremona 106–08 Verona Sicilia: 109 Lilybaeum, Marsala

2. HISPANIA (30) Lusitania: 120 Emerita Augusta, Merida 121 Salacia, Alcácer do Sal Baetica: 122–26 Corduba, Cordoba 127 Italica, Santiponce 128 Baelo Claudia, Bolonia 129 Carmo, Carmona 130 Aratispi, Villanueva del Cauche 131 Celti, Peñaflor 132 Astigi, Écija 133 El Portal Citerior: 134–39 Emporion, Ampurias 140–44 Saguntum, Sagunto 145 Barchín del Hoyo 146–47 Bracara Augusta, Braga 148 Cabrera de Mar Locus Incertus 149 3. GALLIAE (55) Gallia Narbonensis: 150–55 Amélie-les-Bains 156 Mazan 157 Montfo 158 L’Hospitalet-du-Larzac Gallia Aquitania: 159 Rauranum, Rom 160 Mediolanum Santonum, Villepouge-Chagnon 161 Les Martres-de-Veyre 162 Murol 163 Arverni, Chamalières (Puy-de-Dôme) 164 Aquae Tarbelliacae, Dax Gallia Lugdunensis: 165 Lutetia, Paris 166 Autun 167–70 Mediolanum, Evreux 171 Vindinum, Le Mans (Sarthe) 172 Autricum, Chartres

86

Sylloge of defixiones from the Roman West (SD) Gallia Belgica: 173 Maar 174 Durocortorum, Reims 175–202 Augusta Treverorum, Trier 203 Dalheim 204 Wolberg

6. RAETIA (6)

4. BRITANNIA (256)

7. NORICUM (1)

520–21 Brigantium, Bregenz 522 Cambodunum, Kempten 523 Veldidena, Wilten 524 Peiting 525 Abusina, Eining

205 Lydney 206–335 Aquae Sulis, Bath 336 Chesterton-on-Fosse 337 Isca Silurum, Caerleon 338–44 Londinium, London 345 Clothall 346 Canonium, Kelvedon 347–48 Bravonium, Leintwardine 349–51 Ratcliffe-on-Soar 352 Wanborough 353 Old Harlow 354–40 Uley 441 Venta Icenorum, Caistor St Edmund 442 Thetford 443–45 Pagans Hill 446 Eccles 447 Brean Down 448 Puckeridge Braughing 449 Brandon 450 Weeting-with-Broomhill 451 Hamble 452 Marlborough Downs 453 Dodford 454 Calleva Atrevatum, Silchester 455 Farley Heath 456–57 Ratae Corieltauvorum, Leicester 458 East Farleigh Locus Incertus 459–60

526 Faviana, Mautern 8. PANNONIA (9) Superior: 527 Poetovio, Ptuj 528 Emona, Ljubljana 529 Siscia, Sisak 530 Carnuntum, Petronell 531 Savaria, Szombathely Inferior: 532–34 Aquincum, Budapest Locus Incertus 535

5. GERMANIA (59) Inferior: 461 Fectio, Vechten 462 Gelduba, Krefeld-Gellep 463 Bodegraven 464 Colonia Claudia Ara Agrippinensium, Cologne 465 Ulpia Noviomagus, Nijmegen Superior: 466–76 Cruciniacum, Bad Kreuznach 477–78 Frankfurt 479 Waldmössingen 480 Roßdorf 481 Aventicum, Avenches 482–83 Groß Gerau 484–17 Mogontiacum, Mainz 518–19 Centum Prata, Kempraten

87

1 Italia them to the second century BCE. Since then, there has been no mention of these two items.

The importance of the Italian corpus of defixiones rests on not only its age —the earliest curses from the Roman West come from Sicily and the Italian Peninsula— but also on its size: after Britannia, Italia has yielded the greatest number of tablets. The Italian group discussed in this catalogue contains a total of 119 texts, in addition to which several other items must be mentioned (see below). The oldest curses from Italia date to the fourth century BCE, when we first find defixiones in Graeco-Oscan contexts (e.g., 74, 77–81). By way of the Oscan context and beginning in the second century BCE, this praxis began to spread throughout the peninsula slowly but relentlessly. By the High Empire, the practice of writing defixiones truly flourished. In Late Antiquity there are only a few isolated examples of curse tablets throughout the peninsula (e.g., 57 and 114), although Rome itself notably maintained this magical practice due to the influence of new Graeco-Egyptian practices, as is clearly evidenced by the exceptional discoveries from Porta San Sebastiano and the sanctuary of Anna Perenna.

[Regio I, Campania] Cumae, Cuma: during an 1880 excavation of the necropolis in Cumae archaeologists discovered a folded lead tablet that had been pierced four times with a nail, which could still be seen in a hole at the tablet’s corner. The text contains five lines of capitals, which run from left to right and read: Barcathes Dasi M(arci) L(ibertus)/ M(arcus) Dassius M(arcus) Allius/ Cerdo salvi nisei ista/ re qua(m) ages qum Q(uinto) Cava(rio?)/ literas perlegerit C(aius) Vitrasi(us) (see CIL X, 8214 = CIL I2, 3128). As Audollent has pointed out, ‘insueto sane modo conscripta, et in v. 3–4 obscurissima’ (DT 197). Recently, Bevilacqua has suggested that the curse is a ‘sorta di “contro-defixio” cautelativa (...) da una minaccia di scongiuro nei confronti dei tre individui nominati’ (2017: 99–100). Although the tablet’s archaeological context, medium and formal aspects all point to an identification as a defixio, its true character is still unclear. For that reason, I have mentioned the item here in this introductory section.

As mentioned above, in addition to the 119 tablets discussed in this chapter, we must add several others that are either still being studied or were found in older excavations and subsequently lost before being properly published (e.g., the tablets from Aquileia and Capua). Here, I provide a brief discussion of relevant tablets that are not included in the general catalogue, which is organized by regiones and the date of any published notification.

[Regio I, Campania] Pompeii: Della Corte (1916: 305, fig. 19) indicated that during the excavation of the necropolis located outside the Porta Stabia another defixio was found, which had been placed on top of cinerary urn no. 29, but ‘in così avanzato stato di ossidazione, da non consentire la lettura pur di una lettera de suo brevissimo testo, consistente forse in un semplice nome: in uno dei suoi capi è tuttora infisso un grosso chiodo di bronzo’. The whereabouts of this curse, which was written on a strip of lead measuring 2.3 × 8 cm (with the excavation inventory no. 97= Inv. SAN 142467), are currently unknown.

Rome: according to Wünsch (1898: 10), among the collection of the so-called Sethian defixiones, ‘Spuren von lateinischer Schrift finden sich auch auf Tafel 14v und 34.’ [Regio I, Latium] Ostia: Vaglieri (1912: 22; Squarciapino 1958: 14; Solin 1968: no. 36) mentioned a defixio, discovered during the excavation of a grave in Ostia, inside of which ‘si raccolsero tre vasetti a vernice marrone e una lamina di piombo in più pezzi, la cui iscrizione, difficile a leggersi, darò in altra relazione. È evidentemente un altra tabella defixionis’.

[Regio III, Bruttium et Lucania] Roccagloriosa: during the excavation in which 76 was found, archaeologists also uncovered the remains of 13 lead sheets, some of which were stilled rolled up, while others were inscribed. Many were extremely fragmentary (see Gualtieri 1990: 139–41). The editors did not determine whether these remains were actually curse tablets. As Poccetti (1990: 145) has rightly pointed out, if these are indeed defixiones, the accumulation of these tablets in a cultic area at such an early date (the fourth century BCE) would chronologically situate the curses from Roccagloriosa just after those from the sanctuary of Demeter Malophoros in Selinunte. Accordingly, this would be an extremely important cache for understanding the evolution of magical practices on the Italian peninsula. That said, Poccetti (per litt.) has recently said, ‘…all’epoca (1990) avevo visto alcuni frammenti di laminette che erano insieme alla defixio che ho pubblicato, ma queste erano così frammentarie che era impossibile

[Regio I, Campania] Capua, Santa Maria Capua Vetere: Von Duhn (apud Bücheler 1878: 3–4, n. 1= DT 194) mentioned the discovery of new defixiones in 1860, which were found in Santa Maria Capua Vetere. The texts were uncovered in a Roman grave (la Conocchia, from which 67 and 68 also came): ‘1) lungo palmi 0.20, largo 0.13, in due pezzi ed ossidata; 2) lungo palmi 0.11, largo 0.30, infissa al muro di una tomba della parte dello scritto mercè due chiodi di ferro irruginiti, uno lungo palmi 0.41 l’altro palmi 0.33.’ Although the poor condition of the tablets precluded a full edition, Audollent nevertheless dated

89

Sylloge of  Defixiones from the Roman West vedere se erano iscritte o no. Mi avevano assicurato che avrebbero fatto un restauro per vedere se si poteva vedere qualcosa, ma questo non solo finora non è stato fatto, ma dopo mie ricerche nessuno mi sa dire che fine abbiano fatto questi frammenti.’ I thank P. Poccetti for this information.

The tablets, apparently not inscribed, were found together with other remains dated between the second and fourth centuries CE. [Corsica] Mariana, Lucciana: Morachini-Mazel has refered to what appears to be another defixio discovered in grave no. 16 in the I Ponti necropolis. This inhumation, whose grave goods are dated to c. 75 CE, contained ‘une plaque de plomb très érodée qui comporta, peut-être, des lettres gravées’ (1974: 27).

[Regio VIII, Aemilia] Mutina, Fiumalbo: Calzolari (2006) has mentioned that in 1883, while preparing for the construction of an observatory on the peak of Monte Cimone, 19 fragments of a lead sheet were found alongside various coins from the Republican and Imperial periods in an area that was thought to be the site of some sort of cultic worship. Eight of these fragments (the largest measuring (6) × (8) cm and the smallest (2.7) × (3.8) cm) preserved the remains of a Latin text. These could very well be the remains of a defixio. The fragments are currently housed in the Museo Civico Archeologico di Modena.

Locus Incertus: Olivieri (1899: 197; Besnier 1920: no. 5), mentioned the existence of a fifth defixio currently preserved in the Museo Archeologico Civico di Bologna, that came together with 117–18 to the museum. Olivieri states, ‘tre frammenti e forse un quarto derivano da altra tavoletta, ma il loro contenuto non si può assolutamente decifrare perché il loro stato è assai deperito’. The autopsies that I have carried out, however, have not confirmed that these fragments belong to a single curse (for a preliminary analysis, see Sánchez Natalías 2017a: 52–54).

[Regio VIII, Aemilia] Sant Ilario d’Enza (Reggio Emilia): during a land survey of the villa I Moruzzi conducted either at the end of the 1990s or in the early 2000s, two fragments of an inscribed lead sheet were discovered. The tablet, which is still being prepared for publication, appears to belong to the group of curses against thieves and to date to the second century CE. For a preliminary note, see Sánchez Natalías 2017b. [Regio X, Venetia et Histria] Iulium Carnicum, Lagole di Calalzo: Pellegrini and Prosdocimi (1967: Ca XIII) mention in their catalogue a square lead sheet that is opisthographic, perforated in one corner and difficult to read. They note that its appearance recalls that of a curse tablet. The perforation in one corner, however, opens up the possibility that this item is actually a label. [Regio X, Venetia et Histria] Altinum, Altino: Scarfí (1972: 55, n. 2) has pointed to the existence of another defixio, mentioned by J. Marcello in La via Annia alle porte di Altino (1956), where among the discovered objects made of lead there was ‘una tavola rettangolare (cm 15 × 10) con graffitti non ancora decifrati (tabula defixionis?).’ Today the tablet’s whereabouts are unknown. [Regio X, Venetia et Histria] Aquileia: G. Brusin (apud Scarfì 1972: 56, n. 7) referenced a ‘una tavoletta di piombo con scarabocchiatevi in corsivo delle imprecazioni contro persona odiata per gelosia od altro. Vi si scorge anche, distintamente sulla copia posta accanto, una figura in piedi ed altra prona, e vi si legge il nome di Cottius.’ According to Scarfì, the tablet’s whereabouts are unknown. [Regio X, Venetia et Histria] Acellum, Montebelluna: F. Sartori (apud Scarfì 1972: 56, n. 9) mentioned that during the excavation of some Roman graves three fragments of a lead sheet were found, which could have belonged to a defixio. [Sardinia] Alghero: La Fragola (2015) has pointed to the discovery of a small cache of lead and bronze fragments found during a rescue excavation at La Purissima sanctuary.

90

Italia 1. Roma Provenance: necropolis. Current Location: private collection. Date: 1st century CE. Material: lead. Measurements: unknown. Reading: 5



Danae · ancilla · noicia Capitonis · `h´anc · (h)ostiam acceptam · habeas · et · consumas · Danaene habes E`u´tychiam Soterichi · uxorem

Bibliography: Henzen 1849: 77–78; CIL VI, 141; Wünsch 1897: 25; DT 138; Diehl 1909: no. 559; ILS, 8747; CIL I², 1013; Jeanneret 1917: no. 138; ILLRP, 1145; Kropp 2008: no. 1.4.4/2; Urbanová 2018: no. 16; Gordon 2019b: 424–25.

attention of an (unspecified) deity by giving the victim’s name in nominative (l. 1, see Gordon 2019: 425), the author of the text offers Danae to the god as a sacrificial victim (ll. 2–5, (h)ostia). In the final part of the curse, Mommsen interpreted the phrase habes E`u´tichia Soretichi uxorem (ll. 5–6) as a reminder to the deity of previous (and deadly) action taken against Eutychia. In the eyes of the defigens, this reminder would encourage the deity to take action against Danae. Concerning onomastics, Danae, Eutychia and Soterichus are all Greek cognomina (see Solin 2003: 583, 1320 and 1476, and 453, respectively).

Commentary: this defixio was discovered in the 1840s in the so-called ‘vigna Aquari’ outside of the Porta Latina in Rome (CIL I²). The inscription, written on a lead sheet of unknown dimensions, contains six lines, which run from left to right and whose words are separated with interpunctuation consisting of a single dot. The curse has been dated to the first century CE on palaeographic grounds.

After being inscribed, the tablet was rolled up and closed with an iron thread (see 52, 61 and 71 for possible parallels).

The text presents the following orthographic and phonetic features: noicia for novicia (l. 1) and ostiam for hostiam (l. 2). We do not know the circumstances that led to the cursing of the victim, defined as an ancilla noicia Capitonis (ll. 1–2, where novicia refers to ‘a slave who has only recently lost his freedom’, Lewis and Short, s.v.). After calling the

Translation (Gordon 2019b: 424): ‘Danaë the new slavegirl of Capito. May you receive this sacrificial offering, and destroy Danaë. You have already (taken) Eutychia, wife of Soretichus.’

91

Sylloge of  Defixiones from the Roman West 2. Roma Provenance: necropolis. Current Location: MNRTD. Date: 1st century BCE. [Inv. No.: 65037] Material: lead. Measurements: 13 × 30 × 0.1 cm. Reading: 5



10



Col. I quomodo · mortuos · qui · istic sepultus · est · nec · loqui nec · sermonare · potest · seic · Rhodine · apud · M(arcum) · Licinium Faustum · mortua · sit · nec loqui · nec · sermonare · possit ita uti mortuos · nec ad deos nec · ad homines acceptus · est seic Rhodine aput · M(arcum) · Licinium accepta sit et tantum valeat quantum · ille · mortuos · quei



Col. II istic sẹpultus est · Dite Pater Rhodi/ne(m) tibẹi commendo uti · semper odio sit M(arco) · Licinio Fausto item M(arcum) Hedium Amphionem item C(aium) · Popillium · Apollonium item Vennonia(m) Hermiona(m) item Sergia(m) Glycinna(m)

Bibliography: Garrucci 1853: 183; CII 2716; CIL VI, 140; CIL I², 1012; Wünsch 1897: 25; DT 139; ILS 8749; Jeanneret 1917: no. 139; Rubio and Bejarano 1955: no. 185; Ernout 1957: no. 141; ILLRP 1144; Astori 2000: 61; Kropp 2008: no. 1.4.4/3; Giovagnoli 2012; Urbanová 2018: no. 17.

is a crack running across the centre that has threatened to break the tablet into two. As I determined during an autopsy in November 2010, the missing portion of the top edge (which coincides with the crack), has resulted in the loss of the first six letters in col. II, l. 1 (underlined in the reading).

Image source: Garrucci 1853: Tab. XIII. See also App. IV.3, SD 2.

The inscription contains 18 lines written in capitals, which run from left to right and are arranged into two columns. This layout is reminiscent of archaic Greek epistles (see Sarri 2018: 89–90). The words, whose letters measure between 0.5 and 1 cm in height, are separated with interpunctuation consisting of a single dot found halfway up the letters (although note that the use of interpunctuation is not consistent). The curse has been dated to the first century BCE on palaeographic grounds. The text presents the following orthographic and phonetic

Commentary: according to Mommsen (CIL I²), this defixio was discovered in 1851 ‘in vinea Manenti, via Latina a sinistra circiter D pass. a porta in ruderibus sepulcrorum’. It was written on a rectangular lead sheet, which today is in good condition despite being quite curved (probably a consequence of how it was deposited) and missing portions of its top and right edges. Furthermore, there 92

Italia women are to be affected by the same curse: ‘Vennonia et Sergia odio sint M. Hedio Amphioni et C. Popillio Apollonio ut est Rhodine M. Licinio Fausto’ (see his comment on DT 139, p. 198).

features: mortuos for mortuus (I, ll. 1, 7 and 11), seic for sic (I, ll. 3 and 9), sermonare for sermonari (I, l. 6), aput for apud (I, l. 10), quei for qui (I, l. 11), Dite for Dis (II, l. 1), tibẹi for tibi (II, l. 2) and the loss of final –m in Rhodine(m), Vennonia(m), Hermiona(m), Sergia(m) and Glicina(m) (II, ll. 1, 6 and 7).

Translation: ‘Just as the dead man who is buried here can neither speak nor hold a conversation, so may Rhodine be dead, not speak to or hold a conversation with Marcus Licinius Faustus. Just as the dead man is neither with the gods nor with human beings, so may Rhodine not be with Marcus Licinius, and may she mean to him as much as this dead man who is buried here. Dis Pater, I hand Rhodine over to you, so that she may always be hated by Marcus Licinius Faustus. Also Marcus Hedius Amphio, also Gaius Popillius Apollonius, also Vennonia Hermiona, also Servia Glycinna.’

Based on its content, this curse should be grouped among the erotic defixiones. In this case, the defigens attempts to separate Rhodine from Marcus Licinius Faustus. To do so, he employs two similia similibus formulae in which the victim is compared to the corpse next to which the tablet was deposited. In this way, he asks Dis Pater that Rhodine be isolated. In the lines that follow, we encounter another four victims, whose names are listed and introduced with item. As Audollent has rightly pointed out, all of these

3. Roma Provenance: unknown. Current Location: MNRTD. Date: 1st century BCE. Material: lamp. Measurements: unknown. Reading:

Helenus · suom · nomen · eimferis mandat · stipem · strenam · lumen suom · secum [·] defert · nequis · eum solvat · nisi · nos · qui · fecimus

Bibliography: Garrucci 1860; CIL XV, 6265; Wünsch 1897: 25; DT 137; Jeanneret 1917: no. 137; Kropp 2008: no. 1.4.4/1; Bevilacqua 2010: 66; Sánchez Natalías 2018: 14; Urbanová 2018: no. 15.

separated with interpunctuation consisting of a single dot found halfway up the letters. We do not know the circumstances that led to the writing of this curse, in which the victim gives his own name and light to the infernal deities (ll. 1–2). This unusual formula has no parallels: even though the defigens is cursing Helenus on his own behalf, the victim is the grammatical subject and literally sends his own name (suom, l. 1), taking his own light with him (suom secum, l. 3). The text presents the following orthographic and phonetic features: eimferis for inferis (l. 1, with ei for long i), and vocalic confusion in suom for suum (ll. 1 and 3). Helenus (l. 1) is a well attested Greek personal name in Rome and the Italian Peninsula (see Solin 2003: 514). The curse’s final phrase is noteworthy, since it is stressed that the victim cannot escape the curse on his own, but only the defigentes can release him (nisi nos qui fecimus, l. 4). This phrase is analogous to ness[i] me intercedeṇte (see 358, ll. 14–16).

Image source: Dressel, apud CIL XV, 6265. Commentary: we do not know the precise archaeological context of this defixio, which was probably discovered in Rome in the middle of the nineteenth century. It was written as titulus pictus on a lamp. Using this type of medium is unparalleled in the collection of Latin curses from the western Empire (for a fourth-century Athenian parallel, see Thomson 1958: 159). The choice of such a medium, however, is hardly casual and is clearly linked to the magical ritual that accompanied the deposition of the curse. The phrase stipem strenam lumen suom secum defert (ll. 2–3) makes it clear that the lamp was used to establish an analogy with the victim, whose life is compared to the light emanating from the lamp (on this, see Sánchez Natalías 2018: 14 and Mastrocinque 2007).

Translation: ‘Helenus entrusts his name to the underworld, he brings with him his own light as an auspicious sign. May ((s)he) not be able to release him (from the curse) except for us who made (it).’

The inscription contains four lines, which run from left to right and were written in capitals. The words are

93

Sylloge of  Defixiones from the Roman West 4. Roma Provenance: necropolis. Current Location: Galleria degli Uffizi. Date: 1st century CE. [Inv. No.: 950] Material: marble. Measurements: 94 × 63 × 51cm. Reading: 5



10

Hic · stigmata · aeterna · Acte libertae · scripta sunt venenariae et · perfidae dolosae duri · pectoris clavom et restem sparteam · ut sibi collum alliget · et picem candentem pectus · malum · commurat suum manumissa grati(i)s secuta adulterum patronum circumscripsit · et ministros · ancillam · et puerum lecto · iacenti patrono abduxit · ut animo desponderet solus relictus · spoliatus senex e Hymno {F} ˹e˺ade(m) stimta secutis Zosimum

Bibliography: CIL VI, 20905; CLE 95; Wünsch 1897: 25; DT: XXXIII; Mansuelli 1958: 213–14; Kleiner 1987: 132–34; Evans Grubbs 2002; Tedeschi Grisanti and Solin 2011: 265–68; Mander 2013: no. 44.

text: he curses the fugitives and even enlists the help of his own deceased daughter as an aoros (i.e., a ‘restless dead’, on this see section I.6.2 and Evans Grubbs 2002: 242). He asks Junia Procula to take vengeance for the wrongs that he has endured. It appears that the placement of the inscription on the back of the altar (without any decoration) was a deliberate choice so that the curse would not be seen; furthermore, the choice to write it in metre may have been an additional means of obscuring the nature of the text, since cursing was illegal. However, the backside of the altar provided the only blank space left.

Commentary: this defixio is found in the funerary altar dedicated to the child Junia Procula, a medium that is unparalleled in the Roman West. Although 72 also forms part of a funerary monument, the present was written on the altar in a second phase (for the use of expensive materials in curses, see 72 and 120; for a general discussion, see Sánchez Natalías 2016b). The monument was discovered under unknown circumstances in the sixteenth century (perhaps along the Via Flaminia in Rome). The altar formed part of various private art collections (Vigna de Giovanni Poggi, Villa Giulia and Villa Medici) before being moved to the Galleria degli Uffizi (at an unknown date). The altar is richly decorated: in the centre of the piece, there is a portrait of the deceased, while on its corners there are two heads of Jupiter-Amon from whose horns garlands hang (for a discussion, see Mander 2013 and Tedeschi and Solin 2011: 266 with further bibliography). The altar bears two inscriptions: the first is found on the front of the monument and is an epitaph dedicated to Junia Procula by her father Euphrosynus and her mother, though the only part of her name that is legible here is the final -e. The mother’s name was defaced, for reasons that are explained on the backside of the altar, where the second inscription, the defixio proper, is found.

The text presents the following orthographic and phonetic features: clavom for clavum (l. 2), commurat for comburat, gratis for gratiis (both in l. 4), e for et, eade for eadem and stimta for stigmata (the last three in l. 8). Evans Grubbs (2002: 235–36) points out that the phrase clavum et restem (l. 2) is found on another epitaph from Rome (CIL VI, 12649, ll. 14–16: Atimeto lib(erto)/ cuius dolo filiam amisi restem et clavom/ unde sibi collum alliget). Furthermore, this phrase finds several interesting parallels in Martial (IV, 70) and Plautus (Poenulus 396 and Persa 815), though these texts only mention the rope and not the nail (see Evans Grubbs 2002: 235–36, n. 21 with further references). That said, it is reasonable to take the mention of a nail as a clue for identifying this text as a defixio, since nails could serve as a metonymy for this magical rite (on this, cf. I.5.4). Translation (Evans Grubbs 2002: 231): ‘Here the eternal marks of infamy have been written for Acte the freedwoman, the poisoner, faithless and deceitful, hardhearted. (I bring) a nail and a rope of broom so that she may bind her own neck, and burning pitch to consume her evil heart. Manumitted free of charge, she cheated her patron, following an adulterer, and she stole away his servants —a slave girl and a boy— while her patron was lying in bed, so that he pined away, and old man left alone and despoiled. And the same marks of infamy to Hymnus, and to those who followed Zosimus.’

This inscription, written in iambic verse, contains 10 lines, which run from left to right and were written in capitals. Several words are separated with interpuncts found halfway up the letters (note that their use is not systematic). Though it is not stated explicitly, there can be little doubt that Euphrosynus, Junia Procula’s father, is the author of this text, which denounces the behavior of Acte, his wife and the girl’s mother. After the death of their child, the liberta Acte chose to abandon the family home, taking two slaves along with her. Euphrosynus’ rage is palpable in the

94

Italia 5. Roma Provenance: necropolis. Current Location: MNRTD. Date: 4th century CE. Material: lead. Measurements: (11) × (15) cm. Reading:

[---]Οὕ filius[---] qui [---σιρι] Μνε qui [-c.15-]+O ab hac (h)ora ab hoc die ab hac noc`te´ φρι T[-c.7-]MTI[-c.10-]TE ααααααα tere contere con fr[in]ge e[t---t]rade 5 εεεεεεε morti fili[u]m Asell(a)es Praẹṣeti[ci]um p{r}istinarium ηηηηηηη qui manet in regione nona ubi videtur arte(m) sua(m) ιιιιιιι (crux) οοοοοοο facere et trade Plutoni praeposito mortuorum υυυυυυυ et si forte te contempserit patiatur febris 10 ωωωωωωω frigus tortionis palloris sudores obbripi [---]nos me Οὐσιρωομευι  lationis meridianas interdianas seru [---]IGIMO Οὐσιριασι [---] Οὐσιριασιρι  tinas nocturnas ab hac (h)ora ab hoc die ab hac [nocte?] +I++++EP Οὐσιρινεμοφρι  et perturba eum ne repra(h)eensionem (h)abeat 15 ++N M+E+ Εὐλάμον κάτεχε  et si forte occansione(m) invenerit praefocato eum ++XERE E Praestetium fili[um Asell](a)es in t(h)ermas [in] valneas in quocumque loco [-c.6-]NV+ cupede frange Pr[aesetici]o Asell(a)es et [si] forte te seducat per aliqua [---]AS Victor Asella filius Prae[se-] (vacat) [? artifici]a et rideat de te et exsultetur tibi [------]     mater   tici[us] (vacat) vince peroccide filium mares Praesete20 [------]       (crux) (crux) (crux) (vacat) cium p{r}istinarium filium [A]sell(a)es p{r}is[trina-]  (vacat) qui manet in regione [nona ed]e ede r[ius] (vacat) tacy tacy [---]+++ene [---]+++ne 25 [---]++is tacy

95

Sylloge of  Defixiones from the Roman West Bibliography: Matter 1852; De Rossi 1880: 6–9; Wünsch 1897: 27; SV 1; DT 140; Diehl 1910: no. 850; ILS 8750; Jeanneret 1917: no. 140; Preisendanz 1926: 22–41; Rubio and Bejarano 1955: no. 186; Solin 1976: 89–90; Solin 1995: 573–74; Solin 2004: 117–19; AE 2004, 201; Rüpke 2007: 169; Kropp 2008: no. 1.4.4/4; Urbanová 2018: no. 18.

This last detail has led Solin (1976: 90 and 1995: 118) to propose that this miller worked at one of the stables of a rival factio. This would squarely place this curse among the agonistic defixiones. Whatever the case may be, the defigens requests that Praeseticius suffer fevers (cf. 117) and die. The text presents the following orthographic and phonetic features: loss of final -m (arte, sua, occa{n}sione, ll. 6 and 15), Aselles for Asellae (ll. 5, 16–17 and 20), pristinarius for pistrinarius (ll. 5 and 20–22), tortionis for tortiones, palloris for pallores, obbripilationis for obripilationes (these three in ll. 10–11), serutinas for serotinas (ll. 11–13), hac for hoc (l. 13), repraecensionem for reprehensionem, abeat for habeat (both in l. 14), occansione for occasionem (l. 15), termas for thermas, valneas for balneas (both in l. 16), exsultetur for exsultet (l. 18) and mares for maris (l. 19). The lexeme serutinas (ll. 11–13), in Solin’s words, ‘paraît ici pour la première fois dans l’acception “du soir”, “vers le soir”’ (1995: 574), while repraecensionem (l. 14) is etymologically connected to Italian ‘reprendersi’ (Hülsen, apud Wünsch 1898: 7). The final phrase ede ede tacy tacy (ll. 21–22 and 25) is a transliteration of the Greek formula ἥδη ἥδη ταχύ ταχύ which is found in other curse tablets belonging to the ‘Sethian’ collection. Another shared characteristic is a similar invocation of Osiris and the appearance of the seven vowels of the Greek alphabet, which were associated with the seven planets (their ritual recitation was thought to summon all the sounds of the cosmos).

Image source: SV 1. Commentary: this defixio belongs to the group of socalled ‘Sethian’ defixiones and was discovered in c. 1850, according to De Rossi (1880: 6), ‘nella prima vigna a mano sinistra uscendo dalla porta S. Sebastiano sull’Appia, [dove] furono rinvenute molte lamine plumbee rotolate, che si disse essere state rotolate entro le olle cinerarie di un colombario.’ Matter, for his part, has noted that ‘[e]n feuillant le sol de la Vigna Marini, aboutissant à la Via Appia (…) on a trouvé, dans un tombeau très-ruiné, sauf en quelques parties: 1º. une mosaïque représentant une femme avec un enfant; 2º. Un sarcophage parfaitement conservé, et 3º. Plusieurs petits sarcophages, les uns en marbre, les autres en terre cuite, contenant une certaine quantité de feuilles de plomb toutes roulées’ (Matter 1852: 28–29). If these small coffins were for children, as seems to be the case, their use for magical ends should be connected to the belief about the victims of untimely deaths as ideal mediators, who could help practitioners communicate with the gods (see section I.6.2). The present curse was written on a lead sheet that has been broken into two fragments that seem to fit together. There are two damaged areas that have led to the loss of text at ll. 8–9 and 14–19. The inscription contains 25 lines of Greek and Latin text written in new Roman cursive. The tablet has been dated to the fourth century CE on palaeographic grounds. The layout of the text, which runs from left to right, is fitted around the piece’s iconography, which consists of four figures. We see the deity invoked, TyphonSeth, according to Audollent, in standing position and with a bow in his left hand. At his feet we find the schematic representation of the curse’s victims.

Translation (Rüpke 2007: 169, modified): ‘...from this hour, from this day, from this night... trample, crush, smash and... consign to death Praeseticius the miller, son of Aselle, who lives in Regio IX, where he can be seen going about his work; and hand him over to Pluto, the King of the Dead (praeposito mortuorum), then if he snubs you, let him suffer from fevers, cold, torments, pallor, sweating, shivering morning and noon, evening and night from this hour, from this day, from this (night); and set him to so that he has no respite, but if he does manage to get better, suffocate that Praestetius [sic], in the hot bath, in the cold bath, wherever; bind him, destroy Praeseticius the son of Aselle, and if he manages to put you off by some trick or other, and mocks you, and laughs you out, overcome him, kill him again and again, the blackguard, Praeseticius the miller, son of Aselle, who lives in Regio IX, now now, quick, quick!’

The curse is principally directed against Praeseticius, who is named six times throughout. He is further identified with a matronymic (Asell(a)e filius, ll. 5 and 16–18), his profession (p{r}istinarius, ll. 5 and 16–18) and even his home (qui manet in regione nona, ll. 6 and 21).

96

Italia 6. Roma Provenance: necropolis. Current Location: MNRTD. Date: 4th century CE. Material: lead. Measurements: 17 × 6 cm. Reading:

Asterius Asterius

Asterius Asterius Aur[i]cin[ta] liberaque nascitor (magical di matre symbol) (crux) [cu]m Samiu CA

Auricinta Auricinta

+++ECO +NNTE

(crux)

CA

liberaque nascitor di matre +ụm Samiu

Auricinta LOV QQ EEC

(crux)

Bibliography: SV 2; DT 141; Jeanneret 1917: no. 141; Preisendanz 1926: 22–41; Solin 1976: 90–91; Solin 2004: 118–19; AE 2004, 201; Kropp 2008: no. 1.4.4/5; Urbanová 2018: no. 19.

swaddled, who could be the victim. Between the two, there are magical symbols and an object that looks like an arrow with a sickle that has proven difficult to analyse. In the text, the names Asterius (for this Greek name, see Solin 2003: 1205 and 1473) and Auricinta (without any known parallels) are repeated four times. These names are found alongside the phrase liberaque nascitor di matre [c]um Samiu, in which Solin (1976 and 2004) has corrected nascitor to nascitur and di to de. According to this author, this phrase refers to an unborn (and hence unnamed child) who is also a victim of the curse (for another curse that targets a victim’s children, see 443). Following this hypothesis, cum Samiu would be a reference to the father of this unborn child. Samius is a rarely attested Latin cognomen (see OPEL IV, 46).

Image source: SV 2. Commentary: this defixio belongs to the group of so-called ‘Sethian’ defixiones and was discovered in c. 1850 inside a cinerary urn placed in the columbarium near to Porta S. Sebastiano on the Via Appia (for the archaeological context, see 5). Generally, the tablet is in good condition, though there are two damaged areas on the top half of the item. The inscription contains 22 lines of capitals, whose layout is extremely varied so that the lines can fit around the iconography. Along the left side of the tablet there are several drawings: the uppermost depicts a figure with a magical charaktêr on his chest, who may be the deity invoked. We also can see a second figure, depicted upside down and

Translation (Urbanová 2018: 239, slightly modified): ‘Asterius, Auricinta, the freedwoman and (the child?) who will be born to the mother (Auricinta) and Samius; Auricinta... who will be born to the mother Auricinta, Auricinta.’

97

Sylloge of  Defixiones from the Roman West 7. Roma Provenance: necropolis. Current Location: MNRTD. Date: 4th century CE. Material: lead. Measurements: 12 × 7 cm. Reading: A EPER[-c.1-]VS OD[-c.2-]V[---] adeoda[-c.2-]T[-c.1-]I[-c.1-]SOSTOR donat ne hui dedit ob TEDO collibertos BO+L++ 5 IDO qui precis et S[-c.2-]da te E+BAGOR[-c.3-]+C [---]AI hospite N[---] [---]TIEREO+S[---] [------] 10 [------] [------] T+[---] PARADI+A++P++ restis QV[---]C 15 +A[---]id[---] B [---]ER+[-c.2-]+OI+[---] ROMP+ADONEAETS+ CRVOCVM et date D +CIBON ab hoc+NXV 5 +C gente os quod QV[---] [---]MOSS bi paci[---] [---]+TISS++TED+es [---]M dicto IPS[---] [--- ? p]ectoris++cor[---] 10 te+NTOR+set IB+S[---] ut omnes cog[n]osc[ant] exempl[um e]or[um]

Bibliography: SV 3; DT 142; Jeanneret 1917: no. 142; Eitrem 1922: 115; Preisendanz 1926: 22–41; Kropp 2008: no. 1.4.4/6.

top-right corners. There is a lost semicircular section along the right edge and three holes (one at l. 1 and two at l. 14). The inscription contains 27 lines, which run from left to right and were written in new Roman cursive. The text has been dated to the fourth century CE based on palaeographic grounds. Despite the cracks and the corrosion of side A, some words can still be read. Thus, collibertos (A, l. 4) appears to allude to the defixio’s victims and the final phrase omnes cog[n]osc[ant] exempl[um e]or[um] (B, ll. 11–12) refers to a public and exemplary punishment that they ought to suffer (for parallels, see 128 and 491).

Image source: SV 3. Commentary: this defixio belongs to the group of so-called ‘Sethian’ defixiones and was discovered in c. 1850 inside a cinerary urn placed in the columbarium near to Porta S. Sebastiano on the Via Appia (for the archaeological context, see 5). This opisthographic curse was written on a roughly rectangular lead sheet, which has lost its bottom-left and

98

Italia 8. Roma Provenance: necropolis. Current Location: MNRTD. Date: 4th century CE. Material: lead. Measurements: (3) × (8) cm. Reading: 5

[---]++A+T[---] C[---]++ B[---]A[---]eda[---] [---]ani+++cis+H[---] [---]CI[---]QV[---]

Bibliography: SV 4; DT 143; Jeanneret 1917: no. 143; Preisendanz 1926: 22–41.

in c. 1850 inside a cinerary urn placed in the columbarium near to Porta S. Sebastiano on the Via Appia (for the archaeological context, see 5). The surface is so corroded that the inscription is hardly legible. The text contains five lines, which run from left to right and were written in new Roman cursive.

Image source: SV 4. Commentary: this fragment of a defixio belongs to the group of so-called ‘Sethian’ defixiones and was discovered

9. Roma Provenance: necropolis. Current Location: MNRTD. Date: 4th century CE. Material: lead. Measurements: (3.6) × (4.6) cm. Reading: 5



[------] [---]R[---]O[---] [---]A[---]I++[---] [---]A[---]N[---]++TE[--[---]CI[---]CO+O++ [------]

Bibliography: SV 5; DT 144; Jeanneret 1917: no. 144; Preisendanz 1926: 22–41.

discovered in c. 1850 inside a cinerary urn placed in the columbarium near to Porta S. Sebastiano on the Via Appia (for the archaeological context, see 5). The tablet is irregularly shaped and contains the remains of six lines, which run from left to right and were written in new Roman cursive.

Image source: SV 5. Commentary: this fragment of a defixio belongs to the group of so-called ‘Sethian’ defixiones and was

99

Sylloge of  Defixiones from the Roman West 10. Roma Provenance: unknown. Current Location: Johns Hopkins University. Date: 75–40 BCE. [Inv. No.: #97.6.1] Material: lead. Measurements: (31.6) × (11.3) cm. Reading: Bona · pulchra Proserpina [P]lut[o]nis · uxsor seive · me · Salviam deicere · oportet eiripias · salutem · c[orpus · co]lorem · vires · virtutes · Ploti · tradas · [Plutoni] viro · tuo · ni · possit · cogitati`onibus´ 5 sueis · hoc · vita[re tradas] illunc · febri · quartan[a]e · t[ertian]ae · cottidia[n]ae quas · [cum illo l]uct[ent deluctent illun] ev[in]cant [vincant] usq[ue dum animam] [eiu]s · eripia[nt quare ha]nc · victimam 10 tibi · trad[o Prose]rpi[na siv]e me Proserpin[a si]ve · m[e ach]er˹ou˺s`iam´ dicere oportet · m[ihi mittas a]rcessitum · canem tricepitem · qui [Ploti] cor · eripiat · Polliciar`us´ illi · te · daturum · t[r]es · victimas · 15 palma[· ca]rica[s] · por[c]um · nigrum · hoc · sei · pe[rfe]cerit [ante mensem] M[artium · haec P]r[oserpina Salvia tibi dabo] cum · compote(m) · fe[cer]is · do tibi · cap[ut] Ploti · {Avon}[{iae} Pr]oserpina S[alvia] 20 do tibi · fron[tem Plo]ti · Proserpina · Salvia do [ti]b[i] su[percilia] Ploti · Proserpin[a] Salvia do [tibi palpebra]s · Plo[ti] Proserpina · Sa[lvia do tibi pupillas] Ploti · Proser[pina Salvia do tibi nare]s 25 labra · or[iculas nasu]m lin[g]uam · dentes P[loti] ni dicere · possit · Plotius · quid [sibi dole]at · collum · umeros bracchia · d[i]git[os ni po]ssit · aliquit · se · adiutare [pe]c[tus io]cinera · cor · 30 pulmones · n[i possit] sentique · quit · sibi · doleat [intes]tina · venter · um[bi]licu[s] latera [n]i p[ossi]t · dormire · scapulas ni · poss[it] s[a]nus · dormire · viscum · sacrum · nei · possit · urinam · facere · 35 natis · anum [fem]ina · genua [crura] · tibias pe[des talos plantas] [digito]s · ungis · ni · po[ssit s]tare · [sua] [vi]rt[u]te · seive [plu]s · seive · parvum scrip[tum fuerit] · quomodo · quicqu[id] 40 legitim[e scripsit] mandavit · seic ego Ploti · ti[bi tr]ado · mando ut · tradas [mandes me]nse · Februari[o] [e]c(c)illunc · mal[e perdat mal]e · exset [mal]e · disperd[at · mandes tra]das · ni · possit 45 [ampliu]s ullum [mensem aspic]ere · [videre contempla]re

100

Italia Bibliography: Fox 1912a: 16–19; AE 1912, 139; Fox 1914; Besnier 1920: no. 33; Fox 1923: 357; Vetter 1923: 65–67; CIL I², 2520; Warmington 1953: 280–85; Lewis and Reinhold 1966: no. 179; García Ruiz 1967: no. 9; Gager 1992: no. 134; Bodel and Tracy 1997: 78; Versnel 1998: 225–27; Adams 2007: 444–51; Kropp 2008: no. 1.4.4/8; Urbanová 2018: no. 20.

Proserpina to enlist the help of Cerberus, to whom palmas, caricas and porcum nigrum (‘dates, figs and a black pig’) are offered in exchange for ripping out the victim’s heart. This request that Cerberus tear out the victim’s heart is also a common trope in other curse tablets from the first century BCE (such as 48 and 103). The second portion of the curse treats the figurative dismemberment of the victim, whose various body parts and organs are singled out and related to the specific vital functions that the defigens hopes to damage. This section of the curse is quite similar to the extremely common non permittas formulae found in British curses. This section begins with the head and its various parts (ll. 18–27), which are introduced with the anaphora Proserpina Salvia do tibi. This portion targeting the head is followed by a longer list ([nare]s labra or[iculas nasu]m lin[g]uam dentes, ll. 24–26) and is punctuated with the negative purpose clause ni dicere possit Plotius quid [sibi dole] at (ll. 26–27). Next, the curse turns to the victims upper extremities (ll. 27–29) so that the victim cannot protect himself (ni po]ssit aliquit se adiutare, ll. 28–29), before continuing with the organs inside the thorax (ll, 29–31), so that he does not know what is hurting (n[i possit] sentique quit sibi doleat, ll. 30–31). The text then turns to the area of the abdomen, flanks and back to prevent the victim from sleeping (ni possit s[a]nus dormire; the adjective sanus is found here and in 13 and 14). This is analogous to the attested phrases (non permittas) somnum (215, 237, 240, 257 and 259) and nec dormire (352, 358 and 443). Next, the defigens targets the viscum sacrum (cf. OLD 3. D) so that the victim cannot urinate (ni possit urinam facere, l. 34; cf. 358, l. 4: ne m˹eia˺t). This section of the curse ends with the lower extremities from buttocks to toe nails (ll. 35–38) so that the victim cannot stand on his own (ni po[ssit s]tare [sua][vi]rt[u]te, ll. 37–38; cf. 366, A, ll. 6–7: [n]on illis permittas nec ṣtare).

Image source: Fox 1912a: Plate II. Commentary: in 1908 the department of classical archaeology at Johns Hopkins University (Baltimore) acquired five defixiones (10–14), all of which were pierced with a single iron nail. We do not know the precise archaeological context for these tablets, although Vetter claimed that Egger had seen these curses when they were still in the possession of a Roman antiquarian, who assured him that they were found in front of Porta Salaria (1923: 65). Although they target different victims (Plotius, Avonia, Vesonia, Secunda and Aquillia), the texts of all five curses are nearly identical. Palaeographic similarities between the five tablets show that the same hand wrote the entire group. Fox has convincingly argued that the differences in letter size, inclination and alignment should be seen as evidence for the author’s haste and exhaustion. Accordingly, the editor was able to determine that the author first wrote the curses against Avonia and Vesonia (11 and 12) ‘when he was fresh and unwearied, for they exhibit the best handwriting and contain fewest errors’ (1912: 55). Next came the curses targeting Plotius and Aquillia (10 and 14) and finally the defixio directed against Secunda (13), in which the errors found in the previous two curses are repeated ‘and the increasingly careless writing manifests the writer’s weariness’ (Fox 1912: 55). Palaeographic (capitals) and linguistics aspects (e.g., ei for ī, the aspirated ch and the use of the active morphology of the deponent verb luctor) all date this collection between 75 and 40 BCE.

The third and final part of the inscription contains the final cursing formula, which establishes an analogy between the depositing of the curse and the handing over of the victim to the infernal deities invoked. All of this is to be carried out by the month of February (outside of this group, we do not have any parallels for this kind of magical deadline; for a general discussion, see Sánchez Natalías 2019a). Next, the defigens uses the anaphoric tricolon asking that the victim male perdat, male exseat, male disperdat, that is to say that he suffers a horrible death (for this idea of a horrible death, cf. 486). As Fox has pointed out, ‘perdat and disperdat seem to owe their intransitive use in this passage to the analogy of pereo’ (1912: 47; cf. the use of the verb in 17).

The text can be divided into three parts: the invocation and general curses, the figurative dismemberment of the victim and finally a final cursing formula. In the first part, the defigens invokes Proserpina, who is called the Plutonis uxsor and qualified with a string of epithets: bona, pulchra (both of these unparalleled in the curses from the Roman West), salvia (‘apparently a translation of σώτειρα, a common Greek epithet of Proserpina’, according to Fox 1912: 34) and acherousia (an adjective derived from the name of the river Acheron, which runs by the entrance of Hades). Having invoked this goddess, the practitioner, who also urges Pluto to participate, asks that eripias salutem, corpus, calorem, vires, virtutes (l. 3). This is a rather generic curse in which the health, the body, the appearance and the faculties of the victim(s) are attacked (as pointed already by Fox 1912: 35). With the phrase ni possit cogitationibus sueis hoc vitare the defigens seeks to ensure that the victim cannot escape the curse. This idea of irreversibility is attested in other curses (e.g., 3, 491). After sending various types of fevers upon the victims (for some parallels, see 5, 48 and 117), the practitioner asks

The five texts present the following orthographic, morphological and phonetic features: uxsor for uxor, seive for sive, deicere for dicere, nei and ni for ne, sueis for suis, luctent for luctentur, deluctent for deluctentur, Acheruosiam for Acherousiam, me for mihi, tricepitem for tricipitem, polliciarus for pollicearis, sei for si, compote for compotem, viscum for viscerem, oriculas for auriculas (diminutive of aures), aliquit for aliquid, quit for quid, venter for ventrem, 101

Sylloge of  Defixiones from the Roman West this victim, Proserpina or Acherusia if you prefer that I call you so. Summon for me the triple-headed hound to snatch away the heart of Plotius. Promise that you will give him three victims (gifts)―dates, figs and a black pig―if he completes this before the month of March. These I will offer you, Proserpina Salvia, when you complete this in an orderly fashion. I give over to you the head of Plotius, the slave of Avonia.

umblicus for umbilicus, natis for nates, ungis for ungues, seic for sic, exset for exeat and contemplare for contemplari. After being inscribed, the tablets were folded and pierced with an iron nail that is 12.7 cm long. When the nail was removed, the tablets broke into roughly 500 fragments. Of these, 210 were inscribed and were used to reconstruct the five defixiones.

Proserpina Salvia, I give over to you the head of Plotius. Proserpina Salvia, I give over to you the forehead of Plotius. Proserpina Salvia, I give over to you the eyebrows of Plotius. Proserpina Salvia, I give over to you the eyelids of Plotius. Proserpina Salvia, I give over to you the pupils of Plotius.

The present tablet is made up of 42 fragments that fit together to form what was a rectangular sheet. The inscription contains 46 lines, which run from left to right and were written in capitals. The words are separated with interpunctuation consisting of a single dot found halfway up the letters (though note that the use of interpuncts is not systematic). We do not know the circumstances that led to the writing of this curse, which targets Plotius, the slave of Avonia, who is the victim of 11. In addition to the general linguistic features discussed above, we must add the following orthographic, morphological and phonetic features: illunc for illum (l. 5), quas for quae (l. 7), sentique for sentire (l. 30), eccillunc for eccillum and exet for exseat (both in l. 43).

Proserpina Salvia, I give over to you the nostrils, lips, ears, nose, tongue and teeth of Plotius, so that he may not be able to say what is causing him pain; the neck, shoulders, arms, and fingers, so that he may not be able to aid himself in anyway; his breast, liver heart, and lungs, so that he may not be able to discover the source of his pain; his intestines, stomach, navel, and sides, so that he may not be able to sleep; his shoulder blades, so that he may not be able to sleep soundly; his “sacred organ” so that he may not be able to urinate; his rump, anus, thighs, knees, shanks, shins, feet, ankles, heels, toes, and toenails, so that he may not be able to stand by his own strength. No matter what he may have written, great or small, just as he has written a proper spell and commissioned it (against me), so I hand over and consign Plotius to you, so that you may take care of him by the month of February. Let him perish miserably. Let him leave life miserably. Let him be destroyed miserably. Take care of him so that he may not see another month.’

Translation (Gager 1992: no. 134, modified): ‘Good and beautiful Proserpina or Salvia, if you prefer that I call you so, wife of Pluto, snatch away the health, the body, the complexion, the strength, and the faculties of Plotius. Hand him over to Pluto, your husband. May he not be able to escape this (curse) by his wits. Hand him over to fevers―quartan, tertian and daily―so that they wrestle and struggle with him. Let them overcome him to the point where they snatch away his soul. Thus I give over to you

11. Roma Provenance: unknown. Current Location: Johns Hopkins University. Date: 75–40 BCE. [Inv. No.: #97.6.2] Material: lead. Measurements: (29) × (11.3) cm. Reading: A [B]ona · pu[lchra P]roserpina · Plutoni[s] · [u]xsor · seive [me Salviam] · decidere · oportet · eripias · salu[tem] corp[us] colorem · vires · virtutes Av[on]ia[e] t[r]adas · Plutoni · 5 viro · tuo [ni possit cogitati]onibus · s[ueis hoc] quicqui[d] vit[are tradas illanc] febri quart[an]ae · t[ertianae cottidianae] quas cum · illa [l]ucten[t deluctent illanc] evincant · vinca[t usque dum animam] 10 eius · eripiant · [qu]are hanc · victimam [tibi] trado [Proserpin]a seive me [Pros]erpina · se[ive me Ach]er˹ou˺siam dicere [opo]rtet · me m[ittas a]rcessitum · cane[m] [tricipi]te[m qui Avoniae]s (!) cor eripiat 15 [pollicearis illi te dat]urum · tres · victim[as] palmas · carica[s porcum ni]grum · hoc · sei · perfecerit · an[te mensem] Martium · haec [Salv]ia tibi · dabo cu[m] · compotem · feceris do tibi caput [A]von[iae]s Pr[ose]rpina · Salvia · d[o] 102

Italia 20



25



30



35



40



tibi frontem Avonia[e Pr]oserpina · Salvia do tibi supercilia · [Avoni]aes · Proserpina [Sa]lvia · do · tibi · palpe[bra]s Avoniaes · Proserpi[na] [S]alv[i]a · do [ti]bi · pupillas [Av]onia[e]s [Proserpina] [Sal]via · do · t[ibi] oricula[s la]bra [nares] nasum [de]ntes · liguam · Avon[iae ne dice]re · possit [Avo]nia · quid · s[i]bi [dol]eat [collum umero]s [br]acchia · digito[s ne] possit · ali[quit] se adiutare · pe[tus ioci]nera · cor pulmones · ni [possit] {quit} · sentire · quit · sibi · dolea[t intes]tina · venter · umbilicus · scapul[as] latera · ni · po[ssit] dorm[i]re · viscum sac[r]um · ni possi[t] urinam · facere [nat]is · femina anum · gen[ua crur]a · tibias · pedes talos [p]la[ntas digi]tos · ungis · ni · [po]ssit [stare] su[a virtu]te seive plus · [seiv]e p[ar]vum [scrip]tum fuerit · quomodo · quicqui[d legiti]me · scripsit · mandav[it s]eic · ego [Avo]niam tibi · trado man[do] ut · tradas [illanc] [m]ensi Februario [male perdat male e]xs[et]

B male · disperd[at] mand[es tra]das nei · po[s]s[i]t ampli[us] ull[um] men[s]em aspicere · vi[dere] contemplare Bibliography: Fox 1912a: 20–22; Besnier 1920: no. 34; Fox 1923: 357; Vetter 1923: 65–67; Rubio and Bejarano 1955: no. 180; Ernout 1957: no. 140; García Ruiz 1967: no. 10; Warmington 1953: 280–85; Bodel and Tracy 1997: 78; Astori 2000: 66–67; Adams 2007: 444–51; Kropp 2008: no. 1.4.4/9; Urbanová 2018: no. 21.

An opisthograph, the inscription contains 45 lines, which run from left to right and were written in capitals. The words are separated with interpunctuation consisting of a single dot found halfway up the letters, although the use of interpuncts is not systematic. We do not know the reasons why this text targets Avonia. The victim’s name belongs to one of Rome’s most important Plebeian gentes. In addition to the general linguistic features discussed in the commentary of 10, we must add the following orthographic, morphological and phonetic features: quicquid for quidquid (l. 6), ligua for linguam (A, l. 25), mensi for mense (A, l. 41), and the Greek genitive Avoniaes (A, ll. 14, 19, 21, 22 and 23), which may give us a clue about the author’s origins.

Commentary: in 1908 the department of classical archaeology at Johns Hopkins University (Baltimore) acquired this curse tablet. Although we do not know the precise archaeological context for this curse, there is reason to think it came from Rome. For questions about its author, palaeographic characteristics, structure and content, see the general commentary for 10.

After being inscribed, the tablet was folded and pierced with an iron nail that is 12.7 cm long together with 10 and 12–14.

This tablet has been broken into 36 fragments that fit together to form a large lead sheet, which was originally rectangular.

12. Roma Provenance: unknown. Current Location: Johns Hopkins University. Date: 75–40 BCE. [Inv. No.: #97.6.3] Material: lead. Measurements: 30.3 × 11.5 cm. Reading: A Bona · pulch[ra Proserpin]a · Plutonis · uxsor seive · me · S[al]viam dicere · oportet · eripias · salutem · c[o]rpus colorem · vires · virtutes Maximae · Vesoniae · tra[das] Pluton[i] 103

Sylloge of  Defixiones from the Roman West 5



10



15



20



25



30



35



40



viro [tuo] ni · poss[it co]gitationibus · su[is hoc] quicq[uid vitare Pro]tinus tra[das illanc] febri · qu[artanae terti]anae [cottidianae] quas · cum illa · luc[tent deluctent illanc] ev[i]ncant · vinca[t u]sque · dum · anima[m] e[ius] eripiant · q[uar]e hanc · victima[m] tibi · trad[o] · P[roserpina · sei]ve me · Pros[erpina] seive · me [Acherousiam] dicere · oporte[t mihi] mitta[s arcessitu]m · canem tr[ic]ep[item] qui [Maximae Vesonia]e cor · eripiat polli[cearis illi te da]turum · tres victimas palm[as caricas porc]um · nigrum hoc ·sei [p]erf[ecerit an]te · mense(m) · Martium · ha[e]c [Salv]ia ti[bi dabo] cum · compote(m) fece[ri]s ·do tibi · ca[put] Max[i]mae · Vesoniae P[roser]pina S[a]lvi[a] do · tibi frontem Ma[x]imae [Vesoni]ae · Proserpina S[alvia] do tibi · super[cilia V]esoniaes(!) · Proserpina Salvia · do tibi · palpetras · Maximae · Vesoniae Proserpina Salvia · do · tibi · pupillas · Vesoniae Proserpina Salvia · do · tibi · oriclas · labras · nares · nasum · lingua(m) · dentes · Maximae Vesoniae · nei · dicere · possit · Maxima · Vesonia · quid · sibi · doleat · collum · umeros bra[cchia] digitos · ni · possit · aliq/uit se [adi]utar[e pectus ioc]inera · cor pulmone[s ni possit] · sentire · quit sibi doleat · i[nte]st[ina] venter · umb[ilicum] scapulae [latera] n[i possit dormire] viscu[m sacrum] n[i possit u]rina[m] face[re natis anum femina] genua [ti]bia[s crur]a pedes talos [plantas digito]s · ungis · ni · possit · sta[re sua vir]tute · seive · plus · seive · par[vum scriptu]m fuerit · quomod[o quicquid legitime] scripsit man[davit seic ego M]ax[im]iam Veso[niam Proserpina tibi]

B trado m[ando ut tradas illa]nc · mensi · Februar[io male] male · male · perdat [male e]x[s]et · male · disp[e]rdat · tr[a]das 5 ni · possit [a]mpliu[s] ullum · [m]emsem · aspi[c]ere vid[e]re · contemplar[e] Bibliography: Fox 1912a: 22–25; Besnier 1920: no. 35; Fox 1923: 357; Vetter 1923: 65–67; Warmington 1953: 280–85; García Ruiz 1967: no. 11; Bodel and Tracy 1997: 78; Adams 2007: 444–51; Kropp 2008: no. 1.4.4/10; Urbanová 2018: no. 22.

its author, palaeographic characteristics, structure and content, see the general commentary for 10. This tablet has been broken into 52 fragments that fit together to form a large rectangular lead, which is missing little more than its left edge. An opisthograph, the inscription contains 49 lines, which run from left to right and were written in capitals. The words are separated with interpunctuation consisting of a single dot found halfway up the letters, although the use of interpuncts is not systematic. We do not know the reasons why this

Commentary: in 1908 the department of classical archaeology at Johns Hopkins University (Baltimore) acquired this curse tablet. Although we do not know the precise archaeological context for this curse, there is reason to think it came from Rome. For questions about 104

Italia text targets Maxima Vesonia. In addition to the general linguistic features discussed in the commentary of 10, we must add the following orthographic, morphological and phonetic features: the Greek genitive [V]esoniaes (l. 22), palpetras for palpebras (l. 23), oriclas for auriculas (l. 25), labras for labra (l. 25), lingua for linguam (l. 26) and scapulae for scapulas (l. 33). The editor (1912: 59–60) has highlighted the cognomen Maxima preceeding the nomen,

which is a common phenomenon in the Late-Republican period. The gens Vesonia is widely attested in central Italy, an area that coincides with what appears to be the curse’s provenance. After being inscribed, the tablet was folded and pierced with an iron nail that is 12.7 cm long together with 10–11 and 13–14.

13. Roma Provenance: unknown. Current Location: Johns Hopkins University. Date: 75–40 BCE. [Inv. No.: #97.6.4] Material: lead. Measurements: (30.4) × (16.5) cm. Reading: A [Bona pulchra Proserpina Plutonis uxor] seive · me · [Salviam dicere oportet eripias salutem] c[o]rpus [colorem vires v]irt[utes ---]I tra[d]as · [Plutoni viro tuo ni possit cogitationib]us su[i]s · hoc · 5 [quicquid vitare tradas illunc febri quar]tan[a]e [tertianae cottidianae qua]s cu[m illo l]uctent · [deluctent illunc e]vin[cant vinca]nt · usque · [dum animam eiu]s [er]ipia[nt qua]re · hanc · [victimam tibi] tra[do Proserpina] seive · me · 10 [Proserpina seive me Acherusiam] dicere · [oportet me mittas arcessitum ca]nem · [tricipitem qui ---i cor eri]pia[t Po]lliciarus [illi te daturum tres victimas pa]lmas · [caricas porcum nigrum hoc sei per]fecerit 15 [ante mensem Martium haec Prosep]ina · tibi · [dabo cum compotem feceri]s · do · tibi · [caput ---i Proserpina Salvi]a · do · tibi [frontem --- i Proserpina Sa]lvia do tibi · [supercilia ---i Proserpina Sa]lvia do tibi · 20 [palpebras --- i Proserpina Salvia] do tibi · pupillas [---i Proserpina Salvia do tibi n]ares · labra [oriculas linguam dentes n]asum · [---i ni dicere possit ---ius]quid [sibi do]leat · [collum umeros br]acc[hia] dig[itos ni] possit · 25 [aliquid se adiutare p]ect[us ioci]nera · [cor pulmones ni pos]sit sen[tire] quit · [sibi doleat intestina ve]nter umbilicus · [latera ni possit dormire s]cap[ul]as · ni [possit sanus dormire viscere]m · sacrum 30 [ni possit urinam] f[acere natis a]num [femina gen]u[a crura tibi]as · pedes B [talos plantas digitos ungis] ni · [possit stare sua vir]tu[te sei]ve [plus seive parvum script]um [fuerit quomodo quic]q[ui]t · legitime 5 [scripsit mandavit] seic ego [--- Proserpina] tibi trado [mando ut tradas i]llun[c] m[e]nsi [Februari]o mal[e p]erdat · male [exet male dispe]rd[a]t m[andes tra]das 10 [ni possit amplius ullu]m m[ense]m aspicere [videre contemplari] 105

Sylloge of  Defixiones from the Roman West Bibliography: Fox 1912a: 25–28; Besnier 1920: no. 36; Fox 1923: 357; Vetter 1923: 65–67; Warmington 1953: 280–85; García Ruiz 1967: no. 12; Bodel and Tracy 1997: 79; Adams 2007: 444–51; Kropp 2008: no. 1.4.4/11; Urbanová 2018: no. 23.

This curse consists of 22 fragments that fit together to form the right edge of the tablet. An opisthograph, the inscription contains 42 lines, which run from left to right and were written in capitals. The words are separated with interpunctuation consisting of a single dot found halfway up the letters, although the use of interpuncts is not systematic. We do not know the reasons why this text targets Secunda.

Commentary: in 1908 the department of classical archaeology at Johns Hopkins University (Baltimore) acquired this curse tablet. Although we do not know the precise archaeological context for this curse, there is reason to think it came from Rome. For questions about its author, palaeographic characteristics, structure and content, see the general commentary for 10.

After being inscribed, the tablet was folded and pierced with an iron nail that is 12.7 cm long together with 10–12 and 14.

14. Roma Provenance: unknown. Current Location: Johns Hopkins University. Date: 75–40 BCE. [Inv. No.: #97.6.5] Material: lead. Measurements: (27) × (15.4) cm. Reading: 5



10



15



20



25



30



35



40

[Bona pulchra] Proser[pin]a · Pl[utonis u]xsor se[ive] [me Salviam dic]ere [oportet erip]ias · s[al]utem [corpus colorem vires virtutes ---]AE · Aqu[illiae] [tradas Plutoni viro tuo ni] pos[sit cogit]ationibus [suis] [hoc quicquid vitare tradas i]lla[nc f]ebri qua[rt]ana[e] [tertianae cottidianae quas cum illa l]uctent [deluctent illanc evincant vi]ncant · u[sque dum animam eius eripiant qua]re hanc [victimam tibi] tra[do Proserpina] seive me [Proserpina seive me Acherusia]m · dic[ere oportet] [me mittas arcessitum canem] trice[pitem] [qui ---ae Aquilliae cor eripiat Polli]ciarus [illi te daturum tres victimas pal]ma[s caricas] [porcum nigrum hoc sei perfecerit ante mensem] [Martium haec Prosepina Salvia tibi dabo cum] [compotem feceris do tibi caput ---]ae [Aquilliae] Pros[erpina Salvia d]o t[ib]i [frontem ---ae] [Aquillae Pro]serpin[a Salvia] d[o tib]i su[percilia] [---ae Aquilliae Proserpina Salvia do tibi palpebras] [---ae Aquilliae P]ros[erpina Salvia] do tibi pu]pilla[s] [---ae Aquilliae Proserpina Salvia d]o t[ibi nares] [labra oriculas nasum linguam dentes ---ae] [Aquilliae ne dicere possit ---a Aquillia quid] [sibi doleat collum umeros bracchia digitos] [ni possit aliquid se a]diuta[re pectus] [cor iocinera pulmon]es [ni possit sentire] [quid sibi doleat i]nt[estina venter umbilicum] [latera ni possit dormire scapulas ni] [possit sana dormire vis]cum [sac]rum [ni possit urinam facere] f[emina] [natis anum genua tibia]s · cru[ra pedes] [talos plantas digi]tos · [ungis ni possit] [s]tare sua [virtute seive plus] seive paru[um sc]riptum [fuerit] [quomodo quicquid] leg[itime scri]psi[t] [mandavit] se[ic ego Aquillliam tibi trado] [mando ut tra]das m[andes illanc mense] [Februario male perdat male exset male] [disp]erd[at mandes tradas ni possit amplius] [ullum mensem aspicere videre contemplari] 106

Italia Bibliography: Fox 1912a: 28–31; Besnier 1920: no. 37; Fox 1923: 357; Vetter 1923: 65–67; García Ruiz 1967: no. 13; Bodel and Tracy 1997: 78; Adams 2007: 444–51; Kropp 2008: no. 1.4.4/12; Urbanová 2018: no. 24.

The curse, written against Aquillia for unknown reasons, consists of 51 fragments, about half of which fit back together. The inscription contains 40 lines, which run from left to right and were written in capitals. The words are separated with interpunctuation consisting of a single dot found halfway up the letters, although the use of interpuncts is not systematic. The text is quite fragmentary and has been reconstructed thanks to the other four defixiones from the collection.

Commentary: in 1908 the department of classical archaeology at Johns Hopkins University (Baltimore) acquired this curse tablet. Although we do not know the precise archaeological context for this curse, there is reason to think it came from Rome. For questions about its author, palaeographic characteristics, structure and content, see the general commentary for 10.

After being inscribed, the tablet was folded and pierced with an iron nail that is 12.7 cm long together with 10–13.

15. Roma Provenance: necropolis. Current Location: MNRPMT. Date: 2nd century CE. [Inv. No.: 82684] Material: terracotta. Measurements: 11.6 × 10 cm. Reading: A comodo isti non qumbe r˹e˺ inter s[e] (vacat) Peculio 5 illi intẹr ṣe CONSVP PII incognoscunt sic ne(c) LO Eufrates Iulius Manilia Bictoria B ego coacta Pecoris Peculi Iuli Eufrates et Maniliae Bictoriae{s} hos ego ubicumque deposue

Bibliography: Mancini 1923: 37–39; Della Corte 1938; Solin 1968: no. 33; Faraone 1991b: no. 21; Kropp 2008: no. 1.4.4/14; Sánchez Natalías 2018: 13; Urbanová 2018: no. 26.

The curse was written on a small sculptural group made of terracotta depicting two adults flanking a child (contra Urbanová 2018, who identifies it as an urn). As I was able to determine during an autopsy in November 2010, the item is in excellent condition and even has preserved traces of paint on the adults’ faces. The sculpture has two inscriptions, one made before it was fired and the other after. The first was stamped on the sculpture ante cocturam and refers to the workshop in which it was produced (it reads: LVCI), while the second, which contains the present curse, was etched on post cocturam. The use of a terracotta sculpture as the base on which to inscribe a curse is unparalleled among the defixiones from the Roman West. Given its unique status, scholars like Faraone have grouped the item among the magical figurines.

Image source: Mancini 1923: 38, figs. 13–14. See also App. IV.3, SD 15. Commentary: this defixio was discovered c. 1920 during the excavations of the area under the basilica of San Sebastian (located on the Via Appia), which resulted in the discovery of an important necropolis. According to Mancini, the curse tablet was found ‘fra il terriccio che riempiva (…) [il] vano C’ (1923: 37), in which there were three burials and the remains of burnt bones.

107

Sylloge of  Defixiones from the Roman West The curse is inscribed on both the front and the back of the terracotta, filling not only the small pedestal but also the chests of the depicted figures. The inscription contains 12 lines written in old Roman cursive, which run from left to right and whose letters measure between 0.3 and 0.9 cm in height. The circumstances that led to the writing of this curse are unknown, though its content suggests that this is an erotic defixio. The text presents the following orthographic and phonetic features: comodo for quomodo (A, l. 1), qumberu for cumber˹e˺ (A, l. 1: the ending in -v is clear; since e in this hand is traced with two vertical strokes, it is possible that the defigens had closed it into an v given the rounded surface of the medium), Bictoria for Victoria (A, l. 8 and B, l. 3), deposue for deposui (B, l. 4). During my autopsy I confirmed the reading CONSVP-/PIIIN (A, ll. 5-6), which is quite clear. Kropp has interpreted the word as consu(m)p(t)ionem, though I propose a new reading with different word breaks: taking

the final –in as part of the subsequent word would result in the reading incognoscunt (l. 6). That leaves CONSVP-/ PII (ll. 5–6), which may be related to a verb like consopio. These changes are not reflected in the drawing of the text, which has been taken from the editio princeps. Finally, it is worth noting the similia similibus formula used to draw an analogy between the victims and the sculptural group (see 3 for parallels). If cumber˹e˺ inter se (A, ll. 2–4) is taken as a synonym for the compound concumbo (see OLD, s.v. II, ‘to lie with (for sexual intercourse)’), the inscription could have a sexual connotation. Through this expression, and as I have proposed elsewhere (Sánchez Natalías 2018: 13), the magical practitioner was perhaps establishing an analogy between the victims of the curse and the figures depicted in the sculpture. The victims will become as rigid and immobile as the terracotta figures, so as to hinder any relationship between them.

16. Roma Provenance: unknown. Current Location: MNRTD. Date: 4th–5th centuries CE. Material: cinerary urn. Measurements: 14.5 × 16.5 (Ø) cm. Reading: 5



Col. I deprecor vos sancti Angeli ụṭ quomodo (ha)ec anima intus inc ̣ḷusa tenetur et angustiatur ẹṭ non vede(t) ṇẹque ḷumine ne aḷịquem rẹfṛ ịgerium non (h)abet ṣịc ̣ ụṭ anima ṃẹṇṭẹṣ c ̣ọrpos Collecticii quem pepereṭ Agneḷḷạ



Bibliography: Muzzioli 1939; AE 1941, 138; Silva Neto 1948–49: 72–73; Rubio and Bejarano 1955: no. 195**; García Ruiz 1967: no. 66; Solin 1968: no. 34; Kropp 2008: no. 1.4.4/13; Urbanová 2018: no. 25; Gordon 2019b: 425–26.

Col. II. teneatur ardẹạṭ de{s}tabesc ̣ạṭ usque ạd infernum ṣẹmper dụcite Collecticium quem peperet Agnella

to have come from Rome or nearby the capital. The Museo Nazionale Romano (Terme di Diocleziano) received the urn as a donation in the 1930s. The curse was written on a ceramic cinerary urn ‘apposta alla parete interna’, according to the editor (1939: 42) like a titulus pictus. Writing a curse on the inside of a cinerary urn is unparalleled in the larger corpus of defixiones from the Roman West, though there are several tantalizing curses that bear similarities. There is, for example, a ‘counter defixio’

Image source: Muzzioli 1939: 49, Tab. I, fig. b. Commentary: we do not have any precise details about this defixio’s archaeological context, though it is thought 108

Italia spirit held within the cinerary urn and the victim. This formula helps explain why the curse was written on the inside of the urn: the defigens wanted the curse to be in direct contact with the deceased. The invocation sancti angeli has been connected to the phrases ἅγιοι ἄνγελοι and the ἐξορκίζω ὑμᾶς ἅγιοι ἄνγελοι, which are found in the so-called ‘Sethian’ defixiones (see Muzzioli 1939: 43–44 and DT 156, 157, 162–64, 168 and 187). As Gordon (2019: 425–26) points out, the text has a Christian ‘flavour’ derived from the use of certain epitaphs: especially, note the terms anima, rẹfṛ ịgerium (both in I, l. 5, which refer to the consolation of the deceased) and angustiare (I, l. 3; common in the earliest translations of the Bible). There are no known parallels for Collecticius, but Agnella, a diminutive from agnus, is an attested cognomen (on these cognomina see Kajanto 1982²: 325).

written against anyone who may disturb the spirit held in a cinerary urn from Lugo (on this see Hervés Raigoso 1995 and Marco Simón 2011). Another point of comparison would be the lead containers from Anna Perenna, if we accept the hypothesis that these are indeed small cinerary urns (see section I.3.2, with references). The text of this curse is arranged in two columns, each of which contains six lines that run from left to right and were written in new Roman cursive. The text has been dated to the fourth–fifth centuries CE on palaeographic grounds. We do not know the circumstances that led to the writing of this curse, which targets the soul, discernment and soul of Collecticius, quem pepereṭ Agneḷḷạ (I, l. 6). This matronymic formula is first attested in North African defixiones (e.g., DT 247, 250 and 253) and also in a LateAntique curse tablet from Rome (see 20). The text presents the following orthographic and phonetic features: ec for haec (I, l. 2); vede for videt, ḷumine for lumen, aliquem for aliquod (the three in col. I, l. 4); pepereṭ for peperit (I, l. 6 and II, l. 5); c ̣ọrpos for corpus, ṃẹṇṭẹṣ for mens (both in I, l. 6) and destabesc ̣ạṭ for detabescat (II, l. 2). Another noteworthy feature of this curse is the use of a similia similibus formula, which draws an analogy between the

Translation (Gordon 2019b: 425 and my own) ‘Holy angels, I beseech you that, as this (dead) soul confined here is secured and fixed and unable to see the light (of day) or enjoy any comfort, so may the soul, the consciousness, the body of Collecticius, who Agnela bore, be secured, burn, rot away. Drag Collecticius, who Agnela bore, continuously, to the Underworld.’

17. Roma Provenance: domestic setting. Current Location: MNRTD. Date: 1st century CE. [Inv. No.: 538909] Material: lead. Measurements: 1.9–3.3 × 12.2 × 0.1 cm. Reading: A D(i) · M̂anes · com(m)ando · ut perdant B in{n}imicos · meus com(m)and · Domitia Omonia · Menecratis · alius · trado · Nicea · Cyrus · Nice Porista · Demo · Asclepiades Time Ce Philaia Caletic(he) · Menotia itim {M} 5 atversar(ios) annor(um) · menor(um)

Bibliography: Panciera 1968: 332–40; Solin 1968: no. 35; Panciera 2006; Kropp 2008: no. 1.4.4/15; Urbanová 2018: no. 27.

curse was deposited in a domestic setting, placed under the pavement or inside the wall of one of the victim’s homes (for some parallels, see 522, 524, 528, etc.).

Image source: Panciera 1968: Tab.II, figs. 3 and 5.

The curse was written on an irregularly shaped lead sheet, which is curved at certain points and had been cut out from a larger sheet. In November 2010 I was able to confirm during an autopsy that the curse is in good condition, although several small parts of its surface are corroded. An opisthograph, the inscription contains seven lines, which run from left to right and were written in capitals. Several words are separated with interpunctuation consisting of a single dot found halfway up the letters. On side A,

Commentary: this defixio was discovered between 1950 and 1952 during the excavations of the Palatine near the socalled ‘house of Livia’. As Panciera has noted, ‘il terreno appariva profondamente sconvolto ed il materiale trovato con la tabella è risultato estremamente eterogeneo’ (2006: 146). Although we do not know whether this is where the tablet was originally deposited, it is possible that the 109

Sylloge of  Defixiones from the Roman West Solin 2003: 1337). This first victim is separated from the others with the phrase alius trado (B, l. 2; here, alius seems adverbial, meaning ‘again’ or ‘furthermore’). The remaining victims were, judging from their names, slaves, who were probably of Greek origin: Nicea (taken as Nicaea or Nicias, see Solin 2003: 904–05), Cyrus (see Solin 2003: 241), Porista (see Solin 2003: 1079), Demo (see Solin 2003: 1373), Asclepiades, Time (see Solin 2003: 1346), Ce (for Ge, see Solin 2003: 323), Philaia (for Philaea, see Solin 2003: 956), Caletic(he) and Menotia (perhaps for Menothea, see Solin 2003: 113). The curse concludes with the phrase itim atversar(ios) annor(um) menor(um) (B, ll. 4–5), which seeks to extend the curse to the descendants of the listed victims (for parallels, see 57, 68, 246 and 370).

the height of the letters ranges between 0.5 and 0.9 cm, while on side B letters measure between 0.2 and 0.8 cm. The inscription can be dated to the first century CE on palaeographic grounds. We do not know the circumstances that led to the writing of this curse, in which the Manes are invoked and asked to take action against the inimici of the principal. Side A contains the invocation and beginning of the curse, while side B lists the names of the victims. The text presents the following orthographic, morphological and phonetic features: comando for commendo (A, l. 1), innimicos for inimicos, meus for meos and again comand(o) for commendo (the last three in B, l. 1), itim for item (B, l. 4) and atversar(ios) for adversarios (B, l. 5). The verb perdant has been understood in various ways: its subject could either be the Di Manes and its direct object in{n}imicos or, as Panciera has noted, it could be taken as equivalent to pereant (2006: 147; for perdo used for pereo see 10–14).

Translation: ‘Di Manes, I hand over (them) to you so (they) be destroyed. I hand over my enemies to you: Domitia Omonia (wife?) of Menecrates. Furthermore, I deliver: Nicea, Cyrus, Nice, Porista, Demo, Asclepiades, Time, Ce, Philaia, Caletiche, Menotia, and also the fewyears-old opponents.’

When it comes to the victims, Domitia Omonia Menecratis (uxor) probably refers to a liberta (for the cognomen, see

18. Roma Provenance: necropolis. Current Location: MNRTD. Date: 4th century CE. [Inv. No.: 402697] Material: lead. Measurements: 3 × 6.7 cm. Reading: (charaktêres) αυιεογαμβα ισ (charaktêr) υυυμχμριχ (charaktêr) μνα βαγα (charaktêres) οr (charaktêr) ει αυτο υ τατιη ει Tullia (?)

Bibliography: Bevilacqua 1998: 132–34; Bevilacqua 2012: 613.

during an autopsy in November 2010. The item has been damaged in three areas: the first two affect the first line, whereas the third, which is more extensive, affects the beginning of ll. 2–4. The surface of the tablet is corroded, especially its upper half and along its right side.

Image source: Bevilacqua 1998: 134. Courtesy of G. Bevilacqua. Commentary: this defixio was discovered in 1984 during the excavations carried out at the Villa Doria Pamphili. The tablet was found inside a columbarium dated between the mid-first century and second centuries CE. An agonistic Greek defixio was also found in the same place (on this, see Bevilacqua 1998: 114–32).

The inscription contains four lines of texts and strings of charaktêres. The text is surrounded by an incised line running around the perimeter of the tablet (for a parallel, see 183 and 184). We do not know the circumstances that led to the writing of this curse, in which Bevilacqua reads the female personal name Τατίη (for Tatia) in l. 4, which is followed by a word that is ‘incerto (…) se si tratti di charaktêres oppure di caratteri latini corsivi’, in which case Tullia would be a possible reading.

This curse was written on a rectangular lead sheet, which has been broken along its edges, as I was able to confirm 110

Italia 19. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th century CE. [Inv. No.: 475561] Material: lead. Measurements: 6.3 × 10 cm. Reading: A Antonius (crux) Ant oni us 5 An to

B ANTION TOENT VSVNV O

Bibliography: Polakova and Rapinesi 2002: 46; Piranomonte 2005: 98–99; Blänsdorf 2010a: 216–17, no. 1; AE 2010, 216; Blänsdorf 2012b: 23; Blänsdorf and Piranomonte 2012b; Piranomonte 2012a: 165.

representation of Antonius in a standing position with his arms sticking out from his sides. On the victim’s body and between his legs there are two inscriptions that read: Antonius and Anto, respectively. Side B has four lines, which again contain the victim’s name, though here the letters have been jumbled. Although Blänsdorf (2010a: 217) has taken the disordered letters as a sign of the practitioner’s poor skills, the mixed-up name could be better understood as a symbolic representation of the victim’s dismemberment. In fact, it is worth stressing how membrum can be taken as a synonym for syllaba (see TLL s.v. membrum [8.0.634.30]). As an alternative to the given reading for side B, Blänsdorf and Piranomonte (2012b) have suggested the following: (col. I) An/to/niu // (col. II) TION/ENT/NV. However, given that the text is written in scriptio continua, the proposed separation of words does not seem justified.

Commentary: in 1999, emergency excavations carried out in the Piazza Euclide (Rome) uncovered a fountain of the ancient goddess Anna Perenna and her nymphs. This sanctuary was in use between the fourth century BCE and the sixth century CE. During the excavation, archaeologists brought to light an extraordinary range of objects: 500 coins, 74 lamps, organic remains (pine cones, egg shells, pieces of wood, etc.), a copper caccabus, containers of lead and terracotta and 23 defixiones (for this deposit, see most recently Piranomonte 2015). The present curse was written on an irregularly shaped lead sheet, which is fully preserved. An opisthograph, the inscription contains nine lines, which run from left to right and were written in capitals by a poorly trained hand (for palaeographic considerations, see Blänsdorf 2012b). The name of the victim, Antonius, is centred along the top edge of side A. Below we can see an anthropomorphic

After being inscribed, the tablet was folded twice and placed inside of a lamp (Inv. 445002) in the place where the wick would be inserted. According to Mastrocinque (2007), the lamp’s light would be a symbol for the life of the victim, which would then be put out in a sort of “symbolic homicide” (for parallel types of deposit, see 20, 30–31).

20. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th century CE. [Inv. No.: 475565 A, B, C] Material: lead. Measurements: vid. reading. Reading: IB IA (2.7 × 4.2 cm) charaktêres [V]ictor quem ΩΩΩ charaktêr ΩΩΩΩ peperit Pri+ NNNABPDEX (traces) RAIA sua

IIA (2.3 × 3.7 cm) V[-c.1-]LI[-c.2-]I (h)oc++rit[-c.1-]I

IIB (charaktêres) (charaktêres) ΩΩ NNNΩ (charaktêres)



IIIA (2.8 × 4.2 cm) Victor quem peperit Pr[iva]ta

IIIB (charaktêres) (charaktêres) (charaktêres) ΩΩΩΩNN ΩIΩPQ+LQQV 111

Sylloge of  Defixiones from the Roman West Bibliography: Polakova and Rapinesi 2002: 38–47; Piranomonte 2005: 98; Blänsdorf 2006: no. II.7; Mastrocinque 2007; AE 2008, 219–21; Blänsdorf 2010a: 221, no. 6; Blänsdorf 2012d: 154–55; Blänsdorf 2012e: 631; Piranomonte 2012a: 165–66.

North African curses, e.g., DT 247, 250 and 253 and also 16). On each side B, the tablets contain between three and four lines of charaktêres, mostly an X inside a circle as well as N, Ω and X with a horizontal line on top. After being inscribed, the tablets were folded and placed inside of a lamp, replacing the wick (Inv. 445018; for some parallels, cf. 19, 30 and 31). Finally, the lamp was placed in the spring of Anna Perenna as a sort of ‘symbolic homicide’ (on this, see 19 and the interpretation of Mastrocinque 2007).

Commentary: these three defixiones, discovered in the fountain of Anna Perenna (see 19 for the archaeological context), were inscribed on three small lead sheets and placed together in the same lamp. Opisthographs, these sheets have texts written in new Roman cursive, which can be dated to the end of the fourth century CE. Sheets I and III name their victim, who is then qualified with the matronymic phrase quem peperit (this is well attested in

Translation: ‘Victor who (...) bore (series of charaktêres...) Victor who Privata bore (series of charaktêres).’

21. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th–5th centuries CE. [Inv. No.: 475566] Material: lead. Measurements: 6.3 × 13.5 × 0.1 cm. Reading: 5



Faniu(m) (H)erculiu(m) et Fapricilianu(m) (h)oc divini et (h)os accipi vigent depona(s or -m) [-c.1-](h)oc nume(n) deponam et agite ut il(l)e se {NA} bona nise (or nisi) nati(v)a et quo(d) petimus et qu(a)erimus a te

Bibliography: Piranomonte 2005: 98–99; Blänsdorf 2006: no. II.8; Blänsdorf 2010a: 227–29, no. 8; AE 2008, 222; Blänsdorf 2010c: 50–52; Blänsdorf 2012b: 26–27; Blänsdorf 2012d: 155; Blänsdorf 2012e: 632; Blänsdorf 2015a: 302–03.

–d and –m, accipi for accipe (l. 2), ile for ille (l. 3), nume for nomen (l. 4), natia for nativa (l. 4), quo for quod (l. 5) and querimus for quaerimus (l. 6).We do not know the circumstances that led to the writing of this curse, which appears to be directed against three individuals with whom the text opens: Fanius, Herculius and Fapricilianus (this last cognomen has no known parallels). Next there is a series of intermittent phrases: (h)os accipi is the formula with which ‘the god is asked to accept the targets as an offering, i. e. to kill them’ (Blänsdorf 2010a: 228, cf. 121). The defigens stresses the handing over of the victims with the phrase nomen deponam (l. 3). Then follows another cursing formula, which is incomplete (et agite ut il(l)e se..., l. 3). To conclude, the defigens reiterates his/ her petition with the final phrase et quo(d) petimus et qu(a)erimus a te (l. 6).

Commentary: this defixio, discovered in the fountain of Anna Perenna (see 19 for the archaeological context), was inscribed on an irregularly shaped lead sheet, which is intact, though its central area has been damaged. The inscription contains six lines, which were written in new Roman cursive (measuring about 0.8 cm in height; but note the capital shape of certain letters, such as n in depona(s)). The text has been dated to the third or fourth century CE on palaeographic grounds. Judging from the handwriting, vulgarization of several words and the incoherency of some phrases, the editor has suggested that the text’s author was barely literate. Such a judgement may go too far, since it did not keep the author from writing part of the text ‘in code’: some words in ll. 3 and 6 are written in an unusual order (in l. 3 the transcription depomante etigautilese ought to be taken as deponam et agite ut il(l)e se, while in l. 6 et rimus que a te is for et qu(a)erimus a te).

After being inscribed, the tablet was rolled up. Translation (Blänsdorf 2010a: 242): ‘Fanius, Herculius and Fapricilianus —(Do as I request in performing) this divine (rite)— and receive them (as a ritual offering). They are strong. Take them down below. I’ll take this name down below. And act in order that they shall (...) the property except the physical one. And what we demand and request from you.’

The text presents the following orthographic and phonetic features: the loss of initial h– as well as final

112

Italia 22. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th century CE. [Inv. No.: 475567] Material: lead. Measurements: 7.6 × 7.1 × 0.05 cm. Reading: sacras san(c)tas F a supteris et angilis a[---] quod rogo et peto magnam virtutem vestram (charak- tollatis pertolla{e}tis têres) oculus sive dextrum et 5 sinest{e}ru Surae qui nat(us) maledicta modo e˹st˺ de vulva fiat rogo et peto magnam virtu tem vestra(m) (crux) 10 tollite oculus dextru sinest{e}ru ne possit dura re virtus arbitri Surae qui natu(s) 15 est de vulva maledicta (charaktêres)

Right margin: (crux) dextru/ BLOBES Left margin: (crux) sinest{e}ru/ IRILESVS

Bibliography: Polakova and Rapinesi 2002: 46; Piranomonte 2005: 98–99; Blänsdorf 2006: no. II.9; Blänsdorf 2010a: 221–27, no. 7; AE 2008, 223; Blänsdorf 2010c: 38–43, no. 7; Faraone 2010; Blänsdorf 2012b: 33–35; Blänsdorf 2012c: 374–76; Blänsdorf 2012d: 158; Blänsdorf 2012e: 633; Piranomonte 2012a: 165– 66; Blänsdorf 2015a: 303–05; Blänsdorf 2015b: 32–36; Piranomonte 2015: 76.

a gem, see Gradvohl 2003). There are two long vertical lines that separate the rhombus from four snake-like creatures, which have large eyes. While the top-left ‘serpent’ has a sort of crest, the two bottom creatures are depicted with scales. These serpents, typical in the Late Antiquity, may be representations of the spirits sent to punish the victim (for some parallels, cf. with 33, 117, SV 16, 17, etc.). In between the vertical lines and the two arcs formed by the snakes, the defigens has included two circular symbols (perhaps eyes) that are connected to the central rhombus with arrows and the words dextru/blobes and sin{e}steru/ irilesus. Blänsdorf (2010c: 42f.) has analysed blobs as a Coptic word meaning ‘“l’occhio grande del Bes” or “Baal e Bes”’, while irilesus would be a compound of the Egyptian roots for eye and tongue. Since this interpretation, although suggestive, has no parallels in the defixiones from the Roman West, we should proceed with caution. Two lines of charaktêres surround this large iconographical unit, the first of which has five symbols and is found to the left of ll. 3–5, while the second contains seven symbols and is found at the very bottom of the tablet.

Commentary: this defixio was discovered in the fountain of Anna Perenna (see 19 for the archaeological context). The item, which is intact and in excellent condition, was inscribed on a square lead sheet. The inscription contains 16 lines, which were written in new Roman cursive, whose letters are approximately 0.1 cm in height (cf. Blänsdorf 2012b). The lines run from left to right and fit around the tablet’s iconography, which was undoubtedly drawn before the text was added. In the very centre of the tablet there is a rhombus that surrounds an anthropomorphic figure, who does not have arms or legs and on whose head four locks of hair stand erect. This figure was wrongly identified as the goddess Anna Perenna (Blänsdorf 2010c: 39), though more recently it has been reanalysed as a depiction of the victim Sura, qui natu(s) est de vulva maledicta (ll. 14–16). Accordingly, the rhombus should be interpreted as a schematic representation of a vagina so that the image and text coincide (for an interesting parallel on

The inscription is divided into two paragraphs, the shape of which conforms to the iconography. The content of these two paragraphs is fairly analogous. The text presents the following orthographic, morphological and phonetic features: santas for sanctas, supteris for

113

Sylloge of  Defixiones from the Roman West other body parts (e.g., 56 and 117), whereas here we find them separately. This focus on the victim’s eyes makes more sense given the fact that he was a judge (perhaps the curse was made to prevent him from reaching a decision). Finally, note the repeated phrase qui natu(s) est de vulva maledicta, which is similar to q(uem) p(eperit) vulva (found also on a curse from Constantina, on this, see Jordan 1976) as well as the quem peperit found in other curses (cf. the commentary for 20). Here the matronymic phrase does not actually provide the mother’s name, even though it refers to hers in a pars pro toto.

subteris, angilis for angelis (all three in l. 1), pertollaetis for pertollatis (l. 3), oculus for oculos (ll. 4 and 11), sinesteru for sinistrum (ll. 5 and 11), ets for est (l. 6), vestra for vestram (l. 9), dextru for dextrum (l. 11), natu for natus (l. 14). Based on its content, this curse could be grouped among the juridical defixiones, since the victim was an arbiter (i.e., a sort of judge). The text begins with an invocation of the nymphs (l. 1), who are the titular deities of the sanctuary and here called sacras, the supteris (l. 1, ‘easily intelligible by analogy with superis. It means the same as inferis’, according to Blänsdorf 2010a: 239), and finally the angilis (l. 1; also found in 16). Using the phrase rogo et peto (ll. 2 and 7), the defigens summons these deities and calls upon their magnam virtutem (ll. 2 and 8–9). This latter phrase is unparalleled in the corpus, though it could be seen as analogous to the more common maiestas divina (e.g., 120 and 128). After this invocation, we find the curse proper, in which the defigens makes the following demand: tollatis pertolla{e}tis oculus sive dextrum et sinesteru (ll. 3–5 and 10–11). The violence directed against the victim’s eyes is undoubtedly noteworthy, since in other anatomical curses, the eyes are normally listed among

Sura is a well attested Latin cognomen (see Kajanto 1982²: 226). Translation (Blänsdorf 2010a: 238): ‘The sacred and holy (nymphs), through the infernal gods (?) and the messengers, what I wish and demand from your great virtue: remove, utterly remove the eyes, the right or the left one, of Sura, who was born from a cursed womb. I wish and demand from your great virtue it shall happen: take the eyes, the right and the left one, in order the virtue of Sura the judge may not perish, who was born from a cursed womb.’

23. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th century CE. [Inv. No.: 475722] Material: lead. Measurements: (6.8) × (5.6) cm. Reading: 5



[---]at[---] [---]et h[oc?] [-c.1-]et ille te[---] rogat illi o[mnia? --- u-] xor eam or [u]xoream (?) [-c.1-]RTV[--- con-] fiteatur fem[inam--- sa-] n(c)tas nimfa[s---] SO[-c.2-] vo[s] BON[-c.1-]LE[---]

Bibliography: Blänsdorf 2006: no. II.10; Blänsdorf 2010a: 219–20, no. 5; AE 2008, 224; Blänsdorf 2010c: 52–54; Blänsdorf 2012b: 25–26; Blänsdorf 2012e: 636.

The circumstances that led to the writing of this curse are unknown. The phrase ille te rogat (ll. 3–4) demonstrates that the defixio contained some sort of request that appears to have involved the [sa]n(c)tas nimfa[s] (ll. 6–7, cf. 22 where they are called sacras san(c)tas), who, together with Anna Perenna, were the titular deities of the sanctuary. The verb [con]fiteatur (ll. 5–6) alludes to a possible confession of a wrongdoing that involved several women ([u]xor eam or [u]xoream (?)) and fem[inam], ll. 5–6).

Commentary: this fragment of a lead sheet comes from the fountain of Anna Perenna (see 19 for the archaeological context). The tablet has conserved its left edge and, apparently, part of the bottom edge as well. The inscription contains nine lines, which run from left to right and were written in the uncial script, whose letters measure between 0.4 and 0.6 cm in height. The text has been dated to the fourth century CE on palaeographic grounds (see Blänsdorf 2012b).

114

Italia 24. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th–5th centuries CE. [Inv. No.: 475549] Material: lead. Measurements: vid. commentary. Reading:

C (innermost container) Col. I Col. II Sebe- Θ rinu (crux) s Θ

Col. III Decentias (charaktêres)

Bibliography: Polakova and Rapinesi 2002: 50; Piranomonte 2005: 100; Piranomonte 2006: no. II.12; AE 2008, 225; Piranomonte and Marco Simón 2010: 8; Blänsdorf 2010a: 218–19, no. 3; Blänsdorf 2010c: 43– 44; Piranomonte 2010a: 207–08; Piranomonte 2010b: 26; Blänsdorf 2012d: 155–56; Blänsdorf 2012e: 623; Piranomonte 2012a: 170–71; Blänsdorf 2015b: 29–32; Piranomonte 2015: 79; Sánchez Natalías 2020b.

inscribed with a text but was decorated with horizontal lines. The middle container (B, 7 × 5.5 cm [height × diameter]) is apparently uninscribed and undecorated. The final vessel (C, 6.2 × 4.8 cm [height × diameter]) is decorated with horizontal lines and bears a short inscription arranged in three columns, which run from left to right and were written in new Roman cursive. The curse has been dated to the fourth or the fifth centuries CE on palaeographic grounds. The inscription surrounds the representation of a standing male figure, who has marked pectoral muscles, wears greaves and holds his arms at his sides. This is a depiction of the curse’s victim (contra Blänsdorf).

Image source: Sánchez Natalías 2020b: 115, fig. 1. Commentary: this defixio was inscribed on one of the lead containers discovered in the fountain of Anna Perenna (see 19 for the archaeological context), where, as we will see, a very particular ritual of aggressive magic was carried out by a single specialist. Before turning to the text, it is necessary to mention briefly the characteristics of these artefacts. Generally, they are composed of sets made up of three vessels of different sizes so that they can be nested within each other like Russian dolls. Each set was then covered with a single lid (either flat or conical), which was then sealed with natural resins or even bent along the edges to prevent it from being opened. Inside the smallest container, seven of these sets held an anthropomorphic figure, which in each case clearly represents the cursed victim. The name of the victim, along with the deities invoked (and sometimes a curse proper), appears inscribed on the surface of the two innermost vessels (see Polakova and Rapinesi 2002 and 2012: 177, fig. 3; for the containers and their respective figurines). As I have proposed elsewhere (see section I.3.2 for further references), these sets of containers resemble cinerary urns in miniature and were deliberately made for this ritual. They had two functions: first, to serve as the medium on which one could write a curse tablet; second, to serve as the place in which the victim of the spell could be symbolically buried.

The text in the first column contains the victim’s name in the nominative singular (I, ll. 1–2; contra Blänsdorf, who reads Sete/ Mnu). The name is followed by the letter Θ, which appears again above the victim’s head (II, l. 1). Following the suggestion of Piranomonte (2015: 81), this Θ ought to be understood as a theta nigrum, a symbol which is found in several gladiatorial mosaics, where it always is found next to the name or head of a dead individual. Another aspect of this inscription has a connection to the world of gladiators: the marked difference between the size of the victim’s right and left arms parallels some representations of thraeces or retiarii, who normally wore a manica, a type of armour meant to protect the arm (for a parallel, see CIL VI, 10194). The interpretation of the third column, which reads Decentias, has proven more elusive. Given that in the realm of magic the precise identification of the victim is essential, it seems unlikely that this plural personal name refers to a separate, second victim (contra Blänsdorf). Accordingly, it seems better to take this name as part of the same onomastic formula to which Seberinus belongs. This presents two options: Decentias could be an error for Decentius (with –as instead of –us, see for instance Longinas instead of Longinus, apud Kajanto 19822: 130), which would be an ironic supranomen acquired by the gladiator in the ring; alternatively, Decentias could be part of a matronymic qualifier (either

The vessel currently under consideration belonged to such a set of three lead containers, whose flat lid with a small knob was hermetically sealed (for a picture, see Piranomonte and Marco Simón 2010: 8). The outermost vessel (A, 7.5 × 6.5 cm [height × diameter]) was not 115

Sylloge of  Defixiones from the Roman West a Greek genitive or an elliptical version of the phrase [quem] Decentia [peperit] (for an in depth discussion of the possibilities and their relative merits, see Sánchez Natalías 2020b: 118–21).

who is about to be devoured by a serpent. The figurine is also nailed to an inscribed bronze sheet, which is inscribed with the letter Θ (surely another theta nigrum), several magical symbols and another anthropomorphic figure, which is depicted as standing with his arms laid across his body. This tablet was pierced with two iron nails, which connected it to the head and feet of the figurine.

Finally, we must note that the third container (C) was itself covered with a lid and held an anthropomorphic figure,

25. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th–5th centuries CE. [Inv. No.: 475558] Material: lead. Measurements: vid. commentary. Reading:

B (middle container) Ablatanabla

C (innermost container) (crux) (on the figure’s abdomen) Le- ᾿Ι(ησοῦς) Χ(ριστὸς) Ν(αζωραῖος) ᾿Ι(ησοῦς) Χ(ριστὸς) onti- Ν(αζωραῖος) κ(αὶ) Θ(εὸς) Θ(εὸς) Θ(εός) us AA IV IVS IVE IA VIN Bibliography: Polakova and Rapinesi 2002: 52; Piranomonte 2006: no. II.15; Piranomonte and Marco Simón 2010: 10; AE 2008, 226; Blänsdorf 2010a: 217, no. 2; Piranomonte 2010a: 208–09; Piranomonte 2010b: 26; AE 2010, 108–09; Piranomonte 2012b: 625; Németh 2015: 43–44.

together with the palindrome ablanathanabla, have led several scholars to identify it as the anguipede Abraxas (see Piranomonte 2012b: 625 with further references, Blänsdorf 2010a: 217, no. 2). Nevertheless, in my opinion, their proposal ought to be revised, given that several features of Abraxas are missing and the fact that the abbreviated Christian formula, which is written on its abdomen, was not taken into account in making such an identification. Here, I argue that the acronym not only serves to invoke a divine power but also to ‘name’ said power, as is the case in the representations of other numina found in the larger corpus of defixiones (see DT 283, for the example of Baitmo Arbitto). Thus, and following Németh (2012b and 2015), this could be a new (and undoubtedly unique) image of Christ. These containers were probably made by a single specialist in the fourth or fifth century CE, when basic knowledge of Christianity was well spread. In such a context, all the elements (the palindrome, the invocation of Christ and the iconography of Abraxas) would fit together, as is documented in some Gnostic magical Gems (on this, see Sánchez Natalías 2020d: 110–11).

Commentary: this defixio, discovered in the fountain of Anna Perenna (see 19 for the archaeological context), was inscribed on the surface of a container (C), which was nested inside of two larger vessels (for a discussion of these types of containers, see the commentary for 24). The largest vessel in the set is not inscribed (A, measuring 3.5 × 6.2 cm [height × diameter]); while the middle container (B, measuring 3.6 × 5.6 cm [height × diameter]) and the innermost were inscribed (C, measuring 3.3 × 5.5 cm [height × diameter]). The smallest vessel was found upside down compared to the other two and held an inscribed bone and a fragment of parchment, both of which are today illegible. Upon discovery, this set of containers was covered with a flat lid (for an image, see Polakova and Rapinesi 2002: 39, fig. 4). The inscription on container C has four lines of capitals written by a poorly trained hand. Ll. 1–3 are arranged in two columns that fit around the curse’s iconography, which depicts a standing figure with a rooster’s head, a crest and long tail. On the demon’s chest a short Greek text was inscribed and has been interpreted as the abbreviation of the Christian formula ᾿Ιησοῦσ Χριστὸς Ναζωραῖος, ᾿Ιησοῦς Χριστὸς Ναζωραῖος, καὶ Θεός, Θεός, Θεός (for this formula, see Németh 2012b and 2015). The iconographic characteristics of the figure,

Finally, to the left of the demon we can read Leontius, the victim’s name. Blänsdorf (2010 a: 217) has interpreted the fourth line of the inscription, which contains a series of capitals without any known meaning, as well as the bird head that appears next to the figure as previous attempts to inscribe the container. Translation (Németh 2015): ‘Ablanathanabla. Leontius. Jesus Christ of Nazareth, Jesus Christ of Nazareth, and God, God, God.’ 116

Italia 26. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th–5th centuries CE. [Inv. No.: 475539] Material: lead. Measurements: vid. commentary. Reading: B (middle container) (charaktêres)

C (innermost container) Αβλαναθαναλβα (crux, charaktêr)



(on the figure’s abdomen) ᾿Ι(ησοῦς) Χ(ριστὸς) Ν(αζωραῖος) ὁ π(αῖς) Χ(ριστὸς) κ(αὶ) Θ(εὸς) Θ(εὸς) Θ(εός)



(on the bottom of the container) Quirinus Pistor Auctulus Quirius qui natus est de Equi[---] [De]centia Seberi Decentia [C]omeronis VES[---]SA[---]TAM nocturnas quam diernas iam iam cito cito modo modo

Bibliography: Polakova and Rapinesi 2002: 48; Piranomonte 2006: no. II.11; Piranomonte and Marco Simón 2010: 10–11; Piranomonte 2009: 258 and 262–63; Blänsdorf 2010c: 36–38; Piranomonte 2010a: 210; Piranomonte 2010b: 27; Blänsdorf 2012b: 28–30; Blänsdorf 2012d: 155–56; Blänsdorf and Piranomonte 2012a; Németh 2012b; Piranomonte 2012a: 165, 167; Németh 2015: 43–45; Piranomonte 2015: 76.

As mentioned above, the second inscription is found on the bottom of the innermost container (5.4 cm diameter), which lends itself to the spiral layout. The text, written in new Roman cursive, has been dated to the end of the third or beginning of the fourth century CE on palaeographic grounds (see Blänsdorf 2012b). Given that all the containers from the fountain seem to be made by a single specialist, however, it would be reasonable to date this one into the fourth–fifth centuries CE as well. We do not know the circumstances that led to the writing of this curse, which targets four named individuals: two men (the second of whom is qualified by a matronymic) and two women (each is qualified by a patronymic or the name of her husband). The adjectives nocturnas and diernas (for diurnas) generally characterize types of fevers, which could very well be the punishment that these four victims ought to endure (for a parallel of similar date, see 116). Finally, the defigens uses the formula iam, iam, cito, cito, modo, modo to urge the powers invoked to carry out the terms of the curse as soon as possible (for this phrase, cf. 30).

Commentary: this item, discovered in the fountain of Anna Perenna (see 19 for the archaeological context), is made up of three lead containers that were nested inside one another (for a discussion of these types of containers, see the commentary for 24). The group was covered with a single conical lid with a small knob (measuring 6.5 cm in height). The outermost vessel (A, measuring 16.5 × 9.8–6.5 cm [height × diameter]) was not inscribed or decorated, while the outer surface of the middle container (B) was inscribed with a line of charaktêres. Finally, the innermost vessel (C) held an anthropomorphic male figurine (see Piranomonte 2012 a: 172, fig. 35) and contains two inscriptions: the first of which is located on the external surface of the vessel, while the second, shaped like a spiral, took advantage of the flat surface of the container’s bottom.

Translation (Blänsdorf 2012d, slightly modified and Németh 2015): ‘Ablanathanabla. Jesus Christ of Nazareth the child, Christ of Nazareth, and God, God, God.’

The first inscription contains four lines of capitals, which are written in a mix of Greek and Latin. The layout of this text is closely linked to the curse’s iconography, which represents a figure with a rooster’s head and human legs traditionally identified as Abraxas (see the commentary for 25).

Quirinus Pistor, Auctulus Quirius, son of Equi(...). Decentia (wife of) Seberus. Decentia (daughter of) Comerus, [shall suffer by fever] night and day – now, now, quickly, quickly, soon, soon”.

117

Sylloge of  Defixiones from the Roman West 27. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th–5th centuries CE. [Inv. No.: 475555] Material: lead. Measurements: vid. commentary. Reading: B (middle container) (charaktêres) Ablatanabla C (innermost container) (crux) (on the figure’s abdomen) ᾿Ι(ησοῦς) Χ(ριστὸς) Ν(αζωραῖος) ὁ Υ(ἱός) Χ(ριστὸς) Ν(αζωραῖος) κ(αὶ) Θ(εὸς) Θ(εὸς) Θ(εός) Bibliography: Polakova and Rapinesi 2002: 44; Piranomonte 2005: 99, fig. 15; Piranomonte 2006: no. II.16; Piranomonte and Marco Simón 2010: 11, fig. 17; Piranomonte 2010b: 27; Blänsdorf 2012e: 624; Piranomonte 2012a: 167–169; Németh 2012b; Németh 2015: 43 and 45.

was neither inscribed nor decorated, while the middle one (B, measuring 8.4 × 6 cm [height × diameter]) contains a series of charaktêres and the palindrome Ablatanabla (for Ablanatanabla) inscribed on its outer surface. The palindrome has been interpreted as an invocation of Abraxas, the rooster-headed demon with the body of a human, whom scholars have also identified as the figure depicted on the innermost container (C, measuring 8 × 5.3 cm [height × diameter]; for a different interpretation, see the commentary for 25). Even though it is no possible to date the text on palaeographic grounds (it is likely that the palindrome was written in capitals, but this information has not been published yet), it seems very probable that the set under consideration was made in Late Antiquity (fourth–fifth centuries CE) by the same specialist that made the other containers from the fountain.

Commentary: this item, discovered in the fountain of Anna Perenna (see 19 for the archaeological context), is made up of three lead containers that were nested inside one another (for a discussion of these types of containers, see the commentary for 24). The set was covered with a single rounded lid (for an image see Polakova and Rapinesi 2002: 44, fig. 21). In the innermost container, an anthropomorphic female figurine was placed head down. In all likelihood, she represents the victim of the curse.

Translation (Németh 2015: 45): ‘Ablanatanabla. Jesus Christ of Nazareth the son, Christ of Nazareth, and God, God, God.’

Similar to other sets from Anna Perenna, the external container (A, measuring 9 × 7 cm [height × diameter])

28. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th–5th centuries CE. [Inv. No.: 475562] Material: lead. Measurements: 6.7 × 6.1 × 0.2 cm. Reading: VM EΘC (traces, vacat) σθ+     (crux) P (vacat) ω P Φ Y XI (vacat) Δ 5 (vacat) Δ θΑ α (vacat) ω Mα (vacat) ΘΘΘ AΙC και IΘAEIC (vacat) P 10 (vacat) +Θ α Θ Θ Ι P Θ Cassianus quem ponere[---] [---po]terat qui illas malas mulieres [---]surupserunt paup[erem] 118

Italia Bibliography: Piranomonte 2009: 257; Blänsdorf 2010c: 45–46; AE 2010, 219; Blänsdorf 2012b: 37–39; Blänsdorf 2012d: 156–57; Blänsdorf 2012e: 628.

That said, we ought to be more cautious, especially since other texts from the fountain of Anna Perenna employ abbreviated formulae that have been quite difficult to decipher (see 24–27). The second part of the inscription contains three lines (ll. 11–13), which run from left to right and are written in Latin using new Roman cursive (measuring about 0.2 cm in height). The images and drawings that have been published so far do not allow me to corroborate the proposed reading of the text. According to the editor, the curse contains the name of the victim, Cassianus (l. 11), which is followed by the phrase quem ponere (also l. 11), which appears to allude to the very act of depositing the tablet in the sanctuary. Next, we find reference to qui illas malas mulieres (l. 12) and the verb surupserunt (l. 12, for surripuit or surripuerunt) which appear to reference a theft. Nevertheless, the fragmentary nature of the text prevents us from drawing any solid conclusions about the content of this defixio.

Commentary: this defixio, discovered in the fountain of Anna Perenna (see 19 for the archaeological context), was inscribed on a square lead sheet, which has four small, damaged areas on the upper half of the tablet. The text, which contains 13 lines, was written in Latin and Greek and runs from left to right around the drawing. Iconographically, the tablet depicts a standing figure whose hair stands up on end and whose arms are separated from the body. The individual’s left hand is shaped like a claw. The identity of this figure, who could either be the text’s victim or a deity invoked, cannot be securely identified. The opening section of the text (ll. 1–10) is written in a mix of Latin and Greek, which is almost unintelligible except for και (l. 8). Blänsdorf (2010c: 46) has been quick to assume that these lines were not meant to bear any semantic meaning, but rather serve a magical purpose.

After being inscribed, the tablet was folded three or four times and then finally rolled up.

29. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th–5th centuries CE. [Inv. No.: 475721] Material: lead. Measurements: (5.4) × (4.8) cm. Reading: 5



NEO (charaktêr) +ILAN (?) VS Fortun[atus] et Seberu(charaktêr)S et locus san[ctus] DD(?)CF(?)IDSIED(?)PEB(?) II+A[---]ADII[---] ------

Bibliography: Blänsdorf 2010a: 219, no. 4; AE 2010, 217; Blänsdorf 2012b: 24–25; Blänsdorf 2012e: 636.

The circumstances that led to the writing of this curse are unknown. What we do see is a list of personal names in the nominative. While the first name has proven difficult to analyse given the lack of a solid parallel to help with its reconstruction, Fortunatus and Seberus are both well documented. The phrase locus san[ctus] (l. 4) refers to the sanctuary where the curse was deposited. The remaining lines of the text are unintelligible.

Commentary: this defixio, discovered in the fountain of Anna Perenna (see 19 for the archaeological context), was inscribed on a thin lead sheet, which has lost its bottom and right edges. The inscription contains five lines, which run from left to right and were written in capitals measuring 0.8 cm in height (see Blänsdorf 2012b). The text also contains two charâkteres, placed on top of the fourth letter of l. 1 and of the second-last letter in l. 3.

After being inscribed, the tablet was folded three times.

119

Sylloge of  Defixiones from the Roman West 30. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th–5th centuries CE. [Inv. No.: 475564] Material: copper. Measurements: (5.2) × 9.7 × 0.1 cm. Reading: 5



c[o]ndo[---]NTA [---]in[c]umbite in foco [---] I +[p]enuar++ ut p(er)eat in diebus VII[---] (magical symbols) [---] imp(er)at vobis Abrax[as---] exi E[---e]um sau[cietis---] iacbebrarum Solomon (charaktêres) [---] (charaktêres) [---] (magical symbols) (magical symbols) [---] V imperat tibi abl[anatabla] (charaktêres) [---]IA dep(rae)cor vestra(m) virtutem[---] ut eum [pe]ssime p(er)datis iạm iam cit[o] c ̣ito (vacat)

Bibliography: Blänsdorf 2010c: 46–49; AE 2010, 220; Blänsdorf 2012b: 30–31; Blänsdorf 2012d: 157; Blänsdorf 2012e: 630; Piranomonte 2012a: 167; Blänsdorf 2015b: 25–27; Piranomonte 2015: 79.

Based on the curse’s content and the editor’s analysis, it appears that the curse targets a [p]enuarius (l. 2), who is condemned to die in a fire (foco, l. 2) within a prescribed number of days determined by the practitioner (probably seven, ll. 2–3; for parallels see 456–57; on magical deadlines in defixiones see Sánchez Natalías 2019a). Next, with the phrase imp(er)at vobis Abrax[as] (l. 3), the author has Abraxas ask the numina invoked to harm the victim ([e]um sau[cietis], l. 4). After several strings of charaktêres, the expression iacbebrarum Solomon (l. 4) and a palindrome, the defigens repeats the order of attack, perhaps even calling upon Anna Perenna and her nymphs themselves (vestra(m) virtutem; cf. 22), and asks that the victim be killed immediately (iam iam cito cito is the Latin equivalent of the transliterated Greek phrase ede, ede, tacy, tacy, which is also found in 5).

Commentary: this defixio, discovered in the fountain of Anna Perenna (see 19 for the archaeological context) and now broken into four fragments that do not fit together, was inscribed on a thin copper sheet. The use of this metal instead of lead finds only two parallels in the corpus of defixiones from the Roman West: 31 and 45. The inscription contains nine lines, which run from left to right and were written in new Roman cursive, which has been dated to Late Antiquity (see Blänsdorf 2012b). The defigens appears to have been quite familiar with writing as can be deduced from his/her use of abbreviations: p with a horizontal mark down the descender is per (in pereat, imperat and perdatis, ll. 2, 3 and 8, respectively), whereas p topped with a curved mark is used for prae (depraecor, l. 7; contra Blänsdorf who reads (h)oc p(rae)cor).

After being inscribed, the tablet was folded four times over itself and then placed inside of a used lamp, as if it were the wick (Inv.: 475059, cf. 19, 20 and 31). The reddish colour of the copper could evoke the flame of the lamp. Following Mastrocinque (2007), the flame would in turn represent the life of the victim, which was sacrificed in a ritual of ‘symbolic homicide’ already attested in the fountain (on this, see 19 and 31).

The position of the different fragments is secure given the clear connection between the syllables die-/bus (ll. 2–3, which should be followed by a number [VII] rather than the reading vir, contra Blänsdorf) and virtu-/tem (ll. 7–8). The Latin portion of the text is interspersed with five strings of charaktêres (ll. 3, 5, 6, 7 and 9) that are quite similar to those found in 20. Furthermore, there appears to be magical terms like the palindrome abl[anatabla], which is paralleled in other curses from the same sanctuary (26 and 27, among others).

Translation (Blänsdorf 2015b: 26, modified): ‘I hide him (?) (…) put into the fire (…) the head of the supply, in order he may die within seven (?) days (…) Abraxas gives you the command go out (…) in order you shall hurt him (…) Iacbebrarum Solomon (…) he commands you Abla[tanabla] (…) I implore your virtue (…) to destroy him badly, soon, soon, quickly [quickly].’

120

Italia 31. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th–5th centuries CE. [Inv. No.: 475563] Material: copper. Measurements: (6.8) × (7.9) × 0.1 cm. Reading: 5



I conatas suas [-c.1-] person[as ---] ILL[---] et VATICOLOM[-c.1-]L[---]ERIO [---] filio et quisquis [-c.1-]C+RM [roga-] mus cras deas vest[ra]s [-c.2-] et Cristum nostr[um --- qui] gaudent timi[a]nt[---] eum vincam I[---] C[---] [---]SVC[-c.1-]VI[---]

Bibliography: Blänsdorf 2010c: 49–50; AE 2010, 221; Blänsdorf 2012b: 31–33; Blänsdorf 2012d: 158; Blänsdorf 2012e: 629; Piranomonte 2012a: 166–67; Blänsdorf 2015b: 24–25; Piranomonte 2015: 76 and 78.

noting the phrase deas vest[ra]s (…) et Cristum nos[trum] (ll. 4–5), which certainly stands out: if the reading is correct (and I cannot confirm this from the published material), the defigens alludes to (what today we consider) two different religious spheres. On the one hand, deas vestras could refer to the nymphs of the sanctuary, while Cristum obviously refers to Christ, who is also found in other texts from the fountain of Anna Perenna (see 25–27 and 35–36). Finally, the rhetorical phrase [qui] gaudent timi[a]nt (l. 6, for timeant; without known parallel in other curses) nicely serves as a sort of all-inclusive formula, which here takes the place of the victims’ actual names.

Commentary: this defixio, discovered in the fountain of Anna Perenna (see 19 for the archaeological context), has been broken into four fragments that do not fit together (the measurements above are for the largest fragment). The curse was written on a thin sheet of copper, a material that finds only two parallels: 30 and 45. Given the use of the same material, shared palaeographic features and similar modes of deposition (both were placed inside of a lamp), it is reasonable to assume that one individual made both curses. The inscription contains eight lines, which run from left to right and were written in new Roman cursive. The reading of the text, which is controversial (see the different options published by Blänsdorf), cannot be confirmed from the drawings and photographs that have been published to date.

After being inscribed, the tablet was folded and then placed inside a lamp that had already been used (Inv.: 475501, cf. 19, 20 and 30). It is placed in the place reserved for the wick, a choice that could be related to the colour of the copper, which evokes a flame. Following Mastrocinque (2007), the flame would represent the life of the victim, who was killed in a ritual of ‘symbolic homicide’ already attested in the fountain (on this, see 19 and 30).

The phrase [roga]mus cras (ll. 3–4) appears to reference an upcoming encounter with the deities invoked. It is worth

32. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th–5th centuries CE. [Inv. No.: 548784] Material: lead. Measurements: 6.7 × 4.8 × 0.02 cm. Reading: 5



CVVOSLOVO[---]CVE quibus[---] CA[---]O[---]SCAM[---]CVN[---] quomodo qui + ASI[---]LOS[---]COS[---]OLO[---]V [---]RQVI[---]COSVLLOL[---] vos [---] [---] suo [---] [---]S[---]S[---] [---]S[---]V[---]N[---]AQ[---] I[---] erunt +E+ sibi ipsos +rem mihi sur(r)ipsit [so]lveru[nt] salutares meo[s] qui sur(r)ipserunt

Bibliography: Blänsdorf 2010c: 54–55; AE 2010, 222; Blänsdorf 2012b: 35–37; Blänsdorf 2012d: 158–59; Blänsdorf 2012e: 637; Piranomonte 2015: 76.

was inscribed on a thin lead sheet, which is currently broken in six fragments that fit together.  The inscription contains nine lines, which run from left to right and were written new Roman cursive by a trained hand who used various ligatures (such as al-,  -er-,  -rn- and-nt;  for a discussion, see Blänsdorf 2012b).

Commentary: this defixio, discovered in the fountain of Anna Perenna (see 19 for the archaeological context), 121

Sylloge of  Defixiones from the Roman West The published images and drawings are not sufficent for confirming the proposed transcription of the text, which is so fragmentary that we cannot deduce the causa defigendi. The inclusion of the word quomodo (l. 3), however, provides evidence for what would be the only use of a similia similibus formula in the curses from

the fountain of Anna Perenna. As far as orthography is concerned, note the forms suripsit and suripserunt  (ll. 8–9), which come from the verb surripere, which is also found in 28. After being inscribed, the tablet was folded five times.

33. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th–5th centuries CE. [Inv. No.: 561972] Material: lead. Measurements: 9.9 × 7.9 × 0.03 cm. Reading:

(along the upper edge of the tablet) ΔΙΑ+ΧΝΔΡΑΧ ΑΜΡΙΚ+ΑΔΜΑΟ+ΑΧΣΑ (written over the figure) ΘΙΩ+Ω ΝΙΑ+ΑΞΟ [---] Α+Ο ΥΑ+ (crux) (crux) (crux) (between the victim’s right leg and the snake) quem pereo fantasia

Bibliography: Blänsdorf 2012d: 159; Blänsdorf 2012e: 621; Piranomonte 2012a: 165–66; Blänsdorf 2015b: 27–29; Piranomonte 2015: 76, 81 and 83.

inscribed area. The inscription contains (at least) seven lines of Latin and Greek, which run from left to right and are fitted around the iconography. The tablet depicts a standing individual, who wears a headdress filled with feathers and has his hands at his sides, which are tied to his body with a dozen ropes. This is undoubtedly the victim, perhaps the same Petronius Cornigus who is represented with the figurine found inside of the smallest container. This victim is on top of some sort of stool, tripod vel sim., where he is being ‘offered’ to the numina invoked (contra Blänsdorf 2012e). At the victim’s sides, there are two serpentine monsters in profile, which have pointy ears and scales and may represent the spirits invoked that are asked to punish the spell’s victim (for parallels, see 22, 117–18).

Commentary: this defixio, discovered in the fountain of Anna Perenna (see 19 for the archaeological context), was found inside of a set of three containers (Inv. 475541), which was covered with a single flat lid at the time of discovery (for a discussion of these types of containers, see the commentary for 24). Unlike the sets of containers described in the previous entries for this catalog, all the vessels in this set were not inscribed. That said, within the set archaeologists found two tabellae defixionum (the second curse is 34) and a magical male figurine, which was found in the innermost container (C, measuring 6.3 × 4.8–4 cm [height × diameter]) and is made of bee’s wax. Furthermore, the figure is inscribed with magical symbols, Greek letters, and the name Petronius Cornigus, who is undoubtedly the victim of the curse (see Blänsdorf 2012e: 621 and 2015b: 27–29).

The Greek portion of the inscription, which according to Blänsdorf (2012e) is incomprehensible, is found on the upper part of the lead sheet and partially surrounds the individual and snakes. The Latin text is found in the lower portion of the tablet between the victim’s right leg and the snake prepared to strike. Blänsdorf has argued that this curse should be grouped among the erotic defixiones, though this identification is problematic due to the lack of information provided in the text. In addition, the transliterated Greek lexeme fantasia cannot be read in the published images of this curse.

Both the Latin text on the figurine and the one of the curse tablet under discussion were written in new Roman cursive, which allows us to date the set to either the fourth or fifth century CE. The tablet had been placed in the space between the outermost and middle containers (A, measuring 6.7 × 5.6– 5 cm; B, measuring 6.4 × 5.6–4.5 cm [height × diameter]). Though complete, the tablet has been damaged along its left side and in two places in the central third of the

After being inscribed, the tablet was placed between the outer and middle lead containers. During this process, the left edge of the curse was folded over the top of the middle container.

122

Italia 34. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th–5th centuries CE. [Inv. No.: 561971] Material: lead. Measurements: 5.4 × 9.1 × 0.03 cm. Reading: (charaktêres) Θω (charaktêres) Ων (charaktêres) RIEI+ECEITA (charaktêres) ΙΑΚΕΩΝ (charaktêres) ΛΩΙαωαΠΛ (?) ΑΙΦΑΨΟΤΩΙ Bibliography: Blänsdorf 2012e: 620–22; Piranomonte 2012a: 165–66; Blänsdorf 2015b: 27–29.

6.4 × 5.6–4.5 cm [height × diameter]) and innermost (C, measuring 6.3 × 4.8–4 cm [height × diameter]) of the three containers. This rectangular sheet is still intact and has the residue of resin on its left edge (presumably the same resin used to seal to outermost container). The inscription contains four lines, which consist of six series of charaktêres that are interspersed with Greek and Latin letters (capitals), whose meaning has defied interpretation. Although the text cannot be dated in palaeographic grounds, on analogy with the rest of the objects from the set, we can date this curse tablet to the fourth or fifth century CE.

Commentary: this defixio, discovered in the fountain of Anna Perenna (see 19 for the archaeological context), was found alongside 33 inside of a set of three containers (Inv. 475541), which was covered with a single flat lid at the time of discovery (for a discussion of these types of containers, see the commentary for 24). Unlike the sets of containers described in the previous entries of this catalog, all the vessels in this set were uninscribed. That said, within the set archaeologists found two tabellae defixionum and a magical male figurine (for more information, see the commentary for 33).

After being inscribed, the tablet was placed between the middle and innermost lead containers.

The defixio currently under consideration had been deposited in the space between the middle (B, measuring

35. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th–5th centuries CE. [Inv. No.: 475547] Material: lead. Measurements: vid. commentary. Reading: B (middle container) (charaktêres) Ablatanbla (charaktêr) C (innermost container) (crux) (on the figure’s abdomen) ᾿Ι(ησοῦς) Χ(ριστὸς) Ν(αζωραῖος) ὁ Υ(ἱός) Χ(ριστός) Ν(αζωραῖος) κ(αὶ) Θ(εὸς) Θ(εὸς) Θ(εός) Bibliography: Blänsdorf 2012e: 622; Németh 2012b; Piranomonte 2012a: 167–68; Németh 2015: 43 and 45.

of charaktêres and the palindrome ablatanbla (for ablanatanabla). The palindrome has been interpreted as an invocation of Abraxas, the rooster-headed demon with the body of a human, whom scholars have also identified as the figure depicted on the innermost container (C, measuring 7.9 × 5.9–4.3 Ø cm; for a different interpretation, see the commentary for 25). The palindrome was written in capitals, which precludes us from dating the text on palaeographic grounds. That said, and since all the containers from the fountain were probably made by a single specialist, it is possible to presume a Late Antique date (4th–5th centuries CE) for the group under consideration.

Commentary: this item, discovered in the fountain of Anna Perenna (see 19 for the archaeological context), is made of a set of three containers. At the time of discovery, the set was covered with a single conical lid with a knob (for a discussion of these types of containers, see the commentary for 24). Inside, an anthropomorphic male figurine was deposited, which in all likelihood represents the curse’s victim. Similar to other sets from Anna Perenna, the external container (A, measuring 8.6 × 7.7–5.4 Ø cm) is neither inscribed nor decorated, while the middle container (B, measuring 8.3 × 6–5.1 Ø cm) is inscribed with a series

Translation (Németh 2015: 45): ‘Ablanatanabla. Jesus Christ of Nazareth the son, Christ of Nazareth, and God, God, God.’ 123

Sylloge of  Defixiones from the Roman West 36. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th–5th centuries CE. [Inv. No.: 475551] Material: lead. Measurements: vid. commentary. Reading: B (middle container) Ablatanabla (charaktêres) C (innermost container) (on the right side of the figure) (charaktêres and magical symbols) (crux) (on the figure’s abdomen) ᾿Ι(ησοῦς) Χ(ριστὸς) Ν(αζωραῖος) ὁ Π(αῖς) Χ(ριστὸς) Ν(αζωραῖος) κ(αὶ) Θ(εὸς) Θ(εὸς) Θ(εός) Bibliography: Blänsdorf 2012e: 624; Németh 2012b; Piranomonte 2012a: 167–68; Németh 2015: 43 and 45.

inscribed nor decorated, while the middle container (B, 9.4 × 5.2–4.8 Ø cm) is inscribed with a series of charaktêres and the palindrome ablatanbla (for ablanatanabla). The palindrome has been interpreted as an invocation of Abraxas, the rooster-headed demon with the body of a human, whom scholars have also identified as the figure depicted on the innermost container (C, measuring 9.4 × 4.8–4.3 Ø cm; for a different interpretation, see the commentary for 25).

Commentary: this item, discovered in the fountain of Anna Perenna (see 19 for the archaeological context), is made of a set of three containers (Inv. 475551). At the time of discovery, the set was covered with a single flat lid with a knob (for a discussion of these types of containers, see the commentary for 24). Inside, an anthropomorphic male figurine was deposited, which in all likelihood represents the curse’s victim.

Translation (Németh 2015: 45): ‘Ablanatanabla. Jesus Christ of Nazareth the child, Christ of Nazareth, and God, God, God.’

Similar to other sets from Anna Perenna, the external container (A, measuring 9.7 × 6.3–5.4 Ø cm) is neither

37. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th–5th centuries CE. [Inv. No.: 475560] Material: lead. Measurements: 7.2 × 6.3 × 0.03 cm. Reading: 5



10

Marcel(l)iane sic oro [---?] in+ CR[---]TV++ se VLICAQR mort+ fali [---]V+SNIA d[o]m(i)n[---] F[-c.1-]R ut SA[-c.2]V[-c.2-] [-c.1-]amum il(l)u + si + [---]IC [---] CRA[---]TT++ se RI[---]A nulla (vacat) E[---] [---]SSIENEDASEL [---] BENO

Bibliography: Blänsdorf 2012e: 626.

written in new Roman cursive (measuring 0.3 cm in height). The text has been dated to the fourth or fifth century CE based on palaeographic grounds. We do not know the circumstances that led to the writing of this curse, which targets Marcel(l)ianus (for this cognomen, see Kajanto 19822: 173) and invokes a deity whose name is not known (l. 1: sic oro). According to the editor (2012e: 626), the word mort+ (l. 3) signals that the victim was

Commentary: this defixio, discovered in the fountain of Anna Perenna (see 19 for the archaeological context), was inscribed on a thin lead sheet. The item is almost intact (there are some small breaks along the right and bottom edges). The inscription, which is extremely fragmentary, contains 10 lines, which run from left to right and were 124

Italia 30–31 for some parallels and also 3). Here, it is worth noting how the curse is placed in a spot reserved for the wick. Following Mastrocinque (2007), the wick would represent the life of the victim, which was thus killed in a ritual of ‘symbolic homicide’, that is attested elsewhere in the fountain (on this, see 19, Mastrocinque 2007, Sánchez Natalías 2018: 14).

condemned to death, who apparently had no means (l. 8: nulla) of escaping said punishment. Though suggestive, this hypothesis cannot be confirmed given the fragmentary nature of the text. Furthermore, I was unable to confirm the reading of the text based on the published images. After being inscribed, the tablet was rolled up and placed inside of a lamp (Inv. 445038; cf. 19, 20 and

38. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th–5th centuries CE. [Inv. No.: 475570] Material: lead. Measurements: 9.3 × 10.3 × 0.1 cm. Reading: (on top of the figure’s head) ΑΠΟYΚΟΛΑΝΘΡΩΔΗΚΙΛΟΜΗΘ   (crux) (on the figure’s flank) Orsic ino Bibliography: Blänsdorf 2012d: 159; Blänsdorf 2012e: 634; Piranomonte 2015: 78.

an anthropomorphic figure, which is depicted in a reclining position along the tablet’s bottom edge. He looks outwards towards the reader with his arms (?) extended forward. His legs (?) are perpendicular to the rest of his body. On the side of the figure, we can read (in capitals) Orsicinus, which is surely the name of the represented individual (cf. 179). The name Orsicinus should probably be taken as Ursicinus (see OPEL IV, 186). Although Blänsdorf has not dated the curse, its iconographic features suggest an approximate date in the fourth or fifth century CE.

Commentary: this defixio, discovered in the fountain of Anna Perenna (see 19 for the archaeological context), was inscribed on an irregularly shaped lead sheet. The curse, which is intact, bears a three-line inscription written in a mix of Greek and Latin, which runs from left to right. The Greek portion of the text is found in the upper portion of the tablet, following the shape of the top edge and framing

39. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th–5th centuries CE. [Inv. No.: 475571] Material: lead. Measurements: 5 × 9.8 cm. Reading: 5

SV[-c.1-]OCNO [-c.1-]OMOS filiolos (magical symbols?)NO (magical symbols?) M[---]OS[---]ON[---]L[-c.1-]C[---] V[-c.1-]L[---]NA Turpione(m)

Bibliography: Blänsdorf 2012e: 634.

from left to right and is comprised of a Latin text (written in new Roman cursive) and two strings of magical symbols (perhaps charaktêres) both in l. 2. The curse’s sorry condition prevents us from learning anything about the circumstances that led to the writing of this text, which appears to target Turpio (l. 5, for this cognomen, see Kajanto 19822: 286).

Commentary: this defixio, discovered in the fountain of Anna Perenna (see 19 for the archaeological context), was inscribed on a roughly rectangular lead sheet. The tablet is extremely corroded and bears an inscription of five lines, which run

125

Sylloge of  Defixiones from the Roman West 40. Roma Provenance: sanctuary. Current Location: MNRTD. Date: unknown. [Inv. No.: 475720] Material: lead. Measurements: 8.5 × 5.4 cm. Reading:

Col. I Ρ ΟΑ ΘΘ VΝV

(crux)

Col. II Θ [---] Φ Ο Ν Ν Ρ Ω Υ

(crux)

Bibliography: Blänsdorf 2012e: 635.

four horizontal lines. To the left of the figure, a snake with pointy ears and scales is depicted in profile. The inscription, which runs from right to left, is composed in two columns. The texts comprise a series of Greek and Latin letters (capitals), whose meaning is opaque (perhaps we have an abbreviated formula similar to those found in 24–27 and 35–36?). Although the editor has not dated the curse, its iconographic features suggest an approximate date in the fourth or fifth century CE.

Commentary: this defixio, discovered in the fountain of Anna Perenna (see 19 for the archaeological context), was inscribed on an irregularly shaped lead sheet. This tablet, which is still folded, contains a drawing of a standing figure, whose head is covered by a crown with six beams. The figure’s body is shaped like a cone (perhaps the representation of a mummy) and is segmented by

41. Roma Provenance: sanctuary. Current Location: MNRTD. Date: 4th–5th centuries CE. [Inv. No.: 475569] Material: lead. Measurements: 7.3 × 12.6 × 0.1 cm. Reading:

S[---]VT[---]S[---]N

Bibliography: Blänsdorf 2012e: 638.

inscribed on a rectangular lead sheet. The inscription is extremely fragmentary and appears to have been written in new Roman cursive, which allows us to date the curse to the fourth or fifth century CE.

Commentary: this defixio, discovered in the fountain of Anna Perenna (see 19 for the archaeological context), was

42–47. Roma Provenance: sanctuary. Current Location: MNRTD. Date: vid. commentary. [Inv. No.: vid. commentary] Material: lead. Measurements: vid. commentary. Reading:

vid. commentary

Bibliography: Blänsdorf 2012e: 638–639.

43 [Inv. 475572]: irregularly shaped lead sheet measuring 4.4 × 7.2 × 0.1 cm. It contains the remains of an inscription and iconography.

Commentary: these six defixiones come from the fountain of Anna Perenna (see 19 for the archaeological context). Given the dearth of published information about these items, they appear here in a joint entry with their inventory numbers followed by relevant information.

44 [Inv. 548873]: rectangular lead sheet measuring 4.9 × 9 × 0.1 cm. It is very corroded, but the remains of an inscription and four holes (the result of being pierced) can still be made out.

42 [Inv. 475568]: irregularly shaped lead sheet of unknown dimensions.

45 [Inv. 569173]: copper sheet measuring 8.2 × 3 × 0.1 cm. It has been folded twice and holds the remains of organic material. For other parallels of copper defixiones, see 30–31. 126

Italia 46 [Inv. 561974]: roughly rectangular lead sheet measuring 9.2 × 12.1 × 0.6 cm and weighing 409 gr. Apparently the tablet is not inscribed.

47 [Inv. 561975]: roughly rectangular lead sheet measuring 6 × 8 cm. Apparently the tablet is not inscribed.

48. Roma Provenance: necropolis. Current Location: SSBAR. Date: 1st century CE. [Inv. No.: 500240] Material: lead. Measurements: 17.5 × 12.8 cm. Reading: A Dite pater Proserpina Dia Canes Orcini Ustores Inferi Ossufragae Larvae Furiae Maniae Aves Nocturnae Aves Harpyiae Ortygiae Virga Ximaera Geryones Siredonas Circe Gegantes Sp(h)inx vos precatur et 5 petit rogat vos numina deum inferum qui suprascripti estis ea[m] Caeciliam Primam sive quo alio nomine est uti eam Dite Pater deprimas malisque doloribus eam adpetas aput te abducas Proserpina Dia tu facias illam Caeciliam Primam 10 sive quo alio nomine est uti eam deprimas adimas illae sanguinem de venis corpus calorem animi illae Caeciliae Primae eripias Canes Or[ci]ni Orcini tricipites vos illius Caeciliae Primae exedit[is] iocinera pulmones cor cum venis viscera membra medullas 15 eius diripiatis dilaceretis lumina eius C[a]e[c]ilia P[ri]mae aripiatis vosque Ustores inferi eius Caeciliae {S?}Primae peruratis lumina stomachum cor eius pulmones adipes cetera membra omnia illius Caeciliae Primae peruratis eduratis vos neque vivere nec valere 20 possit eamque Caeciliam Primam ad vos adducatis tradatis ea sua Ossufragis inferis vos illae Caeciliae Primae ossum frangant medullas exedint iocinera pulmones dirimant vosque Ossufragae inferae tradatis illam Caeciliam 25 Primam Aurorae Orchi sorori Aurora Orchi soror tu illae Caeciliae Primae eripias somnum soporem obiicias illae amentiam dolores stupores malam frontem (vacat?) usque donec Caecilia Prima pereat intereat extabescat deinde tu Aurora 30 Orcini soror tradas illam Caeciliam Primam Larvabus et Furiabus inferis (vacat?) illae Caeciliae Primae obiciatis metum formidines dolores B stupor et ameṇṭịạ ọṃṇịsqụẹ D[-c.8-] faciaṇt \ead[e]ṃ[-c.2-]quaẹ sibi[-c.8-]A[-c.4-]adipiscatur >turpidines P[-c.6-]MA[-c.6-]ananantur eam \Caeciliam Primam [-c.5-]AM[-c.4-6-]ea habeat semper 5 turpid(in)es ut{i} illa Caecilia Prima ạb inimicis suis prematur opprima[tur] desụṃa[tu]r ṇẹque [-c.2-]ADATVR Aves Nocturnae Aves Harpyiae vos illius Caeciliae Primae exeditis c ̣ọ[r] ṃạ[n]uṣ intera[nea] omnia obiciatis illae Caeciliae Primae ọṃṇịbụṣ[-c.8-]ạia 10 aegraque consilia febres cotidianas tertianas quartanas usque dum animam eius Caeciliae Primae eripiatis[-c.3-]CAE QVE[-c.3-]mas[-c.4-]S[-c.1-]IA Virga deum inferum[-c.14-]ac[-c.6-]a tu efficias illae Caeciliae Primae quemadmodum 127

Sylloge of  Defixiones from the Roman West 15



20



25



30



35

tu domas caedis uris [per]uris ad Inferos eos qui ad superos omnia mala sc[el]eraque fecerunt sic tu Caeciliam Primam [ill]aṃ Ṿirga uras peruras caedas domes do[n]ec ṭạṃqu[a]m ạḷ ọraṃ egentes omnibus fortunis mortua[m] ad t[e] ạbducas ////Himaera (vacat?) Caeciliam Primam morsu mordeas ///uti sana nunquam fiat adil[-c.2-3?-]illam F[-c.2-3?-]EM //uti turgeas sic tamquam ṭyṃ[p]anum doloribusque malis intereat usque done ea pẹrẹạṭ uti abducas ṇum qua[m] ea[-c.4-] eam Caeciliam Primam perura[s -c.15-]EM[-c.10-]PRM mortua fiat [-c.4-]A Geryonimes Sirẹdones Circe Solis filia quemadmodum Minerva una tunica [---]E[---]ABAT eos ṃonstrinae Siredenae cantibus homines detinebant Circe ferialis medicamentis Ulixis socios [---]a[---]INEM /[-c.8-]E[-c.1-]ATIS Caeciliam Primam isdem malis isdem [-c.1-]ọḷọris Caeciliam adiuvatis vitamque animam eius [-c.3-4?-]CARGAMA (?) eripiatis inferis tradatis SS[-c.3-4?-] Dite pater Proserpina dia et Virga vos precatur is qui hoc ma[n]datum dedit vestram [---]a [---]faṇọ DETES[---]M[---]eam perficiatis perfectae pati[-c.4-]

128

Italia

Bibliography: Bevilacqua 2006; Bevilacqua 2009; Bevilacqua and Colacicchi 2009: 303–38; AE 2007, 260; AE 2008: 153; Bevilacqua 2010b; Bevilacqua 2012: 614– 16; Watson 2019: 64–70, 76.

This curse, which is directed against Caecilia Prima, has a tripartite structure, which Bevilacqua has compared to typical prayers consisting in an invocation, prayer and request. In the first part, the defigens summons eighteen mythical beings, beginning with Dis Pater and Proserpina (who are also invoked in 10–14, 103 and 160). Next, there are some monstrous beings that are not attested elsewhere in the corpus of Latin defixiones: Ustores inferi (body burners), Ossifragae (bone breakers), Furiae, Maniae, Aves Nocturnae, Aves Harpyiae, Ortigyae (another type of predatory bird), Virga (invoked as a personification of the caduceum, according to the editor 2009: 312), Ximaera (for Chimera), Geryon, Siredonas (Sirens), Circe, Gigantes, Spinx (for Sphinx), Orcini tricipites (the dogs of Orcus) and Aurora (here called the Orchi soror (see Bevilacqua 2009 and 2010b: 81–83). The defigens asks that the invoked beings utterly destroy Caecilia Prima, who is named no less than 24 times in the text.

Image source: Bevilacqua 2006–07: 306–07, figs. 1 and 2. Courtesy of G. Bevilacqua. Commentary: this defixio was discovered 2003 during the excavations carried out near a mausoleum on the Via Ostiensis (modern-day Via Bompiani), which has been dated to the first century CE. The curse was written on a roughly rectangular lead sheet, which is fully intact despite corrosion on the central and right portions of side B. Furthermore, there are two small, damaged areas at ll. 4 and 9–10. An opisthograph, the inscription contains 67 lines, which run from left to right and were mostly written in scriptio continua. The capitals in which the text was written date to the second half of the first century CE. 129

Sylloge of  Defixiones from the Roman West The literary, linguistic and palaeographic features of the inscription suggest that the author came from the upper echelons of Roman society, knew the literature of his time and was quite familiar with writing. The text nevertheless presents the following orthographic, morphological, and phonetic features: Ossufragae for Ossifragae (A, l. 2), Gegantes for Gigantes, Spinx for Sphinx (both in A, l. 4), inferum for inferorum (A, l. 5), corpus for corporis (A, l. 11), illae for illi (A, ll. 12, 22 and 26; B, ll. 9 and 14), eduratis for aduratis (A, l. 19), turpidines and turpides for turpitudines (B. ll. 3 and 5), and uti for ut (B, l. 5). The text’s diction is also noteworthy. So far, there are no parallels for the following lexemes in the defixiones from the Roman West: adpeto (A, l. 8), exedere (A, l. 23 and B, l. 8), stomachum (A, l. 17), adipes (A, l. 18), turpidines (B, ll. 3 and 5, from turpitudo, see below) and monstrinae (B, l. 28). Furthermore, we find several verbs commonly found in poetry: adimo (A, l. 11, cf. Naev. fr. com. 95; Plaut. Epid. 363), diripio and dilacero (A, l. 15, see TLL s.v. diripio and dilacero, pp. 1260 and 1159, respectively). Bevilacqua has connected several of the author’s phrases to known literary works, such as obiciatis metum formidines dolores (A, l. 32; cf. Macr., Sat. III, 9, 7–8: eique populo civitati metum formidinem oblivionem iniciatis), aegraque consilia (B, l. 10; cf. Sall. H. IV, 84: consilii aeger), stupor et ameṇṭịạ (B, l. 1; cf. Quint., Declam. Maior XIX, 11.28: verumtamen stupore, amentia, et in silentium orbitate defeci), etc. Finally, the defigens betrays a familiarity with the writing practices through the use of several diacritics, like oblique strokes (B, ll. 2, 4, 20–22 and 28), diples (B, l. 3) and underlining (B, ll. 7ff.).

her below and assail her with terrible pangs, carry her off to your home. Divine Proserpina, may you bring it about that Caecilia Prima, or whatever other name she goes under, you send her below, take the blood from her veins, steal from that Caecilia Prima her flesh, the heat of her mind from that Caecilia Prima. Hounds of Orcus, Three-headed Creatures of Orcus, may you consume that Caecilia Prima’s liver and lungs, may you tear and rend her heart and veins, entrails, limbs, marrow, may you snatch away the eyes of Caecilia Prima and you, infernal Burners, may you burn and scorch all the other limbs of that Caecilia Prima, you, and let her not be able to live or enjoy good health, and may you carry off that Caecilia Prima to yourselves, hand over those parts of her to the infernal bone-breakers, you, let them break the bones of that Caecilia Prima, let them consume her marrow, let them tear apart her liver and lungs, and may you, infernal Bone-Breakers, deliver over that Caecilia Prima to Aurora, sister of Orcus. Aurora, sister of Orcus, may you take away sleep and slumber from that Caecilia Prima, may you inflict on her madness, pangs, stupefaction, a distressed bow (?)... up to the time when Caecilia Prima dies, perishes, wastes away; then may you, Aurora, sister of Orcinus, deliver over that Caecilia Prima to the infernal Ghosts and Furies... may you [plural] inflict upon that Cecilia Prima fear, frights, pangs. Stupefaction and madness and all ...let them bring about, the same... who for herself... let her get, immoral behaviour ...that Caecilia Prima... let her always have. Immoral behaviour, that that Caecilia Prima may be oppressed, crushed by her enemies, consumed (by illness) and not... Birds of the night, Harpy birds, you, may you consume the heart of that Caecilia Prima, her hands, all her intestines, may you inflict upon all... of that Caecilia Prima anguished thoughts (?), daily, tertian and quartan fevers, until the time when you wrest away the life of that... Virga, of the infernal deities... you [singular], may you bring it about that Caecilia Prima, whom you are quite subduing, lashing, burning, scorching among those persons in the Underworld who have done all kinds of wickedness and crimes in the world above, that even so you, Virga, burn, scorch, lash, subdue that Caecilia Prima until you carry her off dead to you, just like persons upon a shore who have lost all their possessions. Himaera... may you sink your teeth into Caecilia Prima ... so that she never becomes healthy... her... may you (Caecilia Prima) swell up just like a drum, and may she die from terrible pangs until she perishes that you carry off... never... that Caecilia Prima you should scorch... let her be dead... Geryones, Sirens, Circe the dauther of the Sun, just as Minerva with her tunic alone... them, the Sirens, female monsters, detained men with their songs and Circe with her deadly drugs the companions of Ulysses... may you... that Caecilia Prima with the same evils, the same pangs you ‘help’ Caecilia, and may you her life and spirit... take away, may you deliver to the infernal... Dis Pater, divine Proserpina and Virga, he beseeches you who charges you in this matter your [plural]... shrine... her... may you bring to perfection of the perfect...’

The circumstances that led to the writing of this curse are not stated explicitly by its author, who was a man (note B, l. 33: vos precatur is qui..., see Watson 2019: 76). Nevertheless, the editor has rightly argued that the tablet likely belongs to the group of erotic defixiones (Bevilacqua 2009: 78), given the figurative dismemberment of the victim, the repeated use of the verbs uro, peruro and aduro (A, ll. 17–19 and B, l. 17; to make the victim burst into flames) as well as the term turpitudo (B, l. 3 and l. 5, ‘obscenity’). The invocation of Aurora fits well in this context, given that she was believed to be ‘cacciatrice di giovani amanti e distruttrice delle unioni amorose’ (Bevilacqua 2009: 327). After inscribing the tablet, the practitioner attempted to fold the tablet and pierce it with a nail but did not manage to succeed on either front. We can still see the beginnings of the fold line and dents from the nails (A, ll. 9–10). Translation (Watson 2019: 64–65, slightly modified): ‘Dis Pater, divine Proserpina, Hounds of Orcus, infernal Burners, Bonebreakers, Ghosts, Furies, Spirits of Madness, Birds of the night, Harpy birds, Ortygian birds, Virga, the Chimaera, Geryones, Sirens, Circe, the Giants, the Sphinx, he beseeches you and requests, asks you, powers of the infernal deities, you whose names are written above: that that Caecilia Prima, or whatever other name she goes under, you, Dis Pater, should send 130

Italia Regio I [Latium] 49. Nomentum, Mentana Provenance: necropolis. Current Location: MNRTD. Date: 1st century BCE. [Inv. No.: 52185] Material: lead. Measurements: 7.2 × 7.1 cm. Reading: 5



T(itus) · Octavius · T(iti) · l(ibertus) P(ublius) · Fidustius · . . Poṣ[tumius? or -tumia?] (traces) Cavia (vacat) si qui · arṿosarius auṭ[ ·] arvosaria

Bibliography: Borsari 1901: 210; AE 1901, 185; DT 133; Jeanneret 1917: no. 133; Solin 1989: 199–200; Solin 1995: 573; Kropp 2008: no. 1.4.2/1; Pettirossi 2016: no. 10; Urbanová 2018: no. 10.

consisting of a small vertical stroke found halfway up the letters. As Audollent pointed out, this text should be grouped among the juridical defixiones. This curse shares at least one victim with 50, where we also find T(itus) Octavius together with another Fidustius. The absence of any cognomina supports a Late-Republican date for this text. Solin (1989: 198) has proposed that Pos[tumius] or Pos[tumia] Gabia (ll. 4–5; note that my autopsy supports the reading Cavia contra Solin’s Gabia) is the defigens. While this is suggestive, Solin’s proposal is unlikely, since this onomastic formula is in the nominative like the names found in ll. 1–2. Furthermore, the final phrase presents a sort of all-inclusive formula to include anyone else who might be an arṿosarius or arvosaria (ll. 5–6), which Solin (1989: 199) rightly has taken for adversarius and adversaria. Besides in these two curses, the nomen Fidustius is only attested in Latium and Etruria.

Image source: Borsari 1901: 210. Commentary: this defixio was discovered c. 1900 during the excavations of a necropolis in Mentana, which contained various incineration graves. According to the editor, two of the urns contained four lead curse tablets that were corroded and folded and placed among the remaining bones (1901: 207). The curse was written on a square lead sheet, which at the time of my autopsy (October 2010) was broken along the central area and missing a good part of the bottom edge (including the left corner). The surface, uneven in the centre along the fold line, has several reddish spots caused by corrosion. The inscription contains six lines of capitals (measuring between 0.3 and 1 cm in height) that are very similar in shape and size to those found in 50–51. The text runs from left to right, though to read the final four lines, which are separated from ll. 1–2 with two dots, the tablet must be rotated 180 degrees (for a parallel, see 526). The words are separated with interpunctuation

After being inscribed, the tablet was folded in half and pierced with a nail. Translation: ‘Titus Octavius, freeman of Titus, Publius Fidustius, Postumius (or Postumia?) Cavia, and any (male or female) opponent.’

131

Sylloge of  Defixiones from the Roman West 50. Nomentum, Mentana Provenance: necropolis. Current Location: MNRTD. Date: 1st century BCE. [Inv. No.: 52183 and 52184] Material: lead. Measurements: 8.5 × 11.5 cm. Reading: A T(itus) · Octavius · sermoṇe(s) M(arcus) · Fidustius {v} · mutus sermone(s) [·] Fidustium (vacat) [m]uṭus 5 Ịrẹna [·] Plotiaes dficere · ex[t]am [·] ume[r]o(s) nesu(m) · quaeṣtu(m) [·] caput [·] OLAṾS d[e]scribo [·] ocilos · ex · eị

B mem(b)ra omnia laṭus li(n)cua(m) · ilaṭu(m) c ̣onnatụs ex anu novọ cres ex̣ (h)oc TAND MOREO il(l)a con`ano´ 5 MATIE bonus QVADRIN vesticia ilaṭus FACI latụs bona+ RAMAT Ṣ il(l)ae APA NOPRV ṇec OLI+NNIN

Bibliography: Borsari 1901: 208–09; AE 1901, 184; DT 134; Jeanneret 1917: no. 134; Solin 1989: 197–99; Solin 1995: 572–73; Kropp 2008: no. 1.4.2/2; Pettirossi 2016: no. 9; Urbanová 2018: no. 11.

interpunctuation consisting of a small vertical stroke or a single dot found halfway up the letters. This curse should be grouped among the juridical defixiones, as can be deduced from the words mutus and sermone(s) (A, ll. 1–4). The victims may be the same as those found in 49. That said, in the previous curse we find P(ublius) Fidustius, while this text has M(arcus) Fidustius (they were probably members of the same family) and we also find a new victim, the slave Irena Plotia. The text presents the following orthographic and phonetic features: Fidustium for Fidustius (A, l. 3), dficere for defigere (A, l. 6), nesu for nisum, quaestu for quaestum, memra for membra (B, l. 1), licua for linguam (B, l. 2), ilatu for flatum (B, ll. 2 and 6), anu for anno (B, l. 3), oc for hoc (B, l. 4), ila/ilae for illa/ illae (B, ll. 4 and 7), vesticia for vestigia (B, l. 6). Kropp has reconstructed olaṿs (A, l. 7) as oculos, though this seems more appropriate for the following line where we find ocilos. The phrase ex anu novọ (B, l. 3) should be taken as ex an(n)o novọ instead of Audollent’s proposed Crescon(i)ano. The final lines of the text are legible but cannot be satisfactorily explained. Besides these tablets, Fidustius is only attested in Latium and Etruria.

Image source: Borsari 1901: 208–09. Commentary: this defixio was discovered c. 1900 during the excavations of a necropolis in Mentana, which contained various incineration graves. According to Borsari (1901: 207), two of the cinerary urns contained four curse tablets, one of them being the present one. The curse was written on a rectangular lead sheet, which is now broken into two parts that fit together. After its most recent restoration, the curse was mounted on a methacrylate base that runs along the curse’s perimeter so that the majority of both sides can still be seen. The tablet is in fairly good condition, although some areas have been corroded (especially on side B, where some text has been lost, as I confirmed during an autopsy in October 2010). An opisthograph, the inscription contains 16 lines, which run from left to right. The text was written in capitals that are very similar in shape and size (measuring between 0.4 and 1 cm) to those found in 49. This strongly suggests that the same hand inscribed both texts. The words on side A (and occasionally on side B) are separated with

After being inscribed, the tablet was pierced three times and perhaps folded in half, which could have caused the present break. 132

Italia Translation (Urbanová 2018: 237–38, modified): ‘(May) Titus Octavius (be deprived) of speech, (may) Marcus Fidustius (be) deprived of speech, (may) Fidustius (be)

mute, Irena (the slave) of Plautia, I curse (her/their?) guts, shoulders, gait, profit/business, head, (...) I curse the eyes(?), all limbs: hip, tongue, breath,...’

51. Nomentum, Mentana Provenance: necropolis. Current Location: MNRTD. Date: 1st century BCE. [Inv. No.: 52186] Material: lead. Measurements: 5.2 × (9.2) cm. Reading: A Malcio · Nicones · oculos manus · dicitos · brac(h)ias · uncis · capil(l)o(s) · caput · pedes · femus · venter ṇatis · umlicus · pectus · mamil(l)as 5 collus · os · buc(c)as · dentes · labias me[nt]us oclos · fronte(m) · supercili scalas · (h)umerum · nervias · ossu(m) merilas · venter [·] mentula(m) · crus qua(e)stu(m) · lucru(m) · valetudines · defico 10 in (h)as · tabel(l)as · B Rufa · pulica · manus · de(n)teṣ oclos [·] brac(h)ia [·] venter [·] mamil(l)a(s) pẹc ̣ṭụṣ · os(s)u(m) [·] mirilas [·] venter [---]cruṣ os · pedes · frontes 5 uncis · dicitos · venter umlicus · cun(n)us · quas · il(l)ae [·] Rufaṣ (!) [·] pulica(e) [·] difcọ in (h)as [·] ṭabel(l)as Bibliography: Borsari 1901: 207–08; AE 1901, 183; DT 135; Niedermann 1906: 962–63; Audollent 1909: 367–68; Diehl 1910: no. 851; ILS, 8751; Jeanneret 1917: no. 135; Rubio and Bejarano 1955: no. 184; Solin 1989: 196–97; Gager 1992: no. 80; Linderski 1993: 203–04; Solin 1995: 571–72; Versnel 1998: 223; Cimarosti 2005: 452–53; Kropp 2008: no. 1.4.2/3; Pettirossi 2016: no. 8; Urbanová 2018: no. 12; Lucciani and Urbanová 2019.

written in capitals, which measure between 0.2 and 0.8 cm in height, are palaeographically quite similar to those found in 49–50, which suggests that the same hand wrote all three inscriptions. We do not know the circumstances that led to the writing of this curse, although for Audollent (1909: 367) it belongs among the group of erotic defixiones. The text presents the following orthographic, morphological and phonetic features: dicitos for digitos (A, l. 2 and B, l. 5), bracias for brachias (A, l. 2 and B, l. 2), uncis for ungues (A, l. 2 and B, l. 5), capilo for capillus (A, l. 3), venter for ventrem (A, ll. 3, 9 and B, ll. 2, 5), umlicus for umbilicus (A, l. 4 and B, l. 6), mamilas for mamillas (A, l. 4 and B, l. 2), oclos for oculos (A, l. 6 and B, l. 2), scalas por scapulam (A, l. 7), umerum por humerum (cara A, l. 7), quastu por quaestum (A, l. 9), detes for dentes (B, l. 1), pulica for publica (B, l. 1), osu for ossum (B, l. 3), cunu for cunnum (B, l. 6), pulicae for publicae (B, l. 7), difco for defigo (B, l. 8), as for has and tabelas for tabellas (both in A, l. 10 and B, l. 8). Niedermann has interpreted the lexemes merilas and mirilas (A, l. 9 and B, l. 3) as ‘Vulgar’ forms of medullas (see Linderski as well).

Image source: Borsari 1901: 207–08. Commentary: this defixio was discovered c. 1900 during the excavation of a necropolis in Metana, which contained various incineration graves. According to the editor (1901: 207), the tablet was found among the burnt remains of a cinerary urn. The curse, which I examined during an autopsy in October 2010, was written on a rectangular lead sheet, which has lost its bottom right corner and a good deal of the left edge. The surface, which was hammered out before being inscribed, is in good condition, despite some corrosion. An opisthograph, the inscription contains 18 lines, which run from left to right. The words on side A (and occasionally on side B) are separated with interpunctuation consisting of a single dot found halfway up the letters. The text is

The reading of the beginning of  B, l. 7 has proven somewhat controversial: while Borsari read quas il(l)ae (B, l. 7, as 133

Sylloge of  Defixiones from the Roman West We should accept Solin’s suggestion (1989 and 1995) that Malcio Nicones be taken as Malchio son of Nico (instead of Malcius Niconae (filius)). He also rightly argues that Rufa (B, ll. 1 and 7) ought to be interpreted as Rufa publica, a public slave (1995: 572; on this, see specially Cimarosti 2005 and Lucciani 2019: 432-442). Rufas (B, l. 7) should be taken as the possessive dative Rufae in agreement with illae and publicae.

did I during my autopsy, with an interpunct separating the words), Audollent (and later Gager 1992: no. 80, Solin 1995: 571, Versnel 1998: 223, Urbanová 2019: 428) proposed the reading [v]ulva(m). Furthermore, Cimarosti (2005: 452f., notes 14 and 2) reads quasil(l)a{e}(ria), a reference to the slave’s duties, while Kropp (2008: no. 1.4.2/3) reads quastum. While the latter two readings do not take into account the interpunct that divides the two words, the first proposal, ([v]ulva(m)), is impossible for several reasons. First (and most importantly), it ignores the fact that that same corner is preserved in full, which means that the text is also fully preserved and, consequently, that the reconstruction of [v] is artificial. In addition, the ul that should be following the imaginary [v], presents some palaeographical problems: compare this l with the ones found in the rest of the text (which never have a horizontal stroke on top of the vertical stroke). Thus, the reading quas · il(l)ae (B, l. 7) is secure. That said, it is a matter of debate how quas should be analysed: it could either be taken as qua(e)s[tum] (cf. A, l. 9) or the relative pronoun, though its gender does not agree with the immediate antecedents. Finally, note the phrase in (h)as ṭabel(l)as used to refer to the curse tablet (cf. A, l. 10 and B, l. 8).

After being inscribed, the tablet was pierced with a nail on at least one occasion and then likely folded in half. Translation (Versnel 1998, modified): ‘Malchio son/ slave of Nikon: his eyes, hands, fingers, arms, nails, hair, head, feet, thigh, belly, buttocks, navel, chest, nipples, neck, mouth, cheeks, teeth, lips, chin, eyes, forehead, eyebrows, shoulder-blades, shoulders, sinews, guts, marrow (?), belly, cock, leg, trade, income, health, I do curse in this tablet. Rufa the public slave: her hands, teeth, eyes, arms, belly, breasts, chest, bones, marrow (?), belly, guts, mouth,… feet, forehead, nails, fingers, womb, navel, cunt: these things belong to Rufa the public slave I do curse in this tablet.’

52. Ostia Provenance: necropolis. Current Location: unknown. Date: 2nd–1st centuries BCE. Material: lead. Measurements: 10.5 × 10.5 cm. Reading: 5



Agathemeris ··Manliae · ser(va) [Ac]hụlea · Fabiae · ser(va) · ornatrix [Ca]letuche · Vergiliae ser(va) ornatrix Hilara · Liciniae [ser(va) orn]atrix C˹hr˺este · Corn[eliae] ser(va) · ornatrix Hilara · Seiae · ser(va) ornatrix Moscis · ornatrix Rufa · Apeiliae · ser(va) ornatrix Chila · ornatrix

Commentary: this defixio was discovered in 1911 during the excavations carried out by Vaglieri in the necropolis of Ostia, specifically in tomb no. 1, which dates to the Republican period and contained 30 cinerary urns (see Vaglieri 1911: 84ff. and Squarciapino 1958: 21). The curse tablet was found folded like a diptych inside “urn f” (Vaglieri 1911: 87).

Bibliography: Vaglieri 1911: 87; AE 1911, 195; Besnier 1920: no. 32; CIL XIV, 5306; Squarciapino 1958: 21, 55, n. 47 and 137; García Ruiz 1967: no. 63; Kropp 2008: no. 1.4.3/1; Urbanová 2018: no. 13. Image source: Vaglieri 1911: 87, fig. 6. 134

Italia The curse was written on a roughly square lead sheet, which is now missing its two left corners and a good deal of the left edge. Furthermore, there is damage at the lower-right corner and six holes, which probably resulted from the piercing of the tablet after it was inscribed. The inscription contains nine lines of capitals, which run from left to right. The words are separated with interpunctuation consisting of a small vertical stroke (e.g., l. 5) or a single dot (e.g., l. 6) found halfway up the letters. We do not know the circumstances that led to the writing of this curse, which targets a group of women, mostly slaves (except for Moscis and Chilas) and ornatrices (that is, ‘a servant who assists at a person’s toilet, dresses the hair, or sim.’, OLD s.v. 1). This is the only known curse from the Roman West to target such a group of women.

The text presents the following orthographic and phonetic features: Achulea for Achillea (l. 2), Caletuche for Caletyche (l. 3), Crheste for Chreste (l. 5), Moscis for Moschis (l. 7) and Apeiliae for Apiliae (l. 9). Most of the victims had Greek names: Agathemeris (l. 1; cf. Solin 2003: 6 and 1453), [Ac]hulea (l. 2), Caletuche (l. 3; cf. Solin 2003: 103), Chreste (l. 5; cf. Solin 2003: 1006), Moscis (l. 7; cf. Solin 2003: 1142) and Chila (l. 9; cf. Solin 2003: 712). Hilara and Rufa are well attested Latin cognomina (ll. 4, 6 and 8) After being inscribed, the tablet was folded in half and, in the editor’s words, ‘presenta cinque fori per il passaggio del filo che la chiudeva’ (1911: 87). Though Vaglieri does not say so explicitly, it is safe to assume that the ‘thread’ ‘sewn’ through the tablet was made of lead (for parallels, see 61 and 71).

53. Ostia Provenance: necropolis. Current Location: Ostiense Museum. Date: 2nd century CE. Material: lead. Measurements: (11) × (23.5) × 0.2 cm. Reading: A [---]++[inf]ẹrịṣ [lig]o Tucia Crecta e(t) Antonia [---] A[-c.2-]++ e(t) (?) Silio e(t) [---] D[e]xtera mater [---] DN 5 [---] OC [-c.1-]C [-c.1-]NO Mnetus peri(ant) rogo Icimas Mẹvia [---] R+[-c.2-]A[-c.1-]ER+PE[-c.2-]lus oc(c)idạ(nt) Ru(fa?) Pap(i)ria Unaesi(me) [---] V+[-c.7-] LV[-c.2-]iṭine Lu[p]us [---] NNO[-c.3-]++R+[-c.1-]A col(l)ico mendẹs [---]++TIT [-c.1-]E Moṇtạ +LCE tabes(cant) Primigenia e(t) 10 [---]RM+++ACE [Re]ṣtuti Res{i}pecti Iulia F̣ortnata [---]O[-c.3-] (vacat) +N Epia Scaṇṭ[illa?] Iuḷia Pia pẹrị(ant) [---]++O[-c.8-9-]NTRI+[-c.9-]M[-c.4-]T[---]

135

Sylloge of  Defixiones from the Roman West B Col. I Iucunda Ṃẹvia E[-c.1-]+CE



Col. II Mucạṇ[a] Iucunda [M]aevia Procla Iu[li]a [T]ripe(a) Prote Minucia M[---]

Bibliography: Solin 1968: 3–22, no. 37; Pisani 1969: 369– 70; AE 1995, 247; Kropp 2008: no. 1.4.3/2; Urbanová 2018: no. 14.



Col. III Cuarta ẹṭ F̣adia Nereis PO+[-c.2-]STIDIA Aur(elia) Ṃ[u]sario [---]AEP[---]

and destruction: peria(nt), oc(c)idạ(nt), tabes(cant) and again pẹrị(ant), (A, ll. 5, 6, 9 and 11, respectively). The text presents the following orthographic and phonetic features: Tucia for Tuccia (A, l. 1), ocidant for occidant (A, l. 6), peria(nt) for pereant (A, l. 5 and 11), Papria for Papiria (A, l. 6), Unaesime for Onesime (A, l. 6), colico for colligo (A, l. 8), mendes for mentes (A, l. 8; Pisani 1969: 370 has connected this to a southern pronunciation), Resipectus for Respectus (A, l. 10) and [T]ripe(a) for Trebia (B, II, l. 4).

Image source: Solin 1968: 11–12. Courtesy of H. Solin Commentary: this defixio was discovered still folded up in a tomb dated to c. 100 CE that is located along the Via Ostiensis (nearby the so-called ‘Porta Romana’) during the excavations carried out in 1954. The curse was written on an irregularly shaped lead sheet, which today is broken into seven fragments that fit together (the breaks coincide with the fold lines). The surfaces of the first three fragments are quite damaged, while the final four are much better preserved. An opisthograph, the inscription contains 16 lines, which run from left to right. The lines on side B were arranged in three columns. A trained hand wrote the text in capitals whose height ranges between 0.1 and 0.3 cm. Based on a palaeographic analysis, the text should be dated to the second century CE.

Based on an onomastic analysis, the curse’s victims are in their majority slaves or freedwomen with Greek cognomina. Solin has suggested the reconstructions Monta[nus or -na] (A, l. 9) and Pom[peia] or Pom[ponia] (B, III, l. 3). Note the inversion of the normal sequence of nomen-cognomen in the cases of Icimas Mevia (A, l. 5; Icimas may be for Ichmas) and Iucunda Mevia (B, I, ll. 1–2 and II, l. 2). Of the male names in the text, Mnestus (A, l. 5; cf. AE 2006, 1825) is certainly a slave name, while we cannot be so sure about Silio (A, l. 2).

We do not know the circumstances that led to the writing of this curse, which probably contained an invocation of the infernal gods ([inf]ẹrịṣ [lig]o, A, l. 1), before presenting a list of personal names interspersed with verbs of death

After being inscribed, the curse was probably rolled around a nail, which was protruding from the tablet at the time of discovery (for a parallel, 109 and also Solin 1968, table I).

136

Italia 54. Ostia Provenance: unknown. Current Location: private collection. Date: second half of the 1st century BCE. Material: lead. Measurements: 9.2 × (12.8) × 0.08–0.13 cm. Reading: 5



10 10a



11a



15



[Axia --]I Axi lẹiberta · Gymnasio · Leontinis [mater Axia] [--- Axi] leiberta Leontis · Marcus Vibius · Dio[---] Mar]c ̣us Vibius Quincti Luci Marci leibertus Meṇ[ophilus] [G]alla Vibia Paciaecorum · leiberta · Menophili [uxor] [H]eracleo · Gallae Vibiae filius · et Arco Gallae Vi[biae filius] et Menophilus Gallae Vibiae filius · Anulla · Cạe[por filius?] Secunda · Caeporis · uxor Secundae Caeporis ux̣[oris est pa-] ter · Heracleo · Lucius Vibius `Cripo´ et uxor est Secun[da---] [S?]anucia · uxor · est · illius · Q(uintus) Vibius Tuscus l[eibertus] [P]aciaecorum Attin illius · uxor est Pubḷ[i---] L(ucius) Fulvius Numeri · f(ilius) · Seve[ra? uxo]r or [mate]r (?) et Marcione fratre illius Luipor Paciae[corum leibertus] Numri (?) est [u]xor · est · illius Flora et filius est Luipor · Q(uintus) F[abius---] Annia mater est Fab[ia]e Luxininae Fabi[us---] Anniae filius · Sc[aevol]a Anniae · pater [------] Id[i]e ater · Anniae Scaevola quei [est frater Anniae?] Q(uintus) Axius pater puer er[u]m (traces)

Bibliography: Weiss 2016; AE 2016, 197.

an irregularly shaped lead tablet, which today is somewhat curved and has lost its right edge as well as part of the left edge (the editor has calculated that the tablet originally measured 15.5 cm in length). Though the surface is wrinkled at the top-left corner and has been corroded along the last third of ll. 10–11, the tablet is in relatively good condition. There is also a hole in the centre of the inscription at the height of ll. 13–15.

Image source: Weiss 2016: 229, fig. 2. Courtesy of P. Weiss. Commentary: this defixio was discovered at an unknown date and under unknown circumstances (it was acquired on the art market) and comes from Ostia. It was written on 137

Sylloge of  Defixiones from the Roman West The inscription contains 18 lines of capitals which run from left to right measuring 0.3 cm in height (though note that the added text in between the ll. 7–8 and 11–13 measure 0.15–0.2 and 2–2.5, respectively). The words are occasionally separated with interpunctuation found halfway up the letters. We do not know the circumstances that led to the writing of this curse, which consists of a list of more than 30 names, most of which are in the nominative case. So far, this curse has the longest known list of names in a Latin defixio. Unlike in other lists found in curses, the defigens here defined his victims in relation to their family members (parents, spouses, etc.) in phrases containing the verb est. In the editor’s mind, this structure combined with the use of the pronoun illius provides evidence for colloquial Latin.

with the exception of Id[i]e (l. 15), which the editor has taken as a variant of the known name Idia (see Solin 2003: 1074). Prosopographically, the majority of the victims are unknown. That said, Weiss has recognized some important families from Late-Republican Hispania, such as the Fulvii, the Annii and especially the Vibii Paciaeci. This final gens is Oscan in origin and settled in the southern portion of the Iberian Peninsula at an early date (probably as colonists), where they became important land owners. As Plutarch narrates (Crass. 4.2), when Marcus Licinius Crassus fled Marius and the populares in 85 BCE, he found refuge in southern Hispania thanks to the help of Vibius Paciaecus. His son, Lucius Vibius Paciaecus, joined Caesar more than 40 years later to fight against Pompey and captured Ulia. This deed, according to Weinrib (1990: 21–27), granted him entrance into the senatorial order and led to the family being relocated to Rome. Lucius Vibius Paciaecus as well as his brothers Quintus and Marcus Vibius are all mentioned in this defixio as the patrons of 20 of the victims. More speculatively, Weiss has identified Quintus Axius pater (l. 16) as the senator from the Quirinal tribe, businessman and friend of both Cicero and Varro, who mention him in both the Epistulae ad Atticum and Res Rusticae, where he is the narrator of Book III, respectively.

The text presents the following orthographic and phonetic features: leibertus for libertus (ll.1–4), Quinctus for Quintus (l. 3) and quei for qui (l. 15). Also note the inversion of the normal sequence nomen-cognomen in [G]alla Vibia (for expected Vibia Galla, l. 4). While the vast majority of names are in the nominative, it appears that the author tried to use the accusative Marcione(m) fratre(m) (l. 11), the only exception to the larger pattern (for the mixture of nominative and accusatives, see section I.7.2). The editor has focused on the phrase puer erum (l. 16), since it suggests that the defigens was the slave of Quintus Axius who would hence be cursing his own master. Weiss has analysed the final traces at the end of the text as a magical symbol, which seems rather unlikely, given the curse’s date. The majority of names found in the curse are of Greek origin; that said, there are also Iberian names: Arco (l. 5) and Attin (l. 10); and Latin ones all of which are attested except for Luxinina (l. 13), which is probably derived from the name of the Luxia River. The following names also stand out: Caepor (ll. 6–7; perhaps for Gaepor?), Cripo (l. 8; attested in an inscription from Brescia: CIL V, 4547) and Luipor (ll. 11–12; attested as a nomen in CIL X, 8042, 69). The Greek names often belong to slaves (e.g., Leontis, Menophilus) and are well attested,

The defigens does not mention the reason that led him to write this curse. Given that the defixio was found in Ostia, the editor has astutely suggested that it may deal with a commercial dispute between Rome and Baetica. It seems likely that the Vibii Paciaeci, important landowners from the southern part of the Iberian Peninsula, were involved in the trade of agricultural goods, such as olive oil. Perhaps the text reflects a network of mercatores/negotiatores involved in the olive oil trade between Baetica and Rome, whose main port, of course, was Ostia. After being inscribed, the tablet was folded and pierced, as we can deduce from the small hole found at the end of l. 9.

55. Tibur, Tivoli Provenance: necropolis. Current Location: private collection. Date: 1st–2nd centuries CE. Material: lead. Measurements: 13 × 17 cm. Reading: 5



10



Glucheria Aristia Clara Hermiones Antoni ancil(l)a Hilarus Arbias medicus Prima Arbia ancil(l)a staminar(ia) Limne libera est Nicethor libertus Nota (vacat) Aulia Camula Hilarus libertus Apate (vacat) ancil(l)a Aratus (vacat) servos Mar[-c.4- l]ibertus Gemel(l)a [li]berta Charis liberta 138

Italia Bibliography: Tedeschi 1919; Pacifici 1925–26: 72; II IV, 25; Cristofori 2004: 604, no. 22; Alonso Alonso 2018: no. 241.

which run from left to right and were written in capitals. The text has been dated to the first or second century CE on palaeographic grounds (note the cursive e consisting of two vertical strokes). The text presents the following orthographic and phonetic features: ancila for ancilla (ll. 2, 4 and 9), servos for servus (l. 10) and Gemela for Gemella (l. 12). We do not know the circumstances that led to the writing of this curse, which targets a group of individuals who are defined by their social status (liberti and slaves). In two cases, the profession of the victim is included: Hilarus Arbias is said to be a doctor (l. 3) and Prima Arbia a weaver (l. 4). The following names of Greek and Latin origin ought to be noted: Glucheria (l. 1; see Solin 2003: 944), Aristia (l. 1; see Solin 2003, 778), Limne (l. 5; see Solin 2003: 1218 and 1445) and Apate (l. 9; see OPEL I, 138).

Commentary: this defixio was discovered around 1900 near Grotte Sconce (ancient Tibur) together with a cinerary urn that contained the ashes of a deceased person and a copper coin. The grave had been marked with the base of a column and, in the editor’s words, ‘il graffito era fissato con quattro chiodi ad una assicella di legno quadrangolare… la faccia scritta del graffito inchiodato all’assicella poggiava sulla terra, in un angolo a destra della calcara, e sulla lamina era adagiato il basamento di colonna’ (Tedeschi 1919: 49). The curse was written on a roughly rectangular lead sheet, which has lost its lower-left corner as well as part of the lower edge (which affected the middle of ll. 11–12). On the sheet’s rounded surface, there is an inscription of 13 lines,

After being inscribed, each corner of the tablet was nailed to a piece of wood (for parallels, see 95 and 129).

Regio I [Latium Adiectum] 56. Minturnae, Minturno Provenance: necropolis. Current Location: Museum of Capua. Date: 2nd century CE. Material: lead. Measurements: 18.2 × 9 cm. Reading: 5



10



15



Dii i(n)feri · vobis · com(m)e(n)do · si quic(q)ua(m) sa(n)ctitates · habetes · ac tad{r}o Tic(h)ene(n) Carisi · qu`od´qud agat quod · i(n)cida(n)t omnia in adversa · Dii i(n)feri vobis com(m)e(n)do · il(l)ius · memra colore(m) ficura(m) · caput · capillus umbra(m) cerebru(m) · fru(n)te(m) · supe[rcil]ia os nasu(m) me(n)tu(m) buc(c)as · la[bra · ve]rbu(m) · vo[l]tu(m) col(l)u(m) · iocur umeros cor · fulmo`nes´ i(n)testina{s} · ve(n)tre(m) brac(ch)ia di[g]it`us´ os · manus · ublicu(m) · visica(m) femena cenua · crura talos planta ˹d˺ici˹t˺os Dii i(n)feri · si [illam?] videro tabesce(n)te(m) vobis · sa(n)ctu(m) · iḷ(l)ud ḷibns ob an(n)uversariu(m) facere dibus parentibus il(l)ius ++++TA peculiu(m) tabescas

Bibliography: Stornaiuolo 1881; CIL X, 8249; Mancini 1884–86: 76–81; Wünsch 1897: 27; DT 190; Niedermann 1906: 964–65; Diehl 1910: no. 852; Jeanneret 1917: no. 190; Vetter 1923: 63–65; Rubio and Bejarano 1955: no. 195;

Astori 2000: 63; Kropp 2008: no. 1.4.1/1; Faraone and Kropp 2010: 384–85; Urbanová 2018: no. 9; Gordon 2019: 426–28. Image source: Stornaiuolo 1881: 189. 139

Sylloge of  Defixiones from the Roman West Commentary: this defixio was discovered in August 1879 in a tomb located in the so-called ‘campagna Virilasi’ about 80 m away from the amphitheatre of Minturno. According to the editor, the tablet was found folded and nailed under the skull of a skeleton (1881: 188).

victim with the aim that qu`od´qud agat, quod i(n)cida(n)t omnia in adversa (ll. 4–5). According to Faraone and Kropp, in adversa is ‘a fixed expression, in which the abstract noun adversa serves as standard term for “misfortune”, “bad luck”’ (2010: 385). As Gordon points out, the use of the term sanctitas (ll. 1–2) refers here ‘to the underworld god’s own sense of justice and religious duty’ (2019b: 428). In the second part of the curse, the victim is symbolically dismembered and given to the gods, from head to toes (note caput, in l. 6 and digitos in l. 13, referring to the toes; cf. with the list in 10, ll. 18–37 [from caput to ungis, the toenails]). This is a true ‘anatomical curse’ in which not only the body of the victim is given, but also her external appearance (color, figura, ll. 5–6, compare with 10, l. 3), and her internal organs (cf. with 51). In the third portion of the curse, the defigens pledges to give the divinities invoked an annual sacrifice. Finally, (s)he targets also the parents of the victim.

The curse was written on a rectangular lead sheet, which is intact despite a crack in the central area, which coincides with the fold line. The inscription contains 18 lines of capitals, which run from left to right and are generally in scriptio continua with the occasional use of interpunctuation consisting of a small vertical stroke or a single dot found halfway up the letters. We do not know the circumstances that led to the writing of this curse, in which Tychene, the wife or slave (or freedwoman) of Carisius is handed over to the infernal deities. The text presents the following orthographic and phonetic features: iferi for inferi (ll. 1, 4 and 14), comedo for comendo (ll. 1 and 5), quicua for quicquam (l. 1), sactitates for sanctitates (ll. 1–2), habetes for habetis (l. 2), Ticene for Tychenem (l. 2), incidat for incidant (l. 3), ilius for illius (ll. 5 and 17), memra for membra (l. 5), frute for frontem (l. 7), metu for mentum, bucas for buccas (both in l. 8), vitu for vultum (? l. 8), colu for collum, iocur for iecur, umeros for humeros, fulmones for pulmones (the last four in l. 9), itestina for intestinam, vetre for ventrem, bracia for brachia (the last three in l. 10), ublicu for umbilicum, visica for vesica, femena for femina (the last three in l. 11), cenua for genua (l. 12) and sactu for sanctum (l. 15). Note the general omission of final -m and what looks like careless errors: tadro for trado (l. 2) and ticidos for digitos (l. 13).

Translation (Gordon 2019b: 427 and my own): ‘Gods of the Underworld, I hand over to you, if you have any religious duty, I deliver (to you) Tyche, (wife or slave?) of Charisius, whatever she may do, may all things fall into adversity. Gods of the Underworld, I transfer to you her body, complexion, outward appearance, head, hair, coiffure, brain, forehead, eyebrows, mouth, nose, chin, cheeks, lips, utterance, face, neck, liver, shoulders, heart, lungs, intestines, abdomen, arms, hands, navel, bladder/vagina, knees, legs, heels, soles, toes. Gods of the Underworld, if I see her languish, I will gladly offer sacrifice every year to you (...) Diminish the property of her wealthy parents (?)...’

The curse contains three main parts. In the first, the defigens summons the infernal deities to give them the

57. Fundi, Fondi Provenance: aquatic context. Current Location: unknown. Date: 6th century CE. Material: marble. Measurements: 2.6 × (21) × 1 cm. Reading: 5



qui s[ubbertis]ti libra puerum Elissei ita subber[te] domum Ber[-c.2-]ạtis qui subbertisti Sodomam et [Go-] morra ita subberte domum et animam (?) [---] natis qui apparuisti populo in d[eserto --- ita] appare c[u(m)] opera tua super istas s[edes] ++[---su]bbert[e---]

140

Italia Bibliography: Guarducci 1960; García Ruiz 1967: no. 67; Solin 1968: no. 38; Solin 2014a: 247–49; Solin 2014b: 37–41; AE 2014, 302; Alfayé 2019: 269.

between 0.15 and 0.3 cm in height. They have been dated to the sixth century CE, although they could also belong to the fourth or fifth centuries CE.

Image source: Guarducci 1960: Tav. I, fig. 2.

We do not know the circumstances that led to the writing of this curse, which targets two individuals and employs a series of similia similibus formulae. The defigens, who appears to know the Bible well and was likely a Jew (see Guarducci 1960: 7 and Solin 2014: 249), compares several passages from the Bible with the victims’ destiny. Accordingly, the first formula invokes God and asks that he punish the victims just as he punished the slave of Elisha (l. 1; cf. Kings IV 5, 26–27), who was caught stealing and punished with leprosy (l. 1; libra for lepra). It is worth noting the phrase domum Ber[-c.2-]ạtis (l. 2; or B[-c.1-] t[-c.2-]tis, according to Solin 2014a: 247), which is used to allude not only to the victim but also to his/her family (there are several parallels for this sort of ‘hereditary curse’, e.g., 246 and 370). The second analogy compares the destruction of the victim’s house (l. 3; here domum has a more material meaning than it did in the previous line) and spirit (animam, l. 3) to the destruction of Sodom and Gomorrah (for this event and its presence in other magical texts, see Martín Hernández 2016). Next, God is asked to appear in istas s[edes] (l. 5; reconstruction of Guarducci 1960: 6), just as he had done in the town in the desert (Exodus XVI, 10). The last line of the text is difficult to read, but it might contain a final cursing formula, given the presence of the lexeme subvertere, which is also found in the preceding lines.

Commentary: this defixio was discovered in 1958 in the so-called Grotto of Tiberius located to the south of Circeo on the Tyrrhenian Sea. In the first century CE, the cave was turned into the imperial villa’s nymphaeum, which was decked out with richly decorated pools, in one of which the present curse was found. Even though the editor examined this defixio in the Museo Nazionale Romano (Inv. 6812), its current whereabouts are unknown. The curse was written on a reused marble plaque, which in all likelihood originally belonged to an opus sectile. The choice of this medium, which is certainly unparalleled, seems to be related both with the context in which the curse was finally deposited (i.e., the abandoned pool located in the grotto of Tiberius) and, as we will see, with the content of the text. The cave was probably the place from which the practitioner took the object on which the curse was written, alluding to the destruction of the space in which the defixio was deposited. Therefore, the choice of this medium was probably aimed to reinforce, through persuasive analogy, the purposes pursued in the curse (cf. section I.3.3.2). The plaque, which has been broken into three fragments that fit together, is quite eroded along its bottom edge (which has led to the loss of text) and has lost its right edge. Already in Antiquity, there was damage at the ends of ll. 2–3, as can be deduced from the fact that the author of the text avoided this part of the plaque. The text, written like a titulus pictus with black ink, contains six lines, which run from left to right. Palaeographically, the letters are written in new Roman cursive and measure

Translation: ‘You who destroyed the servant of Elisha with leprosy, so you destroy the house of Ber(...)as; you who destroyed Sodom and Gomorrah, so you destroy the house and spirit (?) of (...)nas; you who appeared to the people in the desert..., so you appear with your work on these (...) destroy (...).’

58. Ardea Provenance: domestic space. Current Location: AM Ardea. Date: vid. commentary. Material: lead. Measurements: (9.1) × (13.6) cm. Reading:

vel uθraś{i} mlaχ

141

Sylloge of  Defixiones from the Roman West Bibliography: Colonna 2003; Mario 2003; Mario 2007: 15–17, 22–23; Massarelli 2014: 213–14; Massarelli 2016: 530–31; Massarelli 2019: 366.

Vel and the nomen Uthra(ś). The latter is likely derived from the name Uta, which is attested as a nomen already in the seventh century BCE, which in turn led to the form *uta-ra and then by the middle of the fifth century BCE to ut/θra. Colonna has not analysed mlaχ as an Etruscan adjective ‘beautiful’, but rather as a parallel for Latin vir bonus and Oscan medekon/-an (see 78). On this interpretation, this epithet would place the victim among the Etruscan aristocracy of Ardea. Based on the identification of the victim as an aristocrat, Colonna has argued that the curse should be dated to the mid-fifth century during the conflicts described in Livy (IV, 9–10). This date, however, is extremely early when compared to the rest of the Etruscan curses (see 87–91 and 94). Given the lack of palaeographic evidence for such a date, and the fact that the defixio was found together with materials dating from the eighth to the fourth centuries BCE, the date of the curse tablet should accordingly be reconsidered.

Image source: Colonna 2003: 342, fig. 3. Commentary: this Etruscan curse tablet was discovered during the excavations carried out between 2001 and 2002 in the archaeological area of Colle della Noce. It was found together with materials dating from the eighth to the fourth centuries BCE (Massarelli 2019: 366), inside a large rectangular limestone building of unknown purpose, which is located near a sanctuary (Mario 2003: 23). At the time of discovery, it was still folded. The text was written on an oblong lead sheet, which has lost its top-right corner as well as the left and bottom edges. The inscription contains two lines, which run from right to left. The circumstances that led to the writing of this curse are unknown. The text consists of a bipartite onomastic formula followed by mlaχ.

After being inscribed, the tablet was pierced twice with a nail before being folded longitudinally twice so that the text was not visible.

Following Colonna (2003: 345), the onomastic formula, written in scriptio continua, is made up of the praenomen

59. Antium, Anzio Provenance: necropolis. Current Location: M. di Villa Adele (Anzio). Date: 1st century BCE–1st century CE. [Inv. No.: 515809] Material: lead. Measurements: 8.5 × 10.5 × 0.2 cm. Reading: A B C(aius) · Clodiu(s) · Rufus C(aius) · Clodius · C(ai) · f(ilius) · Ru«f»us

Commentary: this defixio was discovered alongside 60 during the excavations carried out between 1960 and 1975 in a section of modern Bergasi street (Anzio), which has provisionally been identified as the ancient Via Severiana.

Bibliography: Calderini 2007; Solin 2016; AE 2016, 229. Image source: Calderini 2007: 58–59, Tab. II. Courtesy of A. Calderini. 142

Italia In this area, archaeologists uncovered the remains of a necropolis, which contained inhumation and cremation graves, and which has been dated between the first and third centuries CE. The absence of specific information concerning the excavation keeps me from providing details about the specific archaeological context for these two tablets (for more general information about the archaeological context, see Jaia 2007: 45–47).

words are separated with rounded interpunctuation placed halfway up the letters. The text is dated between the first century BCE and the first century CE on palaeographic and onomastic grounds. The circumstances that led to the writing of this curse are unknown. The text merely consists of the victim’s name. Perhaps, as Solin (2016: 288) has maintained, the curse’s author ‘improved’ the first version of the text (found on side A) on the tablet’s side B where the victim’s patronymic is added and the nomen is spelled correctly. Solin has argued that the nomen found on side A is a nominative singular ending in –i, which is common in this period. In the drawing published by Calderini, however, we can make out traces of what very well could be a u (side A just above the r in Rufus). Therefore, it seems more reasonable that the author of the text ran out of space and simply left out the final s.

The defixio was written on a square lead tablet (with rounded corners), which is preserved in full. The shape of the sheet is quite unusual. Its rounded corners and holes, which were avoided by the author, suggest that this could be a repurposed object, like a scale pan. An opisthograph, the inscription contains two lines, which run from right to left and were written in capitals (measuring between 0.5–1 cm [side A] and 0.5–1.2 cm [side B]). The letters were ‘mirrored’ (that is, written from right to left). The

60. Antium, Anzio Provenance: necropolis. Current Location: M. di Villa Adele (Anzio). Date: 1st century CE. [Inv. No.: 515809] Material: lead. Measurements: 8.5 × 8.2 × 0.2 cm. Reading: 5



vasus Horte(n)si Tatin(a)es f(ilii) com(m)endatus ut {h}eius mem˹br˺a et baletudina(m?) ex aniṃ(o or -a) dep(e)rdaṭ

Bibliography: Calderini 2007; Solin 2016; AE 2016, 230.

1960 and 1975 (for the archaeological context, see the commentary for 60). The curse was written on a rectangular lead sheet, which is intact, despite a break along the left side and a hole in the centre of the tablet (note that the author of the text avoids it). The inscription contains seven lines, which run from left to right and were written in capitals (measuring between 0.5 and 1.2 cm in

Image source: Calderini 2007: 58–59, Tab. III. Courtesy of A. Calderini. Commentary: this defixio was discovered alongside 60 during the excavations carried out in Anzio between 143

Sylloge of  Defixiones from the Roman West if they represent, as I think, miniature cinerary urns). The curse also contains a brief anatomical sequence, which references membra, baletudine and animus/anima (ll. 4-6, see 10, 129). Calderini’s reading of ex aniṃ(o or -a) dep(e)rda/ṭ (ll. 6-7) is attractive (contra Solin 2016: 289– 90), especially since verbs like (de-)perdo are attested in defixiones of similar chronology (see 10–14 and 17, etc.).

height). Note that the letters in ll. 1–2 are much bigger than those found in the rest of the inscription. This suggests that they serve as a sort of heading (for this differentiation between parts of the text, see 61 and 215). The circumstances that led to the writing of this curse are unknown. The text presents the following orthographic and phonetic features: Hortesi for Hortensi (l. 1), Tatines for Tatinaes (l. 2), heius for eius, memrba for membra (both in l. 4) and baletudina for valetudina(m?) (l. 5). This defixio begins with a phrase that is unparalleled in the corpus of curses from the Roman West: vasus Horte(n)si Tatin(a)es f(ilii) (ll. 1–2). In this case, vasus should be understood to mean a ‘funerary urn’ (see Calderini 2007: 64 and n. 19), which here has a metaphorical meaning (i.e., the vase represents the fate of the victim; on this see Solin 2016: 290. Other curses that use or mention cinerary urns with similar purposes are 16 and perhaps the containers from the fountain of Anna Perenna [cf. 24–27 and 35–36]

Finally, the use of the matronymic (l. 2 Tatin(a)es f(ilii)) to identify the victim is noteworthy. The mother’s name is common among slaves and hence its appearance here may be a way to signal the servile status of the victim (see Solin’s commentary on the name Tatina [2016: 290–91] for a general discussion). Translation: ‘The funerary urn of Hortensius, son of Tatina, has been deposited, so that he may be deprived of his limbs and health from his life.’

Regio I [Campania] 61. Cumae, Cuma Provenance: necropolis. Current Location: MAN Naples. Date: 1st century CE. [Inv. No.: 4641] Material: lead. Measurements: 12.8 × 8.3 × 0.1 cm. Reading: 5

nọmen · delatum Ṇaeviae · L(uci) · l(ibertae) Secunda · seive­ ea · alio · nomin̂e   est

Bibliography: Avellino 1845–46; Avellino 1847–48; CIL I, 820= X, 1604; CIL I², 1615; Wünsch 1897: 25; DT 196; ILS, 8746; Jeanneret 1917: no. 196; ILLRP, 1149; Kropp 2008: no. 1.5.3/1; Urbanová 2018: no. 31.

separated with interpunctuation consisting of a single dot found halfway up the letters. Note Secunda for Secundae (l. 3) and the ending of nomine, where the n joins the e in a nexus (this is not reflected in Avellino’s drawing, which is also missing the final e of seive; cf. the photograph in the appendix). We do not know the circumstances that led to the writing of this curse, which contains a simple formula for handing the victim over to the infernal deities. The final phrase seive ea alio nomin̂e est (ll. 3–5) demonstrates that the defigens understood the importance of names in magical practices (see 92 for a parallel and section I.4.1 for a general discussion).

Image source: see App. IV.3, SD 61. Commentary: this defixio was found in a tomb in Cumae (Mommsen, CIL) in c. 1884. There is no further information about the archaeological context. The curse was written on what appears to have been a rectangular lead sheet, which today is missing its top-left corner (as well as the first letter of l. 1). The inscribed surface of the text, which is somewhat curved and quite corroded along its edges and fold line, has been lost at various places. During my autopsy in September 2010, I also observed damage on the left edge at the beginning on l. 2.

According to the editors of CIL, after being inscribed, the tablet was folded, pierced with a nail ‘et clausa filo’ (i.e., ‘tied’ with a thread, which was likely a wire; for this practice, see 71 and 52). Translation: ‘The name of Naevia Secunda, the freedwoman of Lucius, or whichever name she has, has been brought down.’

The inscription contains five lines, which run from left to right and were written capitals (measuring between 0.4 and 1.1 cm in height) by a trained hand. The words are

144

Italia 62. Cumae, Cuma Provenance: necropolis. Current Location: MAN Naples. Date: 1st century BCE. [Inv. No.: without inv. no.] Material: lead. Measurements: 14 × (12) × 0.1 cm. Reading: 5



10



Ṃ(arcum) · Hẹium · M(arci) f(ilii) Caled[um] Blossiam · C(ai) · f(iliam) · P(ublium) Heium M(arci) · f(ilii) C̣ale[dum] [Ch]ilonem · Hei M(arci) s(ervum) M(arcum) Heium C̣(aium) Blossium (mulieris) L(ibertum) Bithum Attoṇ[em] [He]i M(arci) ṣer(vum) · Blossiam L(uci) · f(iliam) [hos (?)] họmines omnes · infereis [de]ịṣ deligo ita ut niq[uis] [e]orum quem dum · VI[---] [---]+ ẹṭ possit ni[---] [---]qụịdq[uam ---] [--- p]ossit id ded[ico deis] [maniu]m ut ea ita faci[ant]

145

Sylloge of  Defixiones from the Roman West Bibliography: Stevens 1891: 80–81; Mancini 1891–93; AE 1894, 113; Wünsch 1900: 237; Paribeni 1903; AE 1903, 349; DT 199; Jeanneret 1917: no. 199; Cugusi 2007: 32–33; Kropp 2008: no. 1.5.3/2; Urbanová 2018: no. 32.

(measuring between 0.3 and 0.8 cm in height). Several words are separated with interpunctuation found halfway up the letters, although the use of interpuncts is not systematic. It is also worth noting the symbol Ɔ (l. 4, for mulieris), which is unparalleled in the curse tablets from the Roman West.

Image source: Stevens 1891: 81.

The text’s structure is quite simple and consists of a list of personal names and a cursing formula. The victims are six men and two women, all of whom belonged to the same household. The list begins with Marcus Heius Caledus (l. 1), his wife (?) Blossia and son Publius Heius Caledus (l. 2), who are followed by their slaves Chilo and Atto as well as their freedmen Marcus Heius, Bithus (all in ll. 3–4), and Blossia (l. 5). Caledus is a cognomen derived from a praenomen (cf. Kajanto 1982²: 178), while Blossia is a personal name well attested in Campania and Latium (e.g., CIL VI, 13615 and 34393). Both Chilo and Bithus are Greek names (see Solin 2003: 1481 and 1454, respectively).

Commentary: this defixio was discovered c. 1890 in a grave located near Cumae. According to Stevens, the grave contained two inhumations, the second of which preserved the remains of a skeleton, some pottery and the curse tablet under consideration (1891: 80). The archaeological context dates to the first century BCE. The curse was written on a lead sheet broken into three fragments that fit together. Today, the tablet is in very poor condition: the surface is cracked and in large part corroded, which has led to the loss of the inscription at several points. During my autopsy in September 2010, I was able to confirm that the tablet has lost sections along its edges since the publication of the editio princeps. Due to this damage, the first letter as well as the last two or three of each line have generally been lost (as signalled in the reading with underlining). The first fragment (measuring (1.5) × (9) × 0.1 cm) still conserves part of the upper edge, while the second ((8.6) × (10.2) × 0.1 cm) shows two important breakages at ll. 5 and 7. The third fragment ((3.4) × (5.9) × 0.1 cm) is very corroded, especially on its left half. The inscription, which Cugusi has claimed is written in iambic verse, contains 12 lines, which run from left to right and were written capitals

The second part of the text, which is quite fragmentary, contains the curse proper, in which the victims ([hos] họmines omnes, l. 6) are bound and probably handed over (deligo, l. 7) to the infernal deities (note the archaic infereis, l. 6; for a parallel, see 122). The repetition of the verb possum (ll. 9 and 11) recalls the non permittas formulae, which were used to impair the victims physically and/or intellectually. That said, the text is too fragmentary to be securely reconstructed. After being inscribed, the tablet was rolled up.

63. Cumae, Cuma Provenance: necropolis. Current Location: MAN Naples. Date: 3rd–2nd centuries BCE. Material: lead. Measurements: vid. reading. Reading: 5



5





Frag. C (7 × 10.1 cm) [---]muṭ[---]lli [---]vsfuvdis mạ[---] ̣ [---]dekisbuttis dekis rahiis maraheis niir kulupu 5 dkuva rạhiis upfalleis marahisrahiispapeis dekis hereiisdekkieis saipinaz maras rufriismaras blaisiismarahee+s dekkieis rahiieisuppiieismuttillieiṣ 10 dekkieisheriieis akkatusinimtrstus sullus iṇim eisun k ụhftis (traces)usullumsụllas

Frag. A (2 × 1.6 cm) p[---] u[---] ga[---] luv[---] pak[---] inim[---] dek[is---] kulu[---] ma Frag. B (1 × 0.6 cm; 0.6 × 0.5 cm) dek gn luv[---] deki[---] kluv[---] mara[---] minaz[---] mara[---] stat[---] m 146

Italia 5



10



5



Frag. D (2 × 3.6 cm) [---]v[---]ṣis nu[---] [---]m[-c.1-]tillis [---]ṭịis gnaivsf[---] [---]pakulliisl[---] kersnuvelehi[---] [---]ạhiis marạ[---] rahiisupf[---] [---]s papeis [---]smarah [---]ṣkavkas [---]utiiṣ [---]ise



Frag. F (fits together with Frag. C) [---]vf[---] [---]ti[---] [---]ii[---]



Frag. G [---]pf[---]



Frag. H [---]s[---]

Frag. E (c. 2.4 × 2.6 cm) [---]maraheis [---]eis [---]ṣ[-c.1-]niir kuḷupu ṣ [---]ḷeis [---]niir [---]gn[---]

Bibliography: Von Planta 1893; Von Planta 1897: no. 119; Wünsch 1897: 24; Conway 1897: no. 137; DT 209; Jeanneret 1917: no. 209; Vetter 1953: no. 5; Pisani 1964: no. 30; Marchese 1976: 297–300; Marchese 1978: 884– 85; Álvarez-Pedrosa Núñez 1997: 107 and 118; Rix 2002: Cm 14; ImIt: Cumae 8; Murano 2009; TDO 3.

(measuring (2) × (3.6)) is in four pieces that fit together and contain 12 lines. The first letter of each line has been lost. Finally, fragment E (measuring at most (2.4) × (2.6) cm) now has seven parts, six of which fit together. In addition to the six ‘big’ fragments (A–F), Von Planta and Conway have claimed that there are another five smaller fragments, which today have been lost (see TDO, p. 118).

Commentary: this Oscan defixio was apparently discovered in the area surrounding ancient Cumae under unknown circumstances and at an unknown date. The Museo archeologico nazionale di Napoli acquired the curse tablet in April 1881, where it is still housed today. Upon discovery, the tablet was broken into six fragments, two of which fit together to form fragment C. Nearly a century later, Marchese’s autopsy revealed that ‘tutti i frammenti indicati dal 1º editore sono tuttora reperibili, ma alcuni si presentano ulteriormente frammentati e, nella frattura, sono andati perduti alcuni frustoli’ (1976: 298). As if this were not enough, Murano’s autopsy (2010) showed that ‘Lo stato di conservazione dei vari frammenti dell’iscrizione è inoltre peggiorato nel tempo: rispetto alla situazione riscontrata da Marchese… è quasi del tutto perduto il fram. B, ridotto in più frammenti di piccolissime dimensioni’ (TDO, p. 117).

Originally, the fragments were arranged as follows: A, B and C1 made up the tablet’s left edge, while D, E and C2, in turn, made up the right edge. At the same time, A and C constituted the top and bottom edges. While the positioning of the fragments is necessary for any reading, Murano has explained that none of the various hypotheses about these fragments are conclusive, with the exception of Von Planta’s reconstruction of fragment C (for a detailed discussion, see TDO, pp. 120–22). To make matters worse, the current state of the fragments has made it impossible to reconstruct the curse. We cannot even say with certainty how many lines of text there were in total. The inscription, written in the Oscan language and alphabet, had at least 12 lines in scriptio continua, which run from left to right. Occasionally, words are divided with interpunctuation consisting of two high dots. The text is mostly comprised of a long list of the victims, who are named with bipartite onomastic formulae in the nominative case. In three cases, a patronymic (consisting of the father’s praenomen) qualifies the bipartite formula (C, ll. 4–7) and in one case an individual is referred to with a patronymic and place of origin (C, l. 7). In fragment C several of the personal names are repeated, which could very well be a method of stressing who the most important victims were (for repetition of names, see 104 and 123).

Based on Murano’s analysis, fragment A ((2) × (1.6) cm) has been broken into five parts that fit together, though three lines documented in the editio princeps have been lost (ll. 1, 2 and 9). Furthermore, it appears that fragment A fits together with another fragment, which is uninscribed and formed part of the tablet’s left edge. Fragment B is in six fragments, two of which have preserved signs of writing. Though the inscription was originally opisthographic, currently only the first letters of A, ll. 3–4 and 9–10 are visible, while nothing can be made out on side B. Fragment C (measuring at most (7) × (10.1) cm) is made of two parts: C1and C2 (today broken into two parts that fit together, C2-1 and C2-2) to which fragment F must be added; this final piece is currently lost and belonged between the other two fragments at ll. 2–4. The break of C2 led to the loss of text along the edge in ll. 3–4 and 9–11. Fragment D

The list of names is occasionally interrupted by the sequence niir kulupu (C, l. 4; E, l. 3; A, l. 8 and perhaps E, l. 5), which has been interpreted in a wide variety of ways: 147

Sylloge of  Defixiones from the Roman West a cursing formula (Von Planta), an honorific title (Vetter, who identifies it as a collegiate magistrate), a derogatory epithet (Lejeune offers ‘princeps latronum’), but no consensus has been reached (for a summary of positions, see TDO, pp. 126–29). The sequence kersnu velehi[---] (D, l. 5) has proven to be yet another matter of debate: Von Planta takes it as another cursing formula, while Crawford (apud ImIt) translates it as ‘may their… dinner be spoilt’ on the assumption that kersnu (from *kert-s-na) is related to a type of ritual meal of grains. Murano (TDO, p. 130), for her part, has suggested that the word is somehow associated with a ritual. All that said, the impossibility of establishing a sound etymology for velehi[---] has greatly complicated our understanding of this portion of the curse.

The curse concludes with the phrase akkatus inim trstus sullus iṇim eisunk ụhftis sullud [s]ụllas (C, ll. 10–12). In 1897, Skutsch (apud Wünsch) took akkatus inim trstus as an equivalent to Latin testes et advocatos (in which case akkatus is the Oscan adaptation of a Latin loan word). Subsequent editors have more or less accepted this (Vetter has taken the sequence as nominative rather than accusative). Given that sullus is Oscan for omnis, the entire phrase can be construed as advocati et testes omnes et eorum uhftis omnium omnes (for the controversial word uhftis, see TDO, p. 126). The interpretation of this phrase, then, allows us to place this curse in the group of juridical defixiones.

64. Cumae, Cuma Provenance: unknown. Current Location: AK Bonn. Date: first half of the 1st century BCE. [Inv. No.: C143a] Material: lead. Measurements: 7.6 cm diameter. Reading: 5



L(---) · Ha`{ạ}´rines · Her(---?) · M-`aturi(s?)´ C(---) · Eburis · Pomponius · M(---?) · Caedicius · M(---?) · f(ilius) · N(---?) · Andripius [·] N(---?) [·] f(ilius) [·] pus · olu · solu · fancua rectasint · pus · flatu sicu · olu · sit

Bibliography: Bücheler 1907a; Bücheler 1907b: 291–98; Bréal 1908: 146–48; CIL I², 1614; Besnier 1920: no. 13; Vetter 1953: no. 7; ILLRP 1146; Pisani 1964: no. 31; García Ruiz 1967: no. 65; Marchese 1976: 305; Knobloch 1978; Marchese 1978: 886–87; Mancini 1988; ÁlvarezPedrosa Núñez 1997: 109, 114; Rix 2002: Cm 15; Pike 2005; TDO 4; ImIt: Cumae 10; Mancini 2014: 52–53; Estarán Tolosa 2016: O6.

Though written in the Latin alphabet, scholars have traditionally analysed the curse as a sort of hybrid between Latin and Oscan, forming a sort of dialect that was apparently spoken in Cumae at that time (Bücheler, Bréal, Degrassi, Pisani, Marchese and Knobloch). Instead of identifying such a dialect in this tablet, Mancini has compellingly argued that this tablet actually provides evidence for a sort of symbiosis between Oscan and Latin. Following his hypothesis, a Latin speaker has written an Oscan formula that (s)he had heard spoken. This mixture should be understood in the context of the magical practice, i.e., in an attempt to augment the spell’s efficiency. In his words, ‘la lingua in cui viene pronunciata la formula è… intenzionalmente distante dall’uso quotidiano’ (1988: 208; for the debate about this tablet, see TDO, pp. 138–42).

Image source: Knobloch 1978: 166. See also App. IV.3, SD 64. Commentary: this defixio, which probably hails from Cumae, was written on a small lead disc, a form paralleled by 145, 472 and 514. The tablet is intact, though it has several small cracks along its surface as well as a larger one at ll. 4–5 that coincides with the fold line.

The text can be divided into two parts: a list of personal names (ll. 1–6) and the curse proper (ll. 7–9). The first part provides the names of the text’s six victims, all of which are in the nominative case, although different types of onomastic formulae are used. The first three victims (ll. 1–3) are referred to with an Oscan bipartite formula (in -­is) consisting of a praenomen and nomen. The final two (ll. 4–6), on the other hand, are named with a Latin onomastic formula. As Estarán Tolosa (2016: 111 n. 44) has pointed out, praenomina can be

The inscription contains nine lines, which run from left to right. L. 1 is, in fact, a continuation of l. 2 (Bücheler 1907: 555 has interpreted the first sign of l. 1 as an ạ, which would yield the sequence –aturi that could be added to the end of l. 2). Based on its palaeographic characteristics, the text has been dated to the first half of the first century BCE. 148

Italia expressed with both Oscan and Latin abbreviations: L (l. 2) could be for either L(ucius) or lúvkis, while in l. 3 C could be for either C(aius) or gaavis. According to Mancini (1988: 216–17), the praenomen Her(eno) is an abbreviation for Oscan heírens. He has also argued that the nomen harines could be derived from Etruscan harina and that Pomponius (l. 4; already attested in Oscan) is a slave, whose master, C. Eburis, appears in the preceding line.

223). This would explain the forms sint, sit and recta (ll. 7–8), which in Oscan would be sins, *sid and *rehtas, respectively. This same Latinization has affected flatu and sicu (ll. 7–8), though we do not know the corresponding Italic terms. Vetter, in turn, has correctly interpreted fancua (l. 6) as meaning ‘tongue’ and translated recta (l. 7), related to the Latin rectus by Bücheler, as ‘stiff’. Pike is in agreement with this analysis and has compared the Oscan fancua(s) recta(s) sint in this text with the Ovidian lingua ne rigeat (Ars am. I, 45). Furthermore, according to Vetter, flatus (l. 7) should be taken as homologous to Latin flatus (breathing), since the absence of breath would also make the victim speechless. Given the reference to a stiff tongue, we can group this curse among the juridical defixiones (for a comparison, see 50).

The second portion of the text (ll. 7–9) contains the curse proper, which consists of two phrases introduced by pus (equivalent to Latin ut). Murano (TDO, pp. 135–38) has proposed the following reconstruction: pus ol(l)u(m) solu(m) fancua(s) recta(s) sint pus flatu(s) sic(c)cu(s) ol(l)u(m) sit. Phonologically, the curse demonstrates that wherever it has been possible, the writer tried to Latinize the Oscan formula (Mancini 1988:

After being inscribed, the tablet was folded in half.

65. Cumae, Cuma Provenance: necropolis. Current Location: MAN Naples. Date: 1st century BCE. Material: lead. Measurements: 7.5 × 9.2 cm. Reading: 5



stenim · kalauiiúm · trí aginss ·urinss · úlleis fakinss · fangvam biass · biítam · aftiím ·{a} anamú{·}m ·aitatum amirikum · tíf[---]

Bibliography: Maiuri 1913; Ribezzo 1914–915; Besnier 1920: 20; Vetter 1932: 16; Vetter 1940: 147–52; Vetter 1942: 236–37; Vetter 1953: no. 3; Marchese 1976: 293–95; Marchese 1978: 882–83; Álvarez-Pedrosa Núñez 1997: 107–08, 118–19, 115; Rix 2002: Cm 13; ImIt: Cumae 9; TDO 5.

along the tablet’s central portion and the other through the top-right corner. The text, written in the Oscan language and alphabet, contains six lines, which run from left to right. The words are separated with interpunctuation consisting of a single dot found halfway up the letters.

Commentary: this defixio was discovered in April 1913 during the excavations of the Oscan necropolis in the socalled ‘fondo Correale’. The curse was uncovered in a grave made of flat limestone blocks, whose grave goods, consisting of three bronze coins and some pottery, have been dated between the end of the second century and beginning of the first century BCE.

We do not know the circumstances that led to the writing of this curse, which is directed against Stenim Kalauiiúm, whose name appears in l. 1. Vetter has suggested that this bipartite onomastic formula is qualified by the patronymic Trí (l.1), an abbreviation for trebiis. While Maiuri and Ribezzo have interpreted aginss, urinss and fakinss as verbal forms, Vetter has argued that they are actually accusative plurals (from nouns ending in –ion) that are equivalent to Latin actiones, orationes and facinora respectively. This hypothesis is attractive.

The curse was written on a roughly trapezoidal lead sheet, which has lost part of its right edge (including the bottom corner). There are two large cracks, one of which runs

The text next presents an ‘anatomical’ curse that is organized in parallel manner to some Latin curses (e.g., 56 and 51). In this case we find fangvam (l. 3), which Vetter

Image source: Maiuri 1913: 474.

149

Sylloge of  Defixiones from the Roman West and subsequent editors have taken as meaning ‘tongue’ and then biass and biítam (both in l. 4), which Ribezzo equates to Greek βία and Latin vita. The interpretation of aftiím (l. 4) has been more problematic: Ribezzo takes it as equivalent to Latin alloquium, whereas Vetter thinks it means something like spriritum or valetudinem; Crawford (apud ImIt), in turn, translates it as ‘sight’. There is more consensus surrounding aitatum (l. 5), which Maiuri connects to Latin aetatem in the sense of ‘lifespan’ (cf. Watkins apud

TDO, p. 151), and anamúm (l. 5), which is related to Latin animum. Marchese (contra Vetter) has analysed amirikum (l. 6) not as an accusative, but rather a first person singular present indicative active verb (cf. manafum in 68). In this case, amirikum would be the main verb (a sort of cursing verb) and tíf[ei] would be a pronoun referring to the deity invoked. The reference in this curse to the victim’s tongue, actions and speech suggests (with some uncertainty) that this text can be classified as a juridical defixio.

66. Capua, Santa Maria Capua Vetere Provenance: necropolis. Current Location: MAN Naples. Date: end of the 2nd century–1st century BCE. [Inv. No.: 12140] Material: lead. Measurements: 7.2 × 18 cm. Reading: 5



steniklum · vírriis tryhpíu · vírríiis plasis · bivellis úppiis · helleviis lúvikis · úhtavis statiis · gaviis · nep · fạtíum · nep · deíkum · pútían`s´ lúvkis · úhta˹v˺is · núvellum velliam · nep · dẹíkum · nep · faṭíum pútíad nep · memniṃ · nep · úlam · sífeí · heriiad

Bibliography: Minervini 1857: 99–102; CII 2749; Wünsch 1897: 24; Conway 1897: no. 131; DT 192; Jeanneret 1917: no. 192; Vetter 1942: 231–32; Vetter 1953: no. 4; Pisani 1964: no. 29; Knobloch 1975; Marchese 1976: 295–97; Marchese 1978: 883–84; Álvarez-Pedrosa Núñez 1997: 107, 114, 118; Rix 2002: Cp 36; TDO 1.

editio princeps, today its right half has been totally reduced to dust due to improper storage (TDO, p. 24 and table II, whence I have taken the present reading). The text, written in the Oscan language and alphabet, contains nine lines, which run from left to right. With a few exceptions (see ll. 7–8), the words are separated with interpunctuation consisting of a single point placed halfway up the letters.

Image source: Minervini 1857, Tab. VIII, fig. 1.

Between ll. 6 and 7, a horizontal line divides the text into two parts: the first (ll. 1–6) consists of a list of personal names in the nominative, while the second focuses on a single victim. Despite difficulties in reading (l. 2, TDO, p. 27) and although splitting up the onomastic formulae had been a matter of some debate (especially, ll. 1 and 3; see TDO, pp. 25–27), specialists now agree that the victims

Commentary: this defixio was discovered in 1857 at the necropolis of Capua in a grave that had already been destroyed in Antiquity (‘già in epoca remota violato e depredato’, Minervini, 1857: 99). The curse, which was still rolled up when found, was written on a rectangular lead sheet. Though the tablet was intact at the time of the 150

Italia are named with bipartite onomastic formulae (containing a praenomen and nomen). The text’s first part ends with the phrase nep · fạtíum · nep · deíkum · pútían`s´ (l. 6), which Vetter (1953: no. 4) has argued is equivalent to Latin nec fari nec dicere possint. Accordingly, it appears that the defigens sought to silence the victims, who very well may have been witnesses in a trial. If this interpretation is correct, this defixio is a juridical curse (for a parallel, see 131).

repeated here (l. 8) with the further desire that he lose his memory (nep · memniṃ, l. 9). Vetter identified úlam (l. 9), which Minervini wrongly compared to Latin ollam, as a feminine personal pronoun that in this case refers back to velliam (l. 7), equivalent to voluntas and perhaps borrowed from Latin (Marchese 1976: 297 and 1978: 883–84). Finally, núvellum (l. 7) is a nomen in the genitive plural referring to the defigens’ family. As Murano has put it, ‘è possibile che in Lucio Ottavio possa essere ravvisato un procuratore o comunque qualcuno che dovesse testimoniare sugli interessi della famiglia dei Novelli e in particolare riguardo a lla loro voluntas legata a qualche contratto o volontà testamentaria’(TDO, p. 38).

The second portion of the inscription turns exclusively against lúvikis úhtavis. As Vetter has noted, this victim is named twice in the text (ll. 5 and 7) and perhaps was the star (and hence most damning) witness in the trial. Accordingly, the formula used to silence the victims is

After being inscribed, the tablet was rolled up.

67. Capua, Santa Maria Capua Vetere Provenance: necropolis. Current Location: Staatliche Museum Berlin. Date: 1st century CE. Material: lead. Measurements: unknown. Reading: 5



Cn(aeum) · Numidium Astragalum {v}il(l)ius · vita(m) valetudin(em) quaistum ipsuq(ue) · uti · tabescat · mor`buṣ´ +++ C(aius) · Sextiu(s) · TABSI ++ ma(n)do rogo

Bibliography: CIL X, 3824; Henzen 1866: 252–53; Wünsch 1897: 27; DT 195; Diehl 1910: no. 853; Jeanneret 1917: no. 195; Kropp 2008: no. 1.5.2/1; Urbanová 2018: no. 30.

quaestum (l. 4), morbus for morbos (l. 4–5) and mado for mando (l. 7). Mommsen pointed to some doubtful parts of the reading like a possible v at the beginning on l. 3 or TABSI in l.6. Henzen (1866: 253) took the latter as an error for tabe, while Kropp has reconstructed it as tabescat (see Kropp 2008: no. 1.5.2/1; see below). The text contains some common tropes found in other curses: targeting of the victim’s vita and valetudo (cf. 129, 451 and 491) as well as sending morbus upon him/her (cf. 117 and 129).

Image source: Henzen 1886: 252. Commentary: according to Mommsen, this defixio was ‘reperta S. Mariae Capua in sepulcro’ at an unknown time. The curse, inscribed on an irregularly shaped lead tablet, contains seven lines, which run from left to right and were written in capitals. The words are separated with interpunctuation consisting of a single dot found halfway up the letters, although the use of interpuncts is not systematic.

Astragalus (l. 3) is a Greek cognomen (see Solin 2003: 1244). Given its position in the text, we would expect TABSI (l. 6) to be part of another cognomen, rather than a verbal form (contra Kropp). That said, there are no obvious known parallels to help reconstruct such a cognomen in this text.

The circumstances that led to the writing of this curse are unknown. The tablet presents the following orthographic and phonetic features: the loss of final –m in vita(m) and probably in valetudin(em) (l. 2) as well as quaistum for

Translation: ‘(I curse) Gnaeus Numidius Astragalus, his vitality, health, business and himself, may he die of disease. Caius Sextius (...) I hand over (him) to you (and) ask.’

151

Sylloge of  Defixiones from the Roman West 68. Capua, Santa Maria Capua Vetere Provenance: necropolis. Current Location: MAN Naples. Date: 4th–3rd century BCE. Material: lead. Measurements: (8) × (22) cm. Reading: A keri ạrent[-c.5-]anafụm pai pụ[-c.2-]ui heriam ṣuvạm legin[-c.8-]flukad[---] usurs inim malaks nistrus pakiụ : klụvạtiud valamais p[-c.5-] antkạdum damia[---] leginụm : ạflukad idik tfei · manafum vibiiai prebaiam pu[-c.1-]ulum da[-c.2-]d keri ar[---] valạimas puklum inim ulas lẹginei svai neip dadid : lamatir akrid ẹiseis dunṭẹ[---] 5 inim kaispatạr i[-c.3-] krustatar svai neip avt svai tiium idik fifi kus pụst ẹis[---] pun kạhad ạvt n[-c.3-]rnum neip puṭiiad punum kahad avt svai pid perfạ[---] pụṭiiad nip hu[-c.1-]truis nip suprụis aisusis putiians pidum putiians ufteis udf[---] vạlaimas pụklui pun far kahad nip putiiạd eduṃ nip menvum limu pi[---] pạị hụmúns biṿus karanter suluh pakis kluṿatiis valaims puk ṭurumiiad lu[---] 10 vibiiai akviiai svại puh aflạkus pakiṃ kluvatiium vạlaimaṣ puklui supr[---] inim tuvai leginẹi inim sakrim svại puh aflakus hụṇtrus tẹras huntrus a[---] valamais puklu pvt keri aret[-c.4-] avt ulas leginei nuhtirnas trutas tus[---] B [-c.6-]v[---] keri arentika[-c.2-]ai puisuvạh[-c.3-] +lẹgin+++[-?]krus[---]

Bibliography: Bücheler 1878; Pascal 1894; Conway 1897: no. 130; DT 193; Pascal 1923; Kent 1925; Vetter 1942: 228–31; Bonfante 1933; Bolling 1939; Vetter 1953: no. 6; Pisani 1964: no. 28; Marchese 1976: 301–05; Marchese 1978: 885–86; Pugliese Carratelli 1981; Prosdocimi 1989: 525–26; Campanile 1993c; Álvarez-Pedrosa Núñez 1997: 108–09, 118; Poccetti 1998; Rix 2002: Cp 37; Mancini 2006; TDO 2; Mancini 2014: 46–53.

Commentary: this defixio was discovered in 1876 during the excavations carried out at the ancient acropolis of Santa Maria Capua Vetere, located in the so-called ‘fondo Patturelli’, joined to the sanctuary where the famous matres figures were discovered. There is scarcely any information about the archaeological context of this curse, except that it was found in a grave and was still rolled up. Friedrich von Duhn acquired the tablet and in 1876 sent it to Berlin as a present for Franz Bücheler, who published the editio princeps and then donated it to the Museo Archeologico Nazionale di Napoli, where it is still housed.

Image source: Bücheler 1878: 160.

152

Italia The curse was written on a rectangular lead sheet, which, according to Marchese (who conducted two autopsies in 1974 and 2002), is now broken into 11 fragments, most of which fit together. Only the right edge is intact, while between 2 and 22 letters (depending on the line) have been lost on the left side (for this analysis, see TDO, pp. 41–42). This opisthographic inscription is one of the longest Oscan texts and is the longest Oscan text to date. It contains 14 lines, which run from right to left and were written in the Oscan language and alphabet. The majority of the text is in scriptio continua, though several words are separated with interpunctuation made of two dots. Given the lack of regular interpunctuation, scholars have had difficulty separating words from one another, which has led to various problems with its interpretation. Based on palaeography, Poccetti has dated the text between the fourth and third century BCE. The reading provided here is taken from Murano (TDO, p. 44), who carried out an autopsy in 2010 and improved the previous readings of Bücheler, Conway and Marchese.

More controversial is the identification of Vibia, which appears twice in the text (Vibia Prebia, l. 3: vibiiai prebaiam and Vibia Aquia, l. 10: vibiiai akviiai). This has traditionally been understood as the defigens (Bücheler, Poccetti). Nevertheless, as Murano (TDO, p. 51) has signalled, the mention of Vibia is inconsistent. Thus, while in the first case this woman would be referred to with only a praenomen (l. 3: vibiiai; the accusative singular prebaia could be a nickname, as Campanile (1993c: 57) has argued); in the second appearance, in contrast, she is referred to with a complete onomastic formula. Pascal (1894: 140), in turn, has taken Vibia as a priestess from the temple of Ceres, while Vetter (1942: 228 and 1953: no. 6, p. 42) has claimed that Vibiiai Prebai (l. 3) and Vibiiai Akviiai (l. 10) are two secondary deities who carry out the orders of Keri Arentikai. Pisani (1964: no. 28) has suggested that Vibia refers to the dead woman in whose grave the curse was deposited. Picking up on this idea, Prosdocimi (1989: 525–26) has maintained that given the divinization of the dead in the Hellenistic period, Vibia could be a divinized dead woman in whose tomb the curse was placed; in this case, prabai and akviai would be two divine epithets.

Despite the difficulties in interpreting the text, researchers have reached a consensus that the curse targeted Pacio Clovazio, who is accused of having committed a theft. The victim, who may have been a member of the illustrious gens Clovatia (Poccetti 1998: 184), is mentioned six times throughout the text with a bipartite onomastic formula, followed by a matronymic (for this frequent qualifier in defixiones, see TDO, pp. 55–60 and Lazzarini 2014). The defigens summons Keres Arentika (Keri Arentikai) along with the numina that belong to her divine entourage (leginum suvam) as well as the goddess Damia (see TDO, pp. 53–55 and 51–52, respectively). These divinities are charged with paralysing the victim’s vital functions, which are listed in a neip putiiad formula (ll. 6–8), which is analogous to Latin ne possit and hence the non permittas formula (for the Oscan, see Mancini, 2006: 78; for the Latin formula, cf. 443 ex multis). Furthermore, the divine vengeance should also affect usurs inim malaks nistrus (l. 2), that is the wife, children and relatives of the thief (for parallels, see 246, 370 and 57). Pakim Kluvatiium is offered as a sacrificial victim (sakrim, A, l. 11; cf. 449).

Unsurprisingly, the interpretation of the tablet has been rather controversial: while Poccetti has identified it as a ‘prayer for justice’, Mancini has claimed it is an ‘authentic’ defixio. For the former, the text presents the structure of a prayer, whose characteristics are mixed with legalese. Following Bücheler’s lead, Poccetti has compared this tablet to the corpus from the sanctuary of Cnidus on the following points: in both similar deities are summoned; the authors of the tablets from Cnidus are always women, who turn to the deities to state their name and ask for justice (this point is only valid, of course, if Vibia is the defigens); the texts have similar structures, which contain singular verbs in the first, second and third person forms. Mancini, in contrast, has been unwavering in his assertion that this text is a curse, since it contains ‘strutture ricorsive tipiche del genus defissorio’ (2006: 77). Among these features we can count the qualification of the victim with a matronymic, the delivery of the victim to a deity (manafum (l. 1) is equivalent to Latin mando or commendo) and the use of the abovementioned neip putiiad formula. If we reject the rigid separatation between curses and ‘prayers for justice’, the text can be identified as a defixio either way (for the debate about ‘prayers for justice’, see section I.7.6.1).

According to Mancini (2014: 51), the text presents the following orthographic and phonetic features, all of which point towards a date around the end of the fourth century BCE: the loss of n before a consonant (aretikai, l. 12), loss of final m (limu, puklu and suvạ, A, ll. 8, 12 and B, l. 2, respectively), loss of j in the diphthong o:j (puklu for puklui, A, l. 12) and loss of final d (in the subordinating conjunctions svại puh and suluh, A, ll. 10 and 9). Furthermore, the same author has highlighted the position of the demonstrative pronoun ulum (l. 3), which constitutes ‘una variante marcata dell’ordine sintattico attribuibile al registro espontaneo’ (Mancini 2014: 53), thus providing evidence for spoken Oscan.

After being inscribed, the tablet was rolled up. Translation (Murano 2012: 643–44): ‘To Keres Arentika I have entrusted, who, with respect to any their own will, their own cohort (of demons) shall invoke, [she shall smite (?) him (?),]/ the women and the children, the relatives. With respect to Pacius Clovatius, son of Valaima, who for hatred Damia [own will, her own]/ cohort shall invoke. This I to you have entrusted; (I have

153

Sylloge of  Defixiones from the Roman West entrusted) to Vibia Preba so that she may deliver to Keres Arentika, [(him,) Pacius Clovatius,]/ son of Valaima, and to her cohort. If she does not give him, harshly shall be punished her (...)/ and you shall be smitten and wounded, otherwise, but if you have done it, then after that (...)/ when he may start, but may not..., when starts or if must perform something (...)/ may; neither with infernal sacrifices neither with superior ones may they, whatever is good they may [...to Pacius Clovatius]/ son of Valaima. When he shall start the meal, may he not be

able to eat nor to crush the food (??)/ of which living men nourish themselves. May Pacius Clovatius, son of Valaima, absolutely tremble (??)/ to Vibia Aquia, if you absolutely shall have offered Pacius Clovatius, son of Valaima above [the earth; to your will]/ and to your cohort, if as sacrificial victim you shall have offered him under the earth and under the water, [to Pacius Clovatius,]/ son of Valaima or to Keres Arentika or to her cohort (...)/ To Keres Arentika, who with respect to him who her own will, her own cohort.’

69. Salernum, Salerno Provenance: necropolis. Current Location: private collection. Date: 2nd century CE. Material: lead. Measurements: 8 × 6 cm. Reading: 5



Locus · capilloribus (!) · expectat · caput · suum

Bibliography: Garrucci 1866: 28; CIL X, 511; Mancini 1884–86: 81; CLE: no. 205; Wünsch 1897: 27; DT 210; Fox 1912b: 305; Jeanneret 1917: no. 210; Inscr. It. 1, 241; Kropp 2008: no. 1.5.5/1.

other examples of curse tablets that were placed together with something that was associated to the target, such as hair or threads of clothing, see Bevilacqua et al. 2012: 236). The interpretation of ribus (l. 2) has proven more controversial, which has been signaled out by authors like Mommsen who has argued ‘Capillum devoti hominis in rivum demersum necesse est caput sequatur’ (apud CIL; Fox, who understood the text as an example of sympathetic magic, agreed). That said, as Mancini has shown, the presence of the interpunctuation shows that – ribus is the ending of the preceding capillo-. Accordingly, it should be analysed as capilloribus (ll. 1–2, for capillis). In that case, Mancini has argued that the place where the hair will be put (i.e., the tomb) will also be the place in which the head of the victim should be (1884–86: 81).

Commentary: this defixio, according to Garrucci, was found in a tomb in the area of Salerno on an unknown date and under unknown circumstances. The curse is inscribed on a lead sheet. The text contains six lines, which run from left to right and were written in capitals very similar to those found on the seals of tegulae from the Antonine period. The words are separated with interpunctuation consisting of a single dot. We do not know the circumstances that led to the writing of this curse, which, according to Garrucci, ‘forse in origine si faceva con un ciuffo di capelli onde prese il nome di capillo’ (1866: 28). This would imply the use of ousía or ‘substance’ during the deposition of the tablet (for

Translation: ‘The place in which the hair (has been placed) is waiting for his/her head.’

70. Cales, Calvi Risorta Provenance: necropolis. Current Location: M. Capuano. Date: 1st century CE. Material: lead. Measurements: vid. commentary. Reading:

154

Italia A Dite inferi · C(aius) · Babu l(l)ium · et · Fo(n)t{r}ins Tertia · Salvia

B Dite [pr]om[i]s(s)um Quartae Satiae recipite inferis

Bibliography: Sogliano 1883: 518; Mancini 1884–86: 73–75; Wünsch 1900: no. 8; DT 191; Diehl 1910: no. 855; Jeanneret 1917: no. 191; Kropp 2008: no. 1.5.1/1; Urbanová 2018: no. 29. Image source: Mancini 1884–86, Tab. III, fig. 1a and 1b.

(2 cm wide and 9, 9.5 and 8.5 cm long, respectively) that are attached to form a hollow square missing its right side. An opisthograph, the text was written in capitals that run in two continuous lines following the shape of the tablet. On side A, the words are separated with interpunctuation consisting of a short vertical stroke halfway up the letter.

Commentary: this defixio was discovered during the excavations carried out at the necropolis of ancient Cales, which dates to roughly the first century CE. The curse, which Audollent says was found ‘inter rudera humilis sepulcri iuxta monumenta aetatis Tiberianae’ (DT 191, p. 252), was written on a very irregularly shaped lead sheet that had been repurposed: it consists of three long segments

The circumstances that led to the writing of this curse, in which four victims are handed over to Dis Pater, are unknown. The lexeme [pr]om[i]s(s)um, as Wünsch has pointed out, ‘nehmt die Verheißung der Quartia Satia in die Unterwelt’ (1900: 238). Babul(l)ius is a well attested nomen in Campania, while Satia is found in CIL X, 4989a and in CIL XIII, 2125.

71. Pompeii, Pompei Provenance: necropolis. Current Location: Ufficio Scavi di Pompei. Date: 2nd century BCE. [Inv. No.: 17070 and 17071] Material: lead. Measurements: vid. reading. Reading: 5



10

IA (Measurements: 5 × (5.3) × 0.4 cm) ++++++SSVN · (h)oc · prịṃ[um] P(hi)lematio · Hostili · facia(m or -s?) capil(l)u(m) · cerebru(m) · flaṭus · ren[es] ut il(l)ai · non · suc(c)edas+[---] qui · PRAEC++++ · odiu(m) +[---] ut · il(l)ic · il(l)a(n)c · odiat (!) ·como(do) (h)aec · nec · acere · ne(c) · il(l)aic qui(c)qua(m) · acere · pos(s)it · ul(l)ọ[s] [fili]os · P(hi)lematio · Họst[ili] [------]

155

Sylloge of  Defixiones from the Roman West IB ----- 5

IIA (Measurements: 4.9 × 8.8 × 0.2 cm) [---]IDA · fiat+++ nec · acere · nec · lin[---] ul(l)a(s) · res · pos(s)it · pete[re] quas ego · (h)umaṇ[o---] comodo · is · eis · desert[us] il(l)aec · deserta · sit · cun(n)o (?) a(nte) · d(iem) · IV · c(alendas) {C} · N(ovembres?) · I · dificdos a dic · il(l)aec · deṣẹr[ta?---] (Left margin: DIC · ilai)

IIB ------ comodo+++ ------

Bibliography: Della Corte 1916: 304–07; CIL IV, 9251; CIL I², 2541; Sabbadini 1918; Preisendanz 1930: 139; Rubio and Bejarano 1955: no. 186*; García Ruiz 1967: no. 14; Solin 1968: no. 39; ILLRP 1147; Varone 1998: N32; Kropp 2008: no. 1.5.4/1; Urbanová 2018: no. 33.

During an autopsy of this curse in September 2010, I observed that both tablets are so corroded that some letters are now illegible (see the underlined text in IA ll. 6, 8–9 and IIA, ll. 1–3). Furthermore, the first tablet has lost both of its right corners and now its length measures 2.1 cm less than it when first discovered (when it was 7.5 cm across). The second tablet has also lost its two right corners and is now broken into five parts which fit together. The lead band that held the tablets together was broken and is now lost.

Image source: Della Corte 1916: 305, fig. 16. See also App. IV.3, SD 71. Commentary: this defixio was discovered c. 1911 during the excavations of the Pompeiian necropolis located outside of the Porta Stabia. The curse, according to the editor, was found ‘in vestibulo sepulcri’ (apud CIL IV), ‘quasi a fior di terra, davanti alla columella anepigrafe della tomba 10’ (Della Corte 1916: 304), in the tomb of Marcus Epidius Dioscurus.

Opisthographs, both sheets have texts on side A that run longitudinally from left to right along a guiding line. The inscription contains at least 19 lines, written in capitals (measuring between 0.2 and 0.7 cm in height). The words are separated with interpunctuation consisting of a single dot found halfway up the letters. On both reverse sides, the text runs vertically from left to right, though they are so corroded that it is almost impossible to read anything. About these lines, Della Corte (CIL IV) maintained, ‘Pauca verba, quae intellegere potui, modo adnoto: c6 vestigia, 7 nervos; d6 comodo, 9 Hostili, 10 malitico (?), 12 in odiu’, while Degrassi has added that ‘Scripturae exterioris vestigia quaedam agnoscuntur in quibus Vestilia et Hostili’ (apud ILLRP).

The curse was written on two lead sheets, which emulate tabullae ceratae. At the time of their discovery, they were joined together and closed (like a diptych) with an inscribed lead band and two small nails, all of which are now lost. There are no known parallels among the curses from the Roman West for this shape, which adopts a tool of everyday writing and adapts it for a magical purpose in lead. Such a procedure may reflect a different conception of curse tablets themselves: here the practitioner equates the text with a letter. There are several factors that support this hypothesis: first, the shape and dimensions of the medium; second, the fact that the piercing of the tablet does not have a magical purpose (but rather a practical one, to make a diptych); third, the use of guiding lines to ensure that the text is straight (which may reveal that the author was not very used to writing); fourth, the fact that in IIA the author run out of space and had to write in the margins of the main text (as was usual in letter writing, see Sarri 2018: 112). All these considerations would match well with the conception of defixiones as sorts of letters. Of course, several texts explicitly make this connection, since the curse tablet is called a c(h)arta (see 96, A, ll. 1–2 and 484, B, l. 11) or epistula (see 518, B, l. 4 and 534, B, l. 5).

The text presents the following orthographic, morphological and phonetic features: the simplification of the geminate in posit for possit (IA, l. 8 and IIA, l. 3), ula for ulla (IA, l. 8 and IIA, l. 3), capilu for capillu (IA, l. 3), cuno for cunno (IA, l. 5), sucedas for succedas (IA, l. 4); the loss of final -m in facia (IA, l. 2), cerebru, capilu (both in IA, l. 3) and odiu (IA, l. 5); and loss of initial h- in uman (IIA, l. 4), oc (IA, l. 1) and aec (IA, l. 7). Additionally, Plematio is for Philematio (IA, ll. 2 and 9), facia for faciem (IA, l. 3), ilai for illa (IA, l. 4, with simplification of geminate), comodo for quomodo (IA, l. 6, IIA, l. 5 and IIB, l. 1), acere for iacere (IA, ll. 7–8 and IIA, l. 2) and quiqua for quicquam (IA, l. 8, with loss of final -m). Finally, the following forms are of note: odiat (for oderit, IA, l. 6), dificdos (for defixos, 156

Italia The second compares the context of the deposit, a grave, to the curse’s victim: comodo is eis desert[us], il(l)aec deserta sit cun(n)o (IIA, ll. 5–6). This can be understood as either a reiteration of the defigens’ previous request for Philematio’s infertility or as an attempt to isolate her emotionally and romantically. The final phrase in the inscription has defied interpretation due in part to the poor state of conservation.

IIA, l. 8) and the various forms of the pronoun ille: ilic (for illic = ille, IA, l. 5), ilaec (for illaec, IA, l. 6), ilai (for illai, IA, l. 3), and ilac (for illanc, IA, l. 5). Based on its content, this curse should be identified as an erotic defixio. In this case, the text targets Philematio, the slave of Hostilius, and his lover. The defigens attacks both her appearance (facia(m), capil(l)u(m); IA, ll. 2–3) and vital organs (cerebru(m), flaṭus, renes; IA, l. 3) in an attempt to bring her relationship to an end (ut il(l)ai non suc(c)edas; IA, l. 4) and convert love into hate (ut il(l)ic ila(n)c odiat; IA, l. 6). The presence of comodo (IA, l. 6, IIA, l. 5 and IIB, l. 1) suggests that the text contained three similia similibus formulae, of which two have been preserved. As Varone has pointed out, the first seeks to make the victim barren: ne(c) il(l)aic qui(c)qua(m) acere pos(s)it ulọ[s] [fili]os (IA, ll. 8–9).

Translation: ‘To Philematio, (the slave of) Hostilius, (I curse? her) face (...) hair, brain, breath, kidneys, so that (relationship?) may not prosper (...) may he hate her. Just like this one does not lie, may nobody lie with her (...) To Philematio, (the slave of) Hostilius, not lie not (...) cannot reach business (...) just as this (place) is empty for these, may for her (...) empty. The ninth day before the Kalends of November (these were) cursed...’

72. Pompeii, Pompei Provenance: necropolis. Current Location: in situ. Date: 1st–2nd centuries CE. Material: slate. Measurements: 38.8 × 80.3 cm. Reading: 5



hospes · paullisper · morare si · non · est · molestum · et · quid · evites cognosce · amicum · hunc · quem · speraveram · mi · esse · ab eo · mihi · accusato res · subiecti · et · iudicia · instaurata · deis · gratias · ago · et · meae · innocentiae · omni · molestia · liberatus · sum · qui · nostrum · mentitur · eum · nec · Di · Penates · nec · Inferi · recipiant ·

Bibliography: Ciprotti 1963: 279–80; Ciprotti 1964: 304–05; AE 1964, 160; De Caro 1984; Elefante 1985; AE 1986, 166; Pignataro 2006; AE 2006, 291; Kropp 2008: no. 1.5.3/3; Van Andringa et al. 2008: 379–82; AE 2008, 323; Lepetz and Van Andringa 2011: 117–19; Williams 2012: 260–66; Van Andringa et al. 2013: 998; Urbanová 2018: no. 209.

wearing togas. After the monument was erected, a second inscription was added below the titulus sepulcralis. This inscription is the present defixio. The curse was inscribed on a dark stone plaque, which juts out from the wall of the tomb to which the defixio was attached with five iron nails. The text is intact and in good condition, as I was able to verify during an autopsy in 2010. The defigens’ choice of medium here clearly reflects the inscription’s function, which was meant to resemble lead and was destined to be displayed publicly. The only parallel in the Roman West for a publicly displayed curse comes from Merida and was written on a marble plaque (120; for a more detailed discussion of these two items, see Sánchez Natalías 2016b).

Image source: Van Andringa et al. 2013: fig. 252. Courtesy of W. Van Andringa. See also App. IV.3, SD 72. Commentary: this defixio was discovered on the façade of tomb no. 23 OS from the necropolis of Porta Nocera (Pompeii). The funerary monument is an ‘edicola’ and is 6.5 m tall (for the architectural features of the monument, see De Caro). As we learn from the monument’s first inscription, its commissioner was the wealthy freedman Publius Vesonius Phileros, who was a member of the collegium of the Augustales. He built the tomb for sibi et suis Vesoniae P(ubli) F(iliae) patronae et M(arco) Ofellio M(arci) L(iberto) Fausto amico. These three individuals are depicted in the monument’s sculptural group with the patroness flanked by the two freedmen, who are depicted

The curse contains eight lines of capitals, which run from left to right. The size of the letters decreases towards the end of the text (hence in l. 1 they are 3.7 cm in height, whereas after l.5 they measure 2.5 cm). The layout was carefully planned and left room in l. 6 for an iron nail, which, as Elefante has rightly argued, has no functional

157

Sylloge of  Defixiones from the Roman West purposes (i.e., is not meant to hold the inscription to the wall), but rather is related with a magical purpose (1985: 433). This unusual curse begins with the classic formula hospes paullisper morare si non est molestum (ll. 1–2), with which Phileros invites passersby to stop and read his story. He tells how Faustus, whom he had thought was his friend (amicum hunc quem speraveram mi esse, ll. 3–4), betrayed him by making false accusations and dragging him to court. Despite this betrayal, Phileros came out on top and thanks the gods for their good fortune (Deis gratias ago, ll. 5–6). He ends by cursing his old friend: qui nostrum mentitur eum nec Di Penates nec Inferi recipiant (ll. 7–8).

the words meae and innocentiae (l.6). This serves to stress Phileros’ innocence. Recent excavations (Van Andringa et al. 2008: 382; Lepetz and Van Andringa 2011: 119) have revealed that the space originally meant for Faustus was destroyed back in Antiquity: the slab sealing the graves was broken, the stele with his inscription was defaced and the tube for libations as well as his urn were filled with plaster. But that is not all: What had been Faustus’ tomb was sealed off with a new floor, in which black stones with the inscription Phileros were laid. In this way Phileros reclaimed his space as his own. It is likely, as Van Andringa (et al. 2008: 382) has suggested, that the defigens knew that the destruction of a grave was not considered a profane act against the ius sepulcrale (and hence not punishable by law) as long as the dead body had not been laid to rest. So even though Phileros could neither erase Faustus’ name from the epitaph nor take down his statue, nothing kept him from destroying the grave itself.

Accordingly, this defixio seeks to leave Faustus Ofellius unburied, a punishment which—though it is unparalleled in the collection of Latin curses—undoubtedly fits well into the larger realm of magical practices. Even though Phileros was victorious in court, legislation prohibited him from kicking his erstwhile friend out of the tomb. Given the situation, he chose to publicly curse his enemy: according to Elefante, ‘l’iscrizione stessa, studiata in ogni dettaglio per passare come lapide sepolcrale, rivela l’infamia di quel defunto indegno anche di essere chiamato con il suo nome ed indirizza nei suoi riguardi una maledizione’ (1985: 442). As this scholar has rightly pointed out, the focal point of this curse is the central nail, which is bigger than the rest and is given a pre-eminent position right between

Translation (Williams 2012: 264): ‘Stranger, stay a moment if it is no trouble to you, and learn what you should avoid. This man, whom I had hoped was my friend, suborned witnesses against me and initiated proceedings. I thank the gods and my own innocence; I was set free from all associated risks. Whoever of us is lying may both the Household Gods and the gods below reject him.’

73. Aquinum, Aquino Provenance: necropolis. Current Location: unknown. Date: 2nd century CE. Material: lead. Measurements: 11 × 11 × 0.4 cm. Reading: 5



10



15

M(arcus) Satrius M(arci) fili(us) his (h)ostis M(arci) Luteini Simedi M(arci) f(i)l(ii) siu{s} invidia damno tibi sic unio ne sit tibi i his ossis ut acuis [---]RI i is ossis in s˹uis˺ ++[---]++++ Satriu(s) i is ossis sic [-c.5-] nesu(s) ringensib(us) a canibus +I++++++O sit [---?] idem st pinem [---]no memnist(i) mi +N+OSI S+A++G++ sua [---] (unum) (sestertium) dedi S[-c.3-]NE+++[---] duxiseris AS+++[---] omnes [---?]

Bibliography: Zagarola 2012; AE 2012, 325.

which is nearly intact, despite a good deal of corrosion. It has lost its lower-right corner and has two holes in its top half. The inscription contains 15 lines, which run from left to right and were written in capitals, measuring between 0.4 and 1.6 cm in height. The text has been dated to the second century CE on palaeographic grounds.

Commentary: this defixio was discovered in 2007 during the excavations of the western necropolis in ancient Aquinum. It was found, still folded, in grave no. 14. The curse was written on a roughly rectangular lead sheet, 158

Italia The circumstances that led to the writing of this curse are unknown. What we learn is that Marcus Satrius, driven by envy, curses Marcus Luteinus Simedus (ll. 1–4). The editor has proposed that this tablet is erotic in nature based on the reading pinem (l. 10), though there is not sufficient information to support this hypothesis. The reading of the tablet is highly problematic and cannot be confirmed or corrected from the available images. Furthermore, if we accept the editor’s reading, there appear to be grammatical and morphological problems.

Apparently, his body is to be torn apart by dogs (ll. 9–10). This punishment would also appear here for the first time. According to the editor (despite the lack of any parallel), the second person singular verb in the final phrase refers to the dead person in whose grave the curse was deposited and to whom the defigens offers a sestertium (l. 13) for carrying out the curse. Even though this hypothesis is suggestive, this reading seems highly problematic (and should not be accepted without further study and a new autopsy).

The text includes the following orthographic and phonetic features: ostis for hostis (l. 2), sius for suis, is for his (ll. 6–7), nesu for nasus (l. 8), pinem for penem (l. 10) and duxiseris for duxeris (l. 14). The word ostis (l. 2) is rare in the corpus of curses from the Roman West (cf. 491 and 103) and not used in erotic curses. Here, if we follow the editor, it is applied to a rival in a love affair. The curse is anatomical (ll. 5–10), taking aim at the victim’s bones (ll. 5–7; cf. 51), nose (l. 8; cf. 56) and, perhaps, penis (l. 10; this would be unparalleled).

The nomen Satrius (ll. 1 and 7) is widely attested in Rome, while Luteinus (l. 2) is unparalleled, unless we follow Zagarola and take it as a corrupted form of Latinus. The cognomen Simedus (l. 3) is unparalleled. After being inscribed, the tablet was folded and pierced by an iron nail (measuring 20 cm in length and 0.5 cm in diameter), which resulted in the two holes. It may have been pierced a second time, leaving a single slot on the left side.

Regio III [Bruttium et Lucania] 74. Tiriolo Provenance: necropolis. Current Location: MANP Catanzaro. Date: 4th century BCE. Material: lead. Measurements: 1.5 × 7 cm. Reading:

Τρεβας Τρεβατιες Νυμψιμ Αλαϝιομ

Bibliography: Fiorelli 1882: 395; Pugliese Carratelli 1951; Pugliese Carratelli 1953; Pisani 1952; De Franciscis and Parlangeli 1960: 48–53; Poccetti 1979: no. 190; Rix 2002: Lu 43; TDO 14; ImIt: Teuranus Ager 1; Lazzarini and Poccetti 2017: 230–31.

in the accusative, which have been interpreted as the defigens and his victim respectively (Pisani 1952: 291; Pugliese Carratelli 1953; Poccetti 1979: 142; ImIt: p. 1478). Later on, Poccetti changed view (1990: 140), considering that ‘the morphological opposition does not necessarily reflect a syntactical opposition, so that the morphological forms in these cases are not sufficient for a syntactical understanding of the sentence; the morphological opposition could instead be used to topicalize certain personal names as opposed to others’ as explained by Murano (2012: 649). Murano (apud TDO), for her part, following this idea, considers then that both names can be interpreted as the victims of the curse. For a discussion on the combination of nominatives and accusatives in the curse see section I.7.2. Both Τρεβας (l. 1) and Νυμψιμ (l. 2) are well attested praenomina in the Oscan epigraphy, while the nomen Αλαϝιομ (l. 2) would also be attested in the Paelignian Alafis (see Vetter 1953: no. 216).

Image source: Pugliese Carratelli 1951: 2. Commentary: this defixio was discovered in 1880 during the excavations carried out in Tiriolo and according to Fiorelli ‘ravvolta in una tomba’ (1880: 395). The curse was written on a lead strip, which has been broken into two fragments that fit together. The text, written in the Oscan language and the Greek alphabet, contains two lines, which run from left to right. The inscription contains two bipartite onomastic formulae consisting of a praenomen and nomen. The first name is in the nominative, while the second is

159

Sylloge of  Defixiones from the Roman West 75. Crimisa, Cirò Marina Provenance: necropolis (?). Current Location: MAN Reggio Calabria. Date: end of the 4th–beginning of the 3rd century BCE. Material: lead. Measurements: 3.5 × 8.5 cm. Reading:

στατις πολλιες κερρινομ οριομ μαισιμεσμαιμοποπεδ

Bibliography: Terminelli 1972: 103; Poccetti 1979: no. 189a; Poccetti 1984; Rix 2002: Lu 44; ImIt: Crimisa 3; TDO 12; Lazzarini and Poccetti 2017: 231–33.

praenomen κερρινομ (l. 2) could be analogous to the Latin nomen Cerrinus, while the nomina πολλιες and οριομ (ll. 1–2) are related to Latin Pullius/Pollius and Horius. According to Poccetti, who bases his analysis on analogy with Greek defixiones, the combination of nominative and accusative names would not be related with the different status of the individuals (defigens versus victims) but just differentiates between several types of victims: ‘marca non tanto la contrapposizione tra defiggenti e defissi, quanto la distinzione tra diverse categorie di defissi’ (1984: 83; on this, see commentary for 74). For a discussion on the combination of nominatives and accusatives in the curse see section I.7.2. If we follow the editor, l. 3 contains a cursing formula, in which the final οποπεδ is a verb, which belongs to the group of verba defigendi. The sequence μαισιμεσμαιμ has been analysed as a vox magica (Poccetti 1984: 86; Murano TDO p. 191 agrees). Nevertheless, considering that voces magicae are usually found in much later texts, this interpretation ought to be reconsidered.

Image source: Poccetti 1984: 75, fig. 2. Courtesy of P. Poccetti. Commentary: this defixio was discovered in early 1970s during an illegal excavation near the temple of Apollo Alaios at Punta Alice (Crimissa). The item was found inside of a grave dated between the end of the fourth century and the middle of the third century BCE. The curse was written on a rectangular lead sheet, which has been broken into three parts. According to Murano (TDO, p. 190), it is now missing its bottom-right corner. The text, written in the Oscan language and the Greek alphabet, contains three lines, which run from left to right. We do not know the circumstances that led to the writing of this curse, which consists of two personal names followed by what could be a cursing formula. The

After being inscribed, the tablet was folded in half.

76. Roccagloriosa Provenance: sanctuary. Current Location: DAR. Date: 4th century BCE. Material: lead. Measurements: 3.8 × 11 × 0.8 cm. Reading: 5



δυϝο[-c.2-3-]διμνοπολε`ν´τ[-c.6-7-]ει[-c.3-4-]σμετ[-c.2-3-]ανισδ (vacat) υ[-c.3-4-]ερισπολλ[-c.2-3-]σ (vacat) (vacat) [γ]αϝισ̣φοινι[-c.2-3-]μαχιεσ (vacat) μαμε[ρ]εξ[-c.3-4-]ϝιδισ (vacat) γανα[---] (vacat) (vacat) πακισ[---] (vacat) (vacat) αν̣τ[---] (vacat) (vacat) μα[---] (vacat)

160

Italia Bibliography: Gualtieri 1990; Poccetti 1990; Campanile 1993a; Rix 2002: Lu 45; ImIt: Buxentum 3; TDO 7.

personal names. Thus, the tablet would have been divided into two different sections, according to its content: while the left half was used for the cursing formula; the right half contained the list of victims (Poccetti 1990: 145–46). This layout, with a list of personal names arranged in a column located on the right side of the tablet, would highlight its importance, making it easier to read (Poccetti 1990: 146; for the layout, also see TDO, p. 159).

Image source: Poccetti 1990: 143, fig. 127. Courtesy of P. Poccetti. Commentary: this defixio was discovered in 1977 during the excavations of the so-called ‘complesso A’ in Roccagloriosa. The curse tablet was found rolled up and placed in the south-eastern corner of this structure’s portico. It had been deposited in an area filled with votive offerings, which has been identified as a sanctuary (for the context, see Gualtieri). For the remains of the 13 lead sheets discovered during the same excavation, see the introduction to this chapter (s.v. Roccagloriosa).

When preserved in full, the names of the victims are in the nominative case. Following Poccetti, it is possible to reconstruct μετ[ς] or μετ[ις γρ]ανις in l. 1, while ll. 2 and 4 appear to contain bipartite onomastic formulae consisting of a praenomen and nomen, which Poccetti has reconstructed as επις πολλ[ιε]ς and μαμε[ρ]εξ [---]ϝιδις (the latter could be for [σαλα]ϝιδις or [καλα]ϝιδις, among others). The reconstruction of the onomastic formula in l. 3 has proven to be much more controversial (see Murano apud TDO, pp. 158–59 for a good synthesis). Thus, Poccetti reconstructs [Γ]αϝις φοινι[κις] μαχιες (following the CampanianSamnitic formula) or [Γ]αϝις φοινι[κεις] μαχιες (following the Volsco-Umbran formula). Campanile (1993a: 370), for his part, has taken φοινι[---] as an ethnonym, in which case this formula would designate a foreigner and would not follow the habitual Oscan pattern. This victim would be referred to as ‘Gavio the Phoenician’. Following this line of thinking, Campanile has argued that the term μαχιες is a transliteration of the Semitic mhy’s (‘Iside è la mia guida’), a suggestive hypothesis that given the fragmentary nature of the text cannot be confirmed. Finally, Poccetti has also suggested different reconstructions for γανα[---] (l. 5, either εγανα[τς] or εγανα[τις]), αν ̣τ[---] (l. 7, part of a Greek name like ἀντι, such as Ἀντ[ισθένης] or Ἀντ[ίμαχος]) and μα[---] (l. 8, either μα[μερεξ] or μα[ρασ]). Regarding the praenomina, [γ]αϝισ̣, μαμε[ρ]εξ and πακισ (ll. 3, 4 and 6, respectively) are well attested in Oscan, whereas ερισ (l. 2) is derived from the nomen Heriis.

The curse was written on a rectangular lead sheet, which is now broken into five pieces that coincide with the fold lines and fit together. The tablet’s surface is quite corroded. The inscription contains eight lines, which run from left to right in scriptio continua. The text was written in the Oscan language and Greek alphabet (the letters measure 0.3 cm in height) and has been dated to the fourth century on palaeographic grounds. The text’s layout is notable, since only l. 1 runs across the entire upper margin, while the rest forms a list on the right side of the tablet. The interpretation of this first line has proven so controversial that Poccetti himself has actually wondered whether this line belongs to a different inscription (i.e., the tablet had been reused), or it contains a cursing formula. The first hypothesis, which claims that the tablet was reused, is not only based on the strange layout of the text, but also on the fact that several Greek words can be picked out of l. 1 (also see the interpretation of Crawford apud ImIt). Rix has actually gone so far as to not print l. 1 in his edition (cf. also Poccetti 2015: 386). If it is part of this inscription, the first line would contain in the first half the imprecatory formula, while its second half could actually contain the beginning of the list of

After being inscribed, the tablet was pierced twice and then rolled up.

77. Castiglione di Paludi Provenance: on the ground (public area). Current Location: MNA Sibaritide. Date: second half of the 4th century BCE. Material: lead. Measurements: 2 × 9.3 cm. Reading: A Πακ Νομψις Τρεβι[-c.1?-]μ Λοικ Αϝιμ[-c.1-]ρτορ[-c.2-3-] μες B Μ[-c.1?-]ις Βιϝιδις Λοικες [---]ομμασιμελ+ομ [---?]

161

Sylloge of  Defixiones from the Roman West Bibliography: Poccetti 1993c; Rix 2002: Lu 47; TDO 11; Lazzarini and Poccetti 2017: 233–37.

between both cases has been explained by Poccetti (and later by Murano) not as a differentiation between the defigens and victims but as a way to mark a difference between the types of victims (on this, see commentary for 74 and section I.7.2). Side A contains two personal names in the nominative, Πακ(ις) and Νομψις (l. 1), and at least three in the accusative: Τρεβι[-c.1?-]μ (l. 1; reconstructed also as Τρεβι[ο]μ, is a personal name already attested as a praenomen and nomen), Λοικ(ιμ) (first found here), Αϝιμ (or αϝιμ) and finally, according to the editor’s reconstruction, [Α]ρτορ[ιομ], which is similar to Latin Artorius (all found in l. 2). Side A concludes with what could be a cursing formula: μες (l. 3), analogous to μασιμες found in 75.

Image source: Poccetti 1993c: 215, fig. 1. Courtesy of P. Poccetti. Commentary: this defixio was discovered on the ground together with uninscribed material in the public area (close to the so-called ‘theatre’) of an ancient Bruttian settlement at Castiglione di Paludi (Cosenza). The time and exact circumstances of discovery are unknown. The curse, written on a strip of lead, is broken into six fragments, five of which fit together. The tablet’s surface, which has recently been restored, shows the remains of corrosion only on side A of the final fragment. Also note that on one of the edges of side A there is a mark which, according to the editor, likely resulted from an attempt to pierce the tablet.

Side B begins with an onomastic formula made of a praenomen, nomen and patronymic in the nominative: Μ[-c.1?-]ις Βιϝιδις Λοικες. L. 2 has proven more difficult to analyse: according to Poccetti, this could either be personal names or a cursing formula, depending on how the words are separated. If we opt for the first option, the line would be broken up as [---]ομ Μασιμ Ελ+ομ (a series of names in the accusative) or [---]ομ Μασ Ιμελ+ομ (an accusative, nominative, accusative). If we choose the second option, Poccetti explains, ‘si può solo annotare, in via puramente ipotetica, che tanto l’elemento iniziale come quello finale del rigo (entrambi con l’uscita [---]ομ), sono teoricamente riconducibili ad un verbum defigendi’ (1993c: 230).

An opisthograph, the inscription contains five lines, which run from left to right and were written in the Greek alphabet (the letters measure 0.4 cm in height) by a trained hand, who was able to draw the same letter in different ways (compare the sigma in Νομψις and μες, A, ll. 1 and 3, respectively), use abbreviations (Πακ(ις) and Λοικ(ιμ), A, ll. 1–2; for a different take, see TDO, p. 185) and present a careful layout. The circumstances that led to the writing of this curse are unknown. The curse is directed against a group of at least seven individuals, whose names appear in nominative or in accusative. Again, the opposition

After being inscribed, the tablet was rolled up.

78. Laos, Marcellina Provenance: necropolis. Current Location: MAN Reggio Calabria. Date: 4th–3rd centuries BCE. Material: lead. Measurements: (14.5) × 6.3 cm. Reading: A Col. I Μαραεν Οϝι Σαβιδιν Νοψιν ϝαρϝαριες 5 Σπεδι Ϝιβιν Μαραειν Λοικιν



Col. II Γαϝιν Οψιν μεδεκον Οψιον Νοψιν Σαβιδιον μεδεκον Ϝιβιν Σπελιν

162



Col. III Στατιν Οψιον μεδεκον Ϝιβιν Βοθρονι [---?]

Italia B Νοψα Ϝαριαν Ϝιβιαν Σπελαν μεδεκαν αραδιαν

Bibliography: Pugliese Carratelli 1992; Campanile 1993b; Campanile 1993c: 56; Cassio 1993; Guzzo 1993; Greco 1993; D’Agostino 1993; Mele 1993; Poccetti 1993a; Poccetti 1993b; Silvestri 1993; Rix 2002: Lu 46; Murano 2006; TDO 8; ImIt: Laos 2; Lazzarini and Poccetti 2017: 250–55.

whereas the females are found on side B. To date, this sort of organizational strategy is only found here. In addition, it is worth noting that all the names are in the accusative, with the exception of Ϝαρ ϝαριες (A, I, l. 4), which is in the nominative (see below and commentary to 74 for a different hypothesis). For a discussion on the combination of nominatives and accusatives in the curse see section I.7.2.

Image source: Poccetti 1993b: 155. Courtesy of P. Poccetti. Commentary: this defixio was discovered in the fall 1963 during the excavations carried out at the necropolis of ancient Laos (Calabria). The grave, which probably held the remains of two individuals, boasted an incredible collection of grave goods, which, as the studies of the pottery and belts have shown, date between 380 and 330 BCE. The most recent of these items (330 BCE) provides a solid terminus post quem for the collection. According to Guzzo, the tomb ‘fu scavata sotto il crollo della copertura della camera sepolcrale e non fu registrata la giacitura degli oggetti’ (1993: 116). For this reason, we do not know the exact place in which the defixio under consideration was deposited.

Within the list of male victims, we find the names Λοικις and Νοψις (A, I, l. 8 and II, l. 5; first attested in the curse from 77). Σπελιν and the feminine Σπελαν (A, II, l. 8 and B, l. 2) could be related to the praenomen Σπελληις and the Latin nomen Spel(l)ius. Βοθρονι[---?] (A, III, l. 3) could be an antecedent of Latin Butronius, which is derived from Greek Βόωρος. Μαραεν and Μαραειν (A, I, ll. 1 and 7) come from the same name, probably the Oscan praenomen Μαραhις (in which case it is necessary to correct at least the first form to Μαραεν). In the sequence Ϝαρϝαριες (A, I, l. 4), we can pick out the nomen Ϝαριες, which is also found at B, l. 1 (Νοψα Ϝαριαν). As Poccetti has pointed out, the list of females is much shorter than that of males. In this second list, we only find two bipartite onomastic formulae (with praenomen and nomen written in scriptio continua). Again, the contrast between nominative and accusative could suggest that Νοψα Ϝαριαν is one of the defigentes and Ϝιβιαν Σπελαν one of the victims (this name is clearly related to Ϝιβιν Μαραειν at A, I, ll. 6–7). The interpretation of μεδεκον (A, II, ll. 3 and 7 and III, l. 2) and μεδεκαν (B, l. 3) has proven more complicated. Two solutions have been suggested: according to Pugliese Carratelli (1992: 18), these words are analogous to meddix (rarely attested in the LucanoBruttian context). In contrast, these words could be forms of a verb equivalent to Greek καταδέω. Pugliese Carratelli (1992: 18–19) has interpreted ϝαρϝαριεσ (A, I, l. 3) as a vox magica, though Poccetti (1993b: 170) has rejected this proposal, rightly pointing out that such voces magicae are used at a later date and never in texts which contain nothing more than a list of names.

The curse tablet was written on a roughly rectangular lead sheet, which no longer has its original edges. An opisthograph, the inscription was written in the Greek alphabet and runs from left to right. The text is organized into three columns on side A, the third of which follows the tablet’s right edge, probably because the author miscalculated the space needed (cf. with 103 and 494, this is typical of letter writing, as pointed out by Sarri 2018: 112). Side B only has a single column. The text was written in Oscan by an individual of Italic origins who was not only bilingual but could also write in Greek (Poccetti 1993b: 179). Poccetti has convincingly argued that the text consists of a list of names of the victims, who are referred to either with a bipartite onomastic formula (praenomen and nomen) or simply with a praenomen. These two ways of referring to the victims, reflect, in Poccetti’s view, different cultural traditions that in this context form a bicultural society, in which two different systems for nomenclature coexisted (1993b: 164). Furthermore, the victims are organized according to gender: side A deals with the male victims,

After being inscribed, the tablet was folded five times. 163

Sylloge of  Defixiones from the Roman West 79. Laos, Marcellina Provenance: unknown. Current Location: MAN Naples. Date: 4th–3rd centuries BCE. [Inv. No.: 119925] Material: lead. Measurements: (3.5–5.5) × 10.5 cm. Reading: 5



Λοϝκις Συριες Πακι Πακιδιες Οϝι Συριες Γναι Ϝαδις Οταϝις Πετιδις [-c.1-]ιρις Ορτοριες [Γ]αϝ[ι-c.3-4-]αλανις [-c.12-14-]νις

Bibliography: Poccetti 2000: 745–61; Rix 2002: Lu 63; TDO 9; ImIt: Laos 3.

written in the Oscan language and the Greek alphabet, contains eight lines of scriptio continua, which run from left to right and whose letters are 0.3 cm in height. The tablet has been dated between the end of the fourth century BCE and the beginning of the third century BCE based on palaeographic grounds.

Image source: Poccetti 2000: 761, fig. 2a. Courtesy of P. Poccetti. Commentary: the Museo Archeologico Nazionale di Napoli acquired this defixio in 1890, where it is still found today. This curse was discovered during an illegal excavation in the area between the towns of Scalea and Verbicano, that is, in the area around ancient Laos. Though we have no secure information about the curse’s archaeological context, it is likely (as Poccetti has noted) that it came from a necropolis or was found on the ground.

The text comprises a list of eight personal names in the nominative, two of which belong to the same family (ll. 1 and 3). The individuals are named with a bipartite formula (consisting of praenomen and nomen). Poccetti has noted the omission of final -ς the praenomina Πακι, Οϝι and Γναι (ll. 2–4) as well as οϝ for the diphthong ou (in Λοϝκις and Οϝι, ll. 1 and 3). We must point out the nomina Συριες (ll. 1 and 3, derived from the Greek ethnonym Σύρος), Πακιδιες (l. 2, which gives us Latin Pacidius and Pacideius), Ϝαδις (l. 4, from which Fadius, attested from the Republican period onwards, is derived) and Ορτοριες (l. 6, likely linked to the attested (H)ortorius, and Greek Ορτωριος). Poccetti has reconstructed the praenomen [-c.1-]ιρις as [Δ]ιρις or [Ϝ]ιρισ, whereas Crawford (apud ImIt) has proposed [Τ]ιρις.

The curse was written on an irregularly shaped lead sheet, which was partly folded so that there was a 90 degree angle about 2 cm from the left edge. Furthermore, the tablet was also pierced near the top-left corner between ll. 2–3, where there is a hole with a diameter of 0.6 cm. Only the left edge of the tablet, which has some scratches affecting ll. 1, 2 and 4, has been conserved. The inscription, which is

80. Laos, Marcellina Provenance: unknown. Current Location: MAN Naples. Date: 4th–3rd centuries BCE. [Inv. No.: 119926] Material: lead. Measurements: 5–5.5 × 4.7–6 cm. 164

Italia Reading: 5

[-c.2-]μψιος Κα[-c.1-]αιος [-c.3-]ιος Ασελλιος Νυμψιος Ποππαλαιος Μινιος Ϝαριος [-c.2-]στος Ϝαριος

Bibliography: Poccetti 2000: 762–69; TDO 10; ImIt: Laos 4; Lazzarini and Poccetti 2017: 237–39.

of this curse, which consists of a list of names. The individuals are named with bipartite onomastic formulae composed of a praenomen and nomen. As the editor has noted, the names end is –os and hence use the Greek morphological ending instead of the Oscan one.

Image source: Poccetti 2000: 761, fig. 2b. Courtesy of P. Poccetti. Commentary: the Museo Archeologico Nazionale di Napoli acquired this defixio (and the previous curse) in 1890. As is the case with 79, we do not know the exact archaeological context in which the curse was found, though it did come from the area between Scalea and Verbicano (i.e., ancient Laos).

It is worth noting that in this text two members of the gens Ϝαρια are cursed (Μινιος Ϝαριος and [-c.2-]στος Ϝαριος, ll. 4–5), as was the case with other members of the same family in 78. Poccetti has proposed different readings of two problematic areas of the inscription. In the first, Κα[-c.1-]αιος (l. 1) is understood as either Κα[λ] αιος or Κα[δ]αιος (nomina similar to Latin Cal(l)ius and Cadius, respectively). The nomen Ποππαλαιος (l.3) can be connected to Latin Populeius/Poppelius/Pop(p)il(l)ius or even Pompalaius, if we accept that mp was simplified to p. The editor has reconstructed the praenomen [-c.2-]στος (l. 5) as [Σε(κ)]στος. Crawford, in contrast (apud ImIt) has proposed [τ]ετος.

The curse was written on an irregularly shaped lead tablet with an uneven surface and fold line between ll. 1 and 2. The inscription, which is written in the Oscan language and the Greek alphabet, contains five lines, which run from left to right in scriptio continua and whose letters measure about 0.2 cm in height. The tablet has been dated between the end of the fourth century or the beginning of the third century BCE based on its palaeographic characteristics. We do not know the circumstances that led to the writing

After being inscribed, the tablet was probably folded (see the abovementioned fold marks between ll. 1 and 2).

81. Petelia, Strongoli Provenance: on the ground. Current Location: MAN Crotone. Date: 4th century–beginning of the 3rd century BCE. Material: lead. Measurements: 2.8 × 18.4 cm. Reading:

Col. I Καϝνοτο Στατιο Πακϝιο Καιαιδο Πακολ Στατιες Μαρα(ς) Στατιες



Col. III Αϝες Αυδαις Νοϝιο Αλαφιο Μινακος Καφιριο Βαντινο Κοσσανω



Col. II Γναυ(ς) Στατιες Ϝιβι(ς) Στατιες Εμαυτο Στατιο Μιναδο Καιδικο τρεδοαυδαδο Μινας Καιδικις



Col. IV παϝελιοσνομο[---]νσετ ηισουσοσαραξ Μ[ι]νας Μινας καρισταππισπιτιμ σολλομ ηισου δέκεο hερμᾶ χθόνιε ταῦτα καὶ κάθεκε αὐτεῖ

5

165

Sylloge of  Defixiones from the Roman West

Bibliography: Lazzarini 2004; Poccetti 2006; Lazzarini 2009; TDO 13; ImIt: Petelia 2; Poccetti 2014: 86–102; Lazzarini and Poccetti 2017: 247–50.

nominative forms follow Oscan morphology, while the genitive forms use Doric Greek endings. As alluded to above, the onomastic formulae in I–II are bipartite and include a praenomen and nomen. It is worth pointing out, as Poccetti (2014: 93) has, that even if the Oscan names use Greek morphology (e.g., I. l. 2: Πακϝιο Καιαιδω), the inverse is not the case. When it comes to onomastics, the Oscan praenomina Καϝνοτο and Αϝες (I, l. 1 and III, l. 1, respectively) are unparalleled. Εμαυτο is a Greek praenomen, while Γναυ(ς) is equivalent to Latin Gnaeus (II, ll. 1 and 3), and Μινακος and Μιναδο are derived from the praenomen Μινας (II, l. 4, III, l. 3 and IV, l. 2). Lazzarini has pointed out that the nomina Αυδαις and Καφιριο are here attested for the first time (III, ll. 1, 3; perhaps analogous to Latin Audius/Audeius and *Cafrius). There are also two names derived from Καιδικις (II, ll. 4 and 6), while Κοσσανω (III, l. 4) could be derived from the toponym Κοσσα.

Image source: Lazzarini 2009: 427, fig. 1. Courtesy of L. Lazzarini. Commentary: this defixio was discovered in 2000 in Cassana near the Fondo Castello necropolis. The tablet, found on the ground, was still rolled up at the time of discovery. The curse was written on a long lead strip and the text is arranged in four columns, which are separated with vertical lines (Murano, apud TDO, p. 193 n. 4 has rightly linked this layout to the conventions of writing on papyrus). The inscription, written in Oscan language (and Greek alphabet I–III, IV, ll. 1–3) and Greek language (IV, ll. 4–5), runs from left to right and has been dated to the fourth or beginning of the third century BCE on palaeographic grounds.

Column IV contains the cursing formula in both Oscan and Greek in ‘uno degli esempi più straordinari di codeswitching e di continuum sociolinguistico che il mondo antico ci metta a disposizione’ (Poccetti 2014: 101). This scholar (2014: 98) has divided IV, l. 2 so that καρισταδ (?) would be a subjunctive verb, while he connects πισπιτ to Latin quicquid, σολλομ to Latin omnium and ηισου to Greek αὐτεῖ. Based on this analysis, Poccetti translates as follows: ‘sia inflitta una pena, una punizione a chiunque tra tutti questi’ or ‘chiunque tra tutti (quanti nominati sopra) sia trascinato o punito qui’. Finally, Poccetti has identified the ιμ (IV, l. 2) as an abbreviation for the particle inim, which either coordinates the Greek and Oscan cursing formula or is an adverb meaning ‘also’.

We do not know the circumstances that led to the writing of this curse, which Poccetti (2009) has argued has a bipartite structure: a list of personal names (I–IV, l. 2; see Lazzarini 2009) and the cursing formula proper (IV, ll. 3–5). The defixio targets a group of people, many of whom belong to the gentes Statia and Caedicia and are listed in the first two columns. The three individuals found in the third column are named with a different type of onomastic formula that includes an ethnonym. The use of this naming device along with their position at the end of the list has led Poccetti (2014: 90) to argue that these final victims were foreigners. According to his analysis, the text presents the following orthographic and phonetic features: confusion between ο and ω, which is common in the Doric dialect (in the Greek portion of the text: χθώνιε for χθόνιε [IV, l. 4] and κοσσανω for κοσσανο [III, l. 3]; in the Oscan part: σολλομ for *sollōm [IV, l. 3]), the confusion of η for ε (in the Oscan portion: ηισου, IV, l. 3), misuse of aspirates (in the Greek part: δεκεω for δεχου and κάθεκε for κάθεχε [IV, ll. 4–5 respectively]; in the Oscan part the use of φ for f: Αλαφιω and Καφιριω [III, ll. 2-3, respectively]) and the use of ϝ and υ for ω (in the Greek part: Εμαυτω [ΙΙ, l. 3]; in the Oscan: Πακϝιω, Νοϝιο, Αϝες, Αϝελιος, Γναυ(ς) and Αυδαις [I, l. 2; III, ll. 2 and 1, respectively; IV, l. 1; II, l. 1 and III, l. 1 respectively]).

Lazzarini, for her part, has focused on the Greek formula δέκεο hερμᾶ χθόνιε ταῦτα καὶ κάθεκε αὐτεῖ, pointing out the Doric forms already signalled (see above) and offering the following translation for this portion of the curse: ‘accogli, o Hermes sotterraneo, tutto ciò e tienilo qui sotto’. Murano has stressed that these phrases complement each other, but one is not a translation of the other (hence it is not bilingual sensu stricto): ‘la commistione linguistica presente nel testo, infatti, mette in luce, ancora più profondamente, la situazione di diglossia presente in area brettia’ (apud TDO, p. 195).

The lists of personal names are found both in the nominative and genitive cases. Furthermore, the

After being inscribed, the tablet was rolled up.

166

Italia 82. Copia Thurii, Sybaris Provenance: necropolis. Current Location: MNA Sibaritide. Date: 1st century CE. Material: lead. Measurements: (6.5) × (7) cm. Reading: 5



Quintia(m) Prima(m) [---] rogo inimica(m) m[ea(m) ---] ma(n)do · suas ṃạ[nus ---] delocata iri tam[quam ---] monimenta et Rusticena(m) m[a(n)do (?) ---] ut in Lamia(m) [---] ṿẹniạt rogo [---] ++

Bibliography: Costabile 2017.

that led to the writing of this curse. What we do know is that it targets two women: Quintia Prima and Rusticina. The defigens asks that the hands (suas ṃạ[nus],...; l. 5 ) of the first, who is qualified with the adjective inimica (l. 2), be dislocated (delocata iri, l. 4). This is a punishment without parallels in the corpus of defixiones from the Roman West. If we accept the editor’s reconstruction, the phrase tam[quam] monimenta (l. 4–5), would be employed to establish a persuasive analogy between the first victim of the curse tablet and the mausoleum in which the artefact was deposited, although the text is too fragmentary to fully reconstruct this. Finally, with the phrase ut in Lamia(m) ṿẹniạt rogo (ll. 7–8), the defigens wants Rusticina (l. 6), who is clearly a slave, to be devoured by the awful monster Lamia (so far, the only other known parallel for the invocation of Lamia is in the Greek text DT 74).

Image source: Costabile 2017: 46, fig. 5a. Courtesy of F. Costabile. Commentary: this defixio was discovered in 2001 during the excavations carried out near the north gate of the necropolis in Sybaris. The tablet was deposited inside of the outer wall of mausoleum US 68, which has been dated to the first century CE. The curse was written on a roughly rectangular lead sheet, which has lost its right and bottom edges. The inscription contains eight lines, which run from left to right and were written in capitals dated to the first century CE. The text presents the following orthographic and phonetic features: Quintia for Quintiam (l. 1), Prima for Primam (l. 1), inimica for inimicam (l. 2), mado for mando (l. 3), delocata for dislocata (l. 4), Rusticena for Rusticina (l. 6) and Lamia for Lamiam (l. 7). The fragmentary nature of the text precludes us from determining the circumstances

Translation: ‘I ask and consign Quintia Prima, my opponent, (so that) her hands may be dislocated (?) just as these tombs (...) and I consign (?) Rusticena, (and) I ask that she comes against Lamia.’

Regio IV [Samnium] 83. Marsi Marruvium, San Benedetto Provenance: aquatic context (?). Current Location: SMBPKA. Date: end 1st century BCE–beginning 1st century CE. [Inv. No.: 7235] Material: lead. Measurements: 6.7 × 10 × 0.2 cm. Reading: Sextus · Pompei{i}us Leonida(s) · L(ucius) · Paqued `ius´ Philo · Septimius Felix Marcius Cela5 dus Heren(n)ius Labi canus Marinus C̣(aius) G̣avius Secuṇ[d]ụ[s] 167

Sylloge of  Defixiones from the Roman West Bibliography: CIL IX, 3748; Hübner 1885: no. 946; Wünsch 1897: 27; DT 132; Jeanneret 1917: no. 132; Letta and D’Amato 1976: no. 76; Kropp 2008: no. 1.3.1/1; Buonocore 2012: 349–52; AE 2012, 425; Urbanová 2018: no. 8; Gordon 2019: 424.

Whatever the case may be, the curse was written on a rectangular lead sheet, which has lost its lower-right corner. The inscription contains seven lines of capitals (measuring between 0.7 and 0.8 cm in height), which run from left to right. The words are separated with interpunctuation consisting of a single dot found halfway up the letters. Letta and D’Amato have stressed that the only letter with an archaic form is the a, which has an oblique stroke. Buonocore has dated the text on palaeographic grounds to the end of the Republic or the beginning of the Principate, a time at which the author could have sought to ‘archaize’ the text orthographically.

Image source: Hübner 1885: no. 946. See also App. IV.3, SD 83. Commentary: this defixio was discovered at an unknown time and under unknown circumstances. It belonged to the collection of Julius Friedländer, who was a numismatist and the director of the Königliche Münzsammlung. Following the view of Mommsen (apud CIL), Audollent (apud DT 132) claimed that the curse was found ‘apud Marsos Marruvii (San Benedetto) ad lacum Fucinum.’ Buonocore (2012: 351) has questioned this long-held theory about the defixio’s provenance and has noted that on side B of the tablet there is the legend ‘Rajano bei Corfinium’, which suggests that this curse could have been found (or at least acquired) in Raiano (in the territory of Corfinium). From an onomastic perspective, the personal names found in the tablet are common in the area of San Benedetto dei Marsi (ancient Marsi Marruvium), but Buonocore (2012: 351) has added that the defixio could come from a neighbouring area (the valle Peligna), where some of the nomina from the curse are also attested.

We do not know the circumstances that led to the writing of this curse, which targets seven individuals, all of whom are liberti except for the sixth (Marinus), who appears to be a slave. The text presents the following orthographic and phonetic features: Pompeiius for Pompeius (l. 1), Leonida for Leonidas (l. 2) and Herenius for Herennius (l. 5). According to Letta and D’Amato, the nomen Paquedius (l. 2) has a Marsian origin and is derived from Paquius, which is a ‘variante grafica di Pacuvius’, a name attested in the area. The nomina Marcius, Herennius and Gavius (ll. 4, 5 and 7 respectively) are also documented in this area. The cognomina Leonidas (l. 2), Philo (l. 3) and Celadus (ll. 4–5; see Solin 2003: 1211) are Greek. Labicanus (ll. 5–6) is derived from the toponym Labici (cf. Kajanto 1982²: 182).

84. Aquilonia, Monte Vairano Provenance: domestic area. Current Location: SBAM. Date: 3rd century BCE. Material: lead. Measurements: 1.5 × 6.7 × 0.1 cm. Reading: A pakis : heleviis · tre B statis : betitis : [---]

Bibliography: De Benedittis 1980; ImIt: Bovianvm 98; TDO 6.

An opisthograph, this two-line inscription dates to the third century BCE and its letters measure between 0.6 and 1.2 cm in height. The words are separated with interpunctuation consisting of two dots (except for the second interpunct, which is made of only one dot). The text, written in the Oscan language and alphabet, contains two onomastic formulae in the nominative, consisting of a praenomen, nomen and patronymic. According to the editor, the patronymic tre (A) is for Trebio, an element that is documented as both a praenomen and nomen. Benedittis has pointed out that Betitis is a nomen already attested in the area.

Image source: De Benedittis 1980: 420. Courtesy of G. De Benedittis. Commentary: this defixio was discovered in 1979 during the excavations of the Samnite settlement of Aquilonia (today Monte Vairano). It was written on a lead strip with rounded edges and today is broken into five fragments that fit together. The tablet is complete, though damaged on side B and also along a fracture running across part of its surface. 168

Italia Regio V [Picenum] 85. Septempeda, San Severino Provenance: unknown. Current Location: private collection. Date: 1st century CE. Material: lead. Measurements: see commentary. Reading: 5



Antestia Sabina eṭ Vibia Politice Clymene Cambosa PIAM A Felicissima Oppia Silvina dicato

Bibliography: Lanzi 1789: 548–49 and 569; CIL IX, 5575; Wünsch 1897: 27; DT 131; Jeanneret 1917: no. 131; Kropp 2008: no. 1.2.1/1; Urbanová 2018: no. 7.

The curse is written in an oblong lead sheet, whose height, according to Lanzi (1789: 569), is ‘circa a un palmo’. The inscription contains nine lines of capitals, which run from right to left and were written ‘con lettere di rilievo come in sigilli romani’ (Lanzi 1789: 656). We do not know the circumstances that led to the writing of this curse, which targets a group of women (l. 9; dicato for dicat˹ae˺ (sunt), ‘are consecrated’?). Politice and Clymene (ll. 3–4) are Greek cognomina (see Solin 2003: 1262 and 600, respectively). Cambosa (ll. 4–5) has no parallels, while the other names are all well documented. Concerning PIAMA (ll. 5–6), Audollent has claimed ‘mihi obscurus est locus et probabiliter corruptus’ (apud DT 131). Given the position of these letters in the text, it is reasonable to assume that they are part of another cognomen, which is currently beyond reconstruction.

Image source: Lanzi 1789, Tav. XVI, fig. 10. Commentary: this defixio was discovered in 1782 in San Severino together with ‘un’olla di terra cotta, nel cui fondo era una posatura di liquor nero, come vinoso’ (Lanzi 1789: 569). Although we do not know the precise archaeological context of the vessel, it is likely that it was linked to some sort of ritual connected to the manufacture of the curse. This type of deposit, which involves the use of the curse tablet as a sort of lid for a ceramic vessel that held organic remains, has at least one parallel from Mautern (see 526) and perhaps a second from the fountain of Anna Perenna at Rome (see Polakova and Rapinesi 2002: 39 and 49, Inv. 475545).

86. Urbs Salvia, Urbisaglia Provenance: on the ground. Current Location: unknown. Date: 1st century CE. Material: lead. Measurements: 2.6 × 3.6 × 0.2 cm. Reading: A Arium rium A B Priamuṃ Demetriụ(m) Priamum 169

Sylloge of  Defixiones from the Roman West Bibliography: Marengo 1989: 55–59; AE 1992, 533.

An opisthograph, the inscription contains six lines, which run from left to right. Two different hands inscribed the tablet as can be seen in the differences between the a and r on each side of the lead sheet. The letters range from 0.3 to 0.6 cm in height. The text is quite simple and consists of the names of the three victims in the accusative case. Of these, Arius and Priamus are repeated (on side A, the second Arium is missing the initial letter, which could have been intentional; for a parallel, see 19). All of the cognomina are Greek, though Arius is rather rare and probably derived from Ἄρειος (see Marengo 1989: n. 23 and Solin 2003: 383). The other names, Demetrius and Priamus, are much more common. All these names suggest that the victims were all slaves.

Image source: Marengo 1989: 56, figs. 10, b and d. Courtesy of S.M. Marengo. Commentary: this defixio was discovered on the ground in Ripe San Ginesio (ancient Urbs Salvia and very close to modern Urbisaglia). The curse was written on a small rectangular lead sheet with an uneven surface. The curve along the left edge of side A and the presence of a small hole found at the top-left corner suggest that the defigens attempted to fold and pierced the tablet.

Regio VII [Etruria] 87. Volaterrae, Volterra Provenance: necropolis. Current Location: unknown. Date: unknown. Material: lead. Measurements: unknown. Reading:

ṿ[------]tanzu(n)i [------]ia vel[uś m]ẹtẹlial · [---]tṇạl [-?-]aχ+[-c.1-]n[-?-]

Bibliography: Gori 1737: 404; CIE 43; Pallottino 1968: no. 400; Rix 1991: Vt 4.4; Massarelli 2014: 196; Massarelli 2016: 518–20; Massarelli 2019: 368–69.

this item claims that it was broken in two, the drawing depicts a roughly rectangular tablet that is intact. To make matters worse, Gori transcribed two lines, which run from left to right, but the picture clearly shows three lines. Massarelli (2016: 519) has offered a reading based on the drawing, which has been followed here. We do not know the circumstances that led to the writing of this curse, which contains three feminine nomina ([---]tanzu(n)i, [m]ẹtẹlial and [---]tṇạl, the final two in the genitive) and a masculine praenomen (vel[uś], l. 2).

Commentary: this defixio was discovered in 1732 in the Etruscan necropolis ‘i Marmini’ (see A. F. Gori, apud Massarelli 2016: 518) ‘super calvarium defuncti hominis’, that is, on top of the skull of a deceased man (Fabretti, apud Massarelli 2014: 196, n. 3). Although the curse’s current whereabouts are unknown, there is a surviving sketch, which raises a series of questions. Though testimony about

88. Volaterrae, Volterra Provenance: necropolis. Current Location: M. Guarnacci. Date: 2nd–1st century BCE. Material: lead. Measurements: 11.8 × c. 33 cm. Reading: 5



10



Col. I v · supni · aṣtnei v · supni · larθi · puineil*l* v · supni · velanial v · supni · ceicnal l · velusna · felmuial v · velusna · v *calati(al)* v · puina · armnial l · larθru · fulnei p ultace · ceicna(l) · icap · linei l · larθru · θepza cure · malave · * p utace* *ḷ · ḷạristna · v · cụreacap · fuluna* *l uṭac ̣e*

170

Col. II *acap · fuluna · mazutiu l · larθ(r)u · **canis** la · armne · alpuz · fuluna l · larθ(r)u a · velan(e) · l · puine a · velusna fastịạ · i larθ(r)u(i) larθi · φlavi urin(a)te(ś) l · armne mas · ve · ceicnei φ(l)ave setra φeḷ(mui) θuścẹ · fẹḷmv · larθ u{-c.1-}tacẹ*

Italia

Bibliography: CII 314; CIE 52a; DT 124; Jeanneret 1917: no. 124; Vetter 1960b: 178–86; Pallottino 1968: no. 401; Rix 1991: Vt 4.1; Massarelli 2014: 183–95; Massarelli 2016: 520–21; Massarelli 2019: 369–70.

The palaeographic characteristics of this text reveal that the authors of the text made use of the type II alphabet from Maggiani’s typology. This type was first popularized in southern Etruria at the end of the fifth century BCE and spread to northern Etruria during the third century BCE. The type II alphabet was in use until the first century BCE. Accordingly, and following Massarelli (2014: 185 and 191, n. 5), it is possible to date this text together with 89 and 90 approximately to a date in either the second or first century BCE, due to palaeographic and prosopographical considerations.

Image source: Massarelli 2014: 183, fig. 8. Courtesy of R. Massarelli. Commentary: this defixio was discovered together with 89 and 90 inside of a hypogeum during the excavations of the necropolis of San Girolamo, which were carried out in December 1755. These two curses were found in the lower chamber among broken urns and ashes (for the archaeological context, see Lami 1756, apud CIE 52a).

We do not know the circumstances that led to the writing of this curse, which contains the names of the victims and the expression cure malave (I, l. 11). This phrase has been taken as a cursing formula with malave being connect to the root mal-, which is associated with the idea of ‘watching’ or ‘observing’ (for a linguistic discussion, see Massarelli 2014: 190 and 189).

This curse was written on a roughly rectangular lead sheet. The inscription contains 13 lines, which were written in Etruscan language and alphabet, and arranged in two columns that run from right to left (as is the normal practice in this language). The layout of the first column is undoubtedly neater than that of the second. That said, both make use of interpunctuation consisting of a single dot found halfway up the letters. The differences between the columns suggest that at least two authors wrote this tablet. The first, who was much more familiar with the conventions of writing, inscribed the first eleven lines of column I; the second, whose contributions are marked in the text provided above with asterisks, added several words in column I (at the ends of ll. 2, 6 and 11) and wrote I, ll. 12–13 as well as most of column II (everything but the term canis, in II, l. 2). As Massarelli has noted, (whose reading I have followed), it is notable that the second author, who was less versed in the mechanics of writing, simplified ‘sistematicamente i nessi consonantici con liquida, omettendo di scriverla’ (2014: 190). Furthermore, Vetter (1960b: 180) was correct to point out that the second hand used different abbreviations for the masculine and feminine praenomina Ve(lia) and V(el). In column II, a third hand has perhaps been detected, whose contributions are marked with two asterisks, in the word canis (II, l. 2). While Vetter (1960b: 185) proposed that this word was a Latin term transliterated into Etruscan writing, Massarelli is right to question this claim, since there are not enough elements to defend such a hypothesis (for a full discussion, see Massarelli 2014: 191).

The elements constituting the curse’s onomastic formulae vary. They can be qualified with a matronymic (I, ll. 3–5, 7 and 9), the names of both parents (I, l. 6), a wife’s name (I, ll. 1, 2 and 8) or a cognomen (I, l. 10). The first portion of the text names 13 individuals. When the contributions of the second author are taken into consideration (and there is no reason to doubt that this is a continuation of the previous text), another six or eight people are named. In col. II, several nomina from col. I are repeated, such as Armni (II, l. 10 and I, l. 7), Velani (II, l. 6, which agrees with the matronymic in col. I, l. 3), Puinei (II, l. 6 and I, l. 7), Velusna (II, l. 7 and I, ll. 5–6), Larθru (II, ll. 2, 5 and col. I, l. 8) and Felmui (II, ll. 12–13 and I, l. 5). In the main list of names, the word θuścẹ (II, l. 13) is found and has been identified as a verb, which Massarelli (2014: 192–93) has translated as ‘they have lied/committed perjury’. If this is correct, this curse would be a juridical defixio. It has proven difficult to explain the additional names found after the verb θuścẹ in II, l. 13, one of which is repeated from I, l. 13. Massarelli has suggested that these final names could have been a last-minute addition. After being inscribed, the tablet was folded in half. According to Lanzi (apud CIE), 89 and 90 ‘erano incluse 171

Sylloge of  Defixiones from the Roman West nella prima che si ripiega a maniera di dittico; e questa era stretta con una striscia di piombo’ (for this arrangement resembling a diptych and the use of a lead strip to close it, see 52, 61 and 71). The lead strip, which is currently

lost, also contained an inscription, of which A.F. Gori published a sketch (see Massarelli 2016, fig. 9 and p. 520, n. 19 for this small text, which is undoubtedly related to the defixio).

89. Volaterrae, Volterra Provenance: necropolis. Current Location: M. Etrusco Guarnacci. Date: 2nd–1st century BCE. Material: lead. Measurements: 8.9 × 11 cm. Reading:

θuśaθur selasva θlu θupit(a) aisece tati

Bibliography: CII 315; CIE 52b; DT 125; Jeanneret 1917: no. 125; Pallottino 1968: no. 402; Rix 1991: Vt 4.2; Massarelli 2014: 186, 195–96; Massarelli 2016: 521–23; Massarelli 2019: 369–70.

contains four lines of text running from right to left (as is the normal practice in this language). We do not know the circumstances that led to the writing of this curse, whose interpretation has been made all the more difficult due to the lack of interpunctuation. Massarelli (2014: 193–94) has argued that θuśθaur (l. 1) could be a verb related to the (possible) verb θuścẹ found in 88 (col. II, l. 13; see the commentary for 88). If this is the case, he has speculated that here the subject of the verb would be the noun selasva (l. 2). Most scholars have taken θlu θupit(a) (l. 3) as an onomastic formula consisting of Θlu (perhaps an abbreviated praenomen?) and Θupit(a) (nomen). Finally, l. 4 may contain the verbal form aisece (whose root is unknown), which only complicates the explanation of the final tati.

Image source: Massarelli 2014: 186, fig. 9a. Courtesy of R. Massarelli. Commentary: this defixio was discovered in 1755 together with 88 and 90 in the lower chamber of a hypogeum. The two smaller tablets (i.e., the present curse and 90) were found inside of the larger curse, which had been folded like a diptych (for the archaeological context, see the commentary for 88). The curse was written in the Etruscan language and alphabet on a roughly rectangular lead sheet, which

90. Volaterrae, Volterra Provenance: necropolis. Current Location: M. Etrusco Guarnacci. Date: 2nd–1st century BCE. Material: lead. Measurements: 9.2 × 13 cm. Reading: 5

[---]herace[---] [---]ṣvu[---] [---]a[-c.1-]ḷ[---] [---]urẹ[---] [---]ein[---]

172

Italia Bibliography: CII 316; CIE 52c; DT 126; Jeanneret 1917: no. 182; Pallottino 1968: no. 403; Rix 1991: Vt 4.3; Massarelli 2014: 186 and 196; Massarelli 2016: 521–22; Massarelli 2019: 369–70.

found inside of the larger curse, which had been folded like a diptych (for the archaeological context, see the commentary for 88). The curse was written on an irregularly shaped lead sheet in Etruscan and it contains five lines, which run from right to left (as is the normal practice in this language). The fragmentary nature of the inscription precludes any satisfactory reading. In fact, it is only possible to make out the word Herace, which is a nomen already attested in Volterra.

Image source: Massarelli 2014: 186, fig. 9b. Courtesy of R. Massarelli. Commentary: this defixio was discovered in 1755 together with 88 and 89 in the lower chamber of a hypogeum. The two smaller tablets (i.e., the present curse and 89) were

91. Volaterrae, Volterra Provenance: unknown. Current Location: M. Etrusco Guarnacci (?). Date: unknown. Material: lead. Measurements: 4.7 × 8.2 Ø cm. Reading:

l · veluṣna · raufial l · ap · cuinui mazuṭi · lautni · cnev(nas) lθ · alpiu · ianzu

Bibliography: CIE 4613; Rix 1991: Vt 4.6; Massarelli 2014: 195; Massarelli 2016: 523 Massarelli 2019: 370.

are separated with interpunctuation found halfway up the letters.

Commentary: this defixio, discovered at an unknown date and in unknown circumstances, was inscribed on a semicircular lead sheet. The inscription contains four lines, which, unlike the normal practice of Etruscan, runs from left to right. This would nevertheless fit in well with the logic of ancient magical practices and should be explained as a persuasive analogy (just as a text is deformed, so too shall be that which the text describes; on this, see section I.4.2.1.1). The words

We do not know the circumstances that led to the writing of this curse, which contains a series of personal names, the first of which is qualified with a matronymic (raufi, l. 1). The nomen Mazuṭi (l. 3), for which Vetter has claimed Celtic in origin, is also found in 88, whereas there are no known parallels for Alpiu (l. 4). Given that this text does not conform to the characteristics of other Etruscan defixiones, Massarelli has called its authenticity into question. Today, the whereabouts of the defixio are unknown.

92. Arretium, Arezzo Provenance: sanctuary. Current Location: M.A. Arezzo. Date: 2nd century CE. Material: lead. Measurements: 11.8 × (8.9) cm. Reading: A Q(uintum) · Letinium Lupum · qui · et vocatur Gau cadiọ qui 5 est fị[lius] Ṣạḷ usti[es Vene-] ries sive Ven[e-] rioses hunc ego aput vos10 trum

B numen de mando devo veo desacri fico uti vos A qu`a´e ferventes siv[e] ṿọs Nimfas [si]ve quo alio no mine voltis adpe [l]lari uti vos eum interemates interficiates intra ann um istum 173

Sylloge of  Defixiones from the Roman West Bibliography: Gamurrini 1869: 50–52; Mommsen 1870; CIL XI, 1823; Wünsch 1897: 27; DT 129; ILS 8748; Jeanneret 1917: no. 129; Rubio and Bejarano 1955: no. 194; Calabi Limentani 1973: no. 96; Marengo 2006; AE 2007, 150; Kropp 2008: no. 1.1.1/1; De Bernardo Stempel and Hainzmann 2009: 78; AE 2009, 70; Urbanová 2018: no. 1; Gordon 2019b: 428–29.

of the text, who was also familiar with the conventions of cursing. For instance, the defigens was well aware of the importance of using a name to identify the victim precisely; accordingly, he named his victim in full (Quintum Letinium Lupum, A, ll. 1–2) as well as used the agnomen (i.e., the ‘nickname’: Gaucadiọ, A, ll. 3–4) and the matronymic (Veneries sive Venerioses, A, ll. 1–8; for a parallel, see 341). The deities invoked in the curse are referred to as Aqu`a´e ferventes, siv[e] ṿọs Nimfas, [si]ve quo alio nomine voltis adpe[l]lari (B, ll. 4–9): this formulation recalls Roman prayers in which the deity is referred to in broad and inclusive terms in order to avoid provoking his/her anger (on this, see Gordon 2019b: 429; for the use of sive, see de Bernardo Stempel and Hainzmann 2009: 78). Following Gordon, it is worth noting the use of aput (A, l. 9; ‘in the presence of’), which fits well with the appeal to the Nymphs, referred to by their numen, a term which ‘combines the sense of divine majesty (...) with that of a supernatural presence’ (2019b: 429). These goddesses are asked to bring the victim’s life to an end intra annum istum, a deadline that is also found in 494 (for magical deadlines in general, see Sánchez Natalías 2019a).

Commentary: this defixio was discovered during a land survey in July 1869 in a spring ‘di acqua acidula minerale’ (Gamurrini 1869: 47), located in the town Bagnoli di Pèrgine Valdarno (nearby Arezzo). A group of corroded coins, which mostly date to the reign of Antoninus Pius, were found in the same spring. The curse was written on an oblong lead sheet, which has been broken along its right edge. An opisthograph, the inscription contains 23 lines, which run from left to right and were written in capitals measuring between 0.4 and 0.85 cm in height. Some words are separated with interpuncts consisting of a single dot. The text presents the following orthographic, phonetic and morphological features: the use of the Greek genitive –es for the Latin –ae in Salusties, Veneries and Venerioses (A, ll. 5–8; for a hybrid genitive in –aes, see 11); vostrum for vestrum (A, l. 10); Nimfas for Nymphas, voltis for vultis (B, l. 8), adpelari for appellari (B, l. 9), interemates for intirimates (B, l. 10) and interficiates for interficiatis (B, l. 11). The text has been dated to the second century based on a palaeographic and onomastic analysis (the agnomen Veneriosa with suffix –osa is common in Imperial period).

After being inscribed, it seems that the defigens made up to three attempts at piercing the tablet from side B, as is suggested by the markings found at l. 3 and between ll. 7–8. Translation (Gordon 2019b: 428, slightly modified): ‘Quintus Letinius Lupus, also called Caucadio, the son of Sallustia Veneria or Veneriosa ―before your active divine power I transfer, curse, dedicate him, that you, the hot waters, or you nymphs or by whatever name you wish to be appealed to, should destroy, kill (him) within the year.’

The wide array of rhetorical figures (alliteration, paronomasia, repetition, anaphora, etc.) found in the text demonstrates the linguistic sophistication of the author

93. Perusia, Perugia Provenance: unknown. Current Location: unknown. Date: unknown. Material: lead. Measurements: 1.3 × 1.5 cm. Reading: A B Felix · Pị- Camụ etron · rinuṣ [------] AV Bibliography: CIL XI, 6722, 2; Wünsch 1897: 28; Wünsch 1900: 270; DT 130; Jeanneret 1917: no. 130; Besnier 1920: 22; Preisendanz 1933: 156; Kropp 2008: no. 1.1.4/1.

the curse was discovered. It was written on a rectangular lead sheet, which has a hole in its bottom-left corner. An opisthograph, the inscription contains six lines, which run from left to right. On side A, the words are separated with interpunctuation consisting of a single dot. We do not know the circumstances that led to the writing of this curse, which contains a list of three personal names. Of these names, Camurinus is notable, since it is only attested in Haidra (Africa Proconsularis) besides in the present curse.

Commentary: this defixio, today lost, formed part of the personal collection of G.B. Rossi-Scotti, who was the director of Museo Archeologico Nazionale dell’Umbria (Perugia) at the end of the nineteenth century. We do not know anything about the archaeological context in which

174

Italia 94. Aquae Populoniae, Campiglia Marittima Provenance: necropolis. Current Location: MAN Florence. Date: 3rd century BCE. Material: lead. Measurements: c. 11 × 15 cm. Reading: 5



10

sθ · velśu · lθ · c · lθ · ve[lśu ·] inpa · θ̣apicu`n´ θapintaś · aθ · velśu `ḷθ̣ c ̣´ lθ · velśu lθ · c · ls · velśu `lθ c´ ḷθ · śuplu aθ · śuplụ · lṣ · hasmunị sθ · cleustẹ · ạθ̣ · c ̣ḷeuste · vl · runs(-c.1-)`au´ θancvil · velśui · cẹś ̣ · zeriś · imś · se mutin · aprenśaiś ̣ · inpa · θ̣apicun θapintaiś · ceuśn · inpa · θ̣apicun · i luu · θapicun · ceś · zeriś titi · setria · lautnita

Bibliography: Gamurrini 1891; Gamurrini 1895; CIE 5211; Wünsch 1897: 24; DT 128; Jeanneret 1917: no. 128; Pallottino 1968: no. 380; Rix 1991: Po 4.4; Massarelli 2014: 196–213; Massarelli 2016: 523–26; Massarelli 2019: 371–73.

the central portion that contains the text. The inscription contains 10 lines, which run from right to left, as we would expect in an Etruscan inscription. The words are separated with interpunctuation consisting of a single dot placed halfway up the letter. Based on palaeographic considerations, the text dates to the end of the third century BCE.

Image source: Massarelli 2014: 197, fig. 10. Courtesy of R. Massarelli.

The circumstances that led to the writing of this curse are unknown. Here, 11 men and 1 or 2 women belonging mostly to the families Velśu, Śuplu and Cleuste are cursed. After the first two members of the list are mentioned, both of whom belonged to the Velśu family, we find the phrase inpa θ̣apicun θapintaś (ll. 1–2), which is repeated (ll. 7–8) after the enumeration of the remaining ten victims (ll. 2–7). In this phrase, the word θapicun (ll. 1 and 7–9) has been identified as a verb analogous to the Latin devoveo (Torp, apud Massarelli 2014: 205). This interpretation has been widely accepted. The text concludes with the mention of the liberta (lautnita) Titi Setria (l. 10), who has alternatively been taken as a final victim or as the defigens. Although the former solution is more likely, the difficulties in understanding this text preclude any definitive solution.

Commentary: this defixio was discovered in 1890 in a necropolis located at Monte Pitti (near Campiglia Marittima and to the site of ancient Populonia). The curse was found inside a grave, though Massarelli, whose reading is reproduced here, has pointed out that we do not know whether it comes from one of the older graves (i.e., fourth or fifth century BCE) or one of the more recent ones, which are contemporary with the curse tablet (third century BCE). Whatever the case may be, the curse was written on an irregularly shaped lead sheet, which is intact except for a breakage along the lower edge (where there is no text). The tablet shows signs of corrosion along its edges and in 175

Sylloge of  Defixiones from the Roman West 95. Caere, Cerveteri Provenance: necropolis. Current Location: MNA Cerite (?). Date: 2nd–1st centuries BCE. Material: lead. Measurements: (24) × 14 cm. Reading: 5



10



15



[-] Mam(i)lius · M(arcus) · f(ilius) C(aius) · Mamilius · Limetanus C(aius) · Mamilius · Atelus C(aius) · Macius · Copo L(ucius) · Laterius · Balbus L(ucius) · Laterius · Corvinus L(ucius) · Laterius · Cor(v)us Q(uintus) Laterius Rabilinus [-c.1- La]terius · Hispanus P(ublius) · Laterius · Luscus M(arcus) · [Late]rius Q(uintus) [· Lat]erius · Paperianus [- L]aterius Balbus [-] Minucius C(aius) f(ilius) M(arcus) · Laterius · Mulus Hileria Midia Patolcia Pineia

Bibliography: Mengarelli 1931: 422–23; Mengarelli 1937: 400–01; AE 1939: 47; CIL I², 2765; Ricci 1956: col. 819– 20; Heurgon 1960; AE 1961, 207; García Ruiz 1967: no. 64; Solin 1968: no. 32; ILLRP, 1148; Kropp 2008: no. 1.1.3/1; Urbanová 2018: no. 6.

Hileria for Hilaria, Midia for Media or Maedia (both in l. 16), and Patolcia for Patulcia (l. 17). Some of the onomastic formulae are composed of Etruscan nomina and Latin cognomina as can be seen in Caius Mamilius Limetanus (l. 2), who Heurgon has identified as one of the Tribunes of the Plebs from 109 BCE. This Tribune of the Plebs put forth a legislative agenda in which he both denounced anyone who supported Jugurtha and also passed an agrarian law regulating boundaries of public land. Due to the latter measure, he received the cognomen Limetanus. Heurgon (1960: 226) has identified the next individual, C(aius) Mamilius Atelius (l. 3), as the cousin of the Tribune. Most of the remaining victims belonged to the gens Lateria. If the identification of the Tribune is correct, this curse belongs to a small group of texts that target ‘famous’ victims (for some other examples, see 118, 134–36, 149 and 478). When it comes to the rest of the people mentioned, Hileria (l. 16) was likely a liberta, whereas Midia, Patolcia and Pineia (ll. 16–18) were ingenuae (Heurgon has conjectured that they were the wives of some of the men cursed in the tablet, but this seems highly speculative).

Image source: Mengarelli 1937: 400. Commentary: this defixio was discovered in c. 1937 during the excavations of the necropolis of la Banditaccia. The curse was found in the dromos of tomb no. 336, which dates to the fifth century BCE based on its grave goods. More precisely, it was ‘nell’andito a destra della porta, a circa m. 1,00 sopra il pianerottolo’ (Ricci 1956: col. 819). The curse was written on a rectangular lead sheet, which has lost its four corners. There is also some breakage along the left edge at l. 10. The inscription contains 18 lines, which run from left to right and were written in capitals. Based on its palaeographic characteristics, the text should be dated to the second-first centuries BCE. The words are separated with interpunctuation consisting of a single dot found halfway up the letters. The circumstances that led to the writing of this curse are unknown, since it only contains a list of personal names in the nominative. The text presents the following orthographic and phonetic features: Atelus for Atellus (l. 3), Macius for Magius (l. 4), Mamlius for Mamilius (l. 1), Corus for Corvus (l. 7)

After being inscribed, the tablet was certainly pierced at three of its corners and probably fixed to the wall ‘nei quali erano stati evidentemente inseriti i chiodi con cui essa era stata fissata alla parete’, according to Mengarelli (1937: 400; for a parallel see 129). 176

Italia 96. Fontanaccia Provenance: necropolis. Current Location: unknown. Date: 2nd–3rd centuries CE. Material: lead. Measurements: 5.3 × 7.7 cm. Reading: A qomodo h(a)ec châr ta cọẹlis abeat{i} in dẹọ Adoṇịne cito iacêt silet langet sica5 ta sic Qụịntus Agrip-

B pini s(ervus) ụṭẹr ṣạḷṭụẹṇṣịṣ languiat aigrotêt ex omologi(s) fẹ̣ ri(s) igni(s) n̂â(tus) an̂n(is) IL devincit 5 non seion fortiôr et sic moriatur

Bibliography: Stanco 2003; AE 2003, 645; Bevilacqua 2010a: 1955, note 39.

on the right side and the top-left corner. An opisthograph, the inscription contains 11 lines, which run from left to right and were written in capitals (though the d and s have the shape of new Roman cursive), measuring between 0.2 and 1 cm in height. Based on palaeographic analysis, the text dates to either the end of the second century CE or the third century CE. Note the pracititioner’s use of nexus (âr, A, l. 1; êt, A, l. 3 and B, l. 2; n̂n as well as ôr, B, ll. 4 and 5) and three abbreviations (s(ervus), na(tus) and ann(is), B, ll. 1 and 4).

Image source: Stanco 2003: 132–33, figs. 11 and 13. Courtesy of E. A. Stanco. Commentary: this defixio was discovered in 1988 during the excavations at Fontanaccia (Allumiere, Rome) in tomb no. 1, which originally was closed with four tegulae. Nevertheless, these tegulae had already been removed at the time of discovery, which suggests that the tomb had been plundered in Antiquity. The individual buried was a child (see section I.6.2.2 and n. 274), whose grave goods date to the second half of the second century CE. Inside the grave, a lamp was found, while outside of the grave at the deceased’s feet there was an as of Faustina Augusta, an iron nail, a ceramic cup and the remains of another made of African terra sigillata. According to Stanco (2003: 128 and n. 7), despite being outside of the grave, all these objects were part of a ritual offering related to it. Outside of the grave but at the same level as the deceased’s chest, excavators found the present defixio.

The text presents the following orthographic, morphological and phonetic features: qomodo for quomodo (A, l. 1), hec for haec (A, l. 1), abeati for abit (A, l. 2), langet for languet (A, l. 4), sicata for siccata (A, ll. 4–5), languiat for langueat, aigrotet for aegrotet (both in B, l. 2), feri igni for ferus ignis (B, l. 3) and devincit for devincat (B, l. 4). The text is also noteworthy for the use of the Greek words omologi(s) (B, l. 3; from ομολογον), and seion (B, l. 5; from θειος), whose presence may suggest that the author was of Greek origin. We do not know the circumstances that led to the writing of this curse, which is here referred to as a chârta (cf. 484). The text targets Quintus, defined as the slave of Agrippinus and as a saltuensis (a woodland watchman). In the curse, the defigens employs a similia similibus formula to compare the way that the curse tablet was deposited and

The curse was written on a rectangular lead sheet, the surface of which is somewhat irregular, since it was hammered before being inscribed. The tablet is intact and in good condition, though there are some small breakages

177

Sylloge of  Defixiones from the Roman West Translation: ‘Just as this letter quickly vanishes from the upper world (as?) the god Adonis, (and it) lies mute, languishes, having been dried up, in the same way may Quintus, the servant of Agrippinus, the one who is a woodland watchman, languish and become sick; and as agreed (?) may a fierce fire overcome him, at the age of forty-nine (since) he (is) not stronger than the divine powers and so (finally) he dies.’

then deteriorated to the demise of the victim (for a similar analogy, see 157 and 518). Given the victim’s profession (he was a woodland watchman), the mention of Adonis is quite apt, since he died after being attacked by a wild boar. The defigens includes a magical deadline for the languishing and then death of Quintus that is unparalleled in the rest of the corpus: when he is 49 years old (for a discussion of this ‘deadline’, see Sánchez Natalías 2019a).

Regio VIII [Aemilia] 97. Classis, Classe Provenance: necropolis. Current Location: unknown. Date: unknown. Material: lead. Measurements: 3.1 × 12 cm. Reading:

Q(uintus) Tiburius Optatus TR

Bibliography: Susini 1971; Solin 1974: 163; Solin 1998: 60; Solin 2004: 123; Kropp 2008: no. 1.8.1/1; Urbanová 2018: no. 43.

The curse contains the name of the victim followed by the letters TR, which Solin (1998 and 2004) has interpreted as an abbreviation for tr(aditur), tr(adatur) or even tr(aditus est), a formula meant to consign the victim to the deities invoked orally when the text was deposited. Although forms of the verb trado are found in other curses (see 10–14), Solin’s suggestion is speculative, since there are no confirmed parallels for this abbreviation (though see 98). Optatus is a well attested Latin cognomen (see OPEL III, 115).

Commentary: this defixio was discovered in the spring of 1971 at the necropolis of Marabina (Classe) inside of a ceramic cinerary urn. The curse was written on a trapezoidal lead sheet and contains one line of text, which runs from left to right. The inscription was written by a hand in ‘caratteri che partecipano della scrittura comune’ (Susini 1971). The letters reach 0.6 cm in height.

After being inscribed, the tablet was rolled up and crushed.

98. Classis, Classe Provenance: necropolis. Current Location: unknown. Date: unknown. Material: lead. Measurements: 1.3 × 13 cm. Reading:

Venustus TR

Bibliography: Susini 1971; Solin 1974: 163; Solin 1998: 60; Solin 2004: 123; Kropp 2008: no. 1.8.1/2; Urbanová 2018: no. 44.

Solin (1998 and 2004) has interpreted TR as an abbreviation for tr(aditur), tr(adatur) or even tr(aditus est), a formula here used to hand Venustus over the deity invoked orally when the tablet was deposited. Although forms of the verb trado are found in other curses (see 10–14), Solin’s suggestion is speculative, since there are no confirmed parallels for this abbreviation outside of these two curses. Venustus is a well attested Latin cognomen (see OPEL IV, 155).

Commentary: this defixio was discovered in the spring of 1971 during the excavations of the necropolis of Marabina (Classe). The tablet was found inside of a ceramic cinerary urn and contains one line, which runs from left to right. It was written in ‘caratteri che partecipano della scrittura comune’ (Susini 1971) and the letters reach a height of 0.7 cm. The curse is very similar to 97: the name of the victim is followed by TR.

After being inscribed, the tablet was rolled up and crushed.

178

Italia Regio X [Venetia et Histria] 99. Concordia Sagittaria, Concordia Provenance: unknown. Current Location: MN Concordiese. Date: 2nd–3rd centuries CE. [Inv. No.: IG 753] Material: lead. Measurements: 6.1 × 8.6 cm. Reading:

Secundula aut qui sustulet

Bibliography: Bertolini 1880: 420–21; Besnier 1920: no. 59; García Ruiz 1967: no. 82; Solin 1977: 148–49; AE 1977, 305; Buonopane and Croce da Villa 2002a; AE 2005, 512; Kropp 2008: no. 1.7.3/1; Urbanová 2018: no. 210.

other two are quite damaged. The lower-left corner has been pierced. The text contains three lines of capitals (measuring between 0.5 and 1.9 cm in height), which run from left to right. Based on its content, this curse should be grouped among the defixiones in fures. The text has a simple formula, in which the main suspect of the crime, Secundula (for the diminutive, see Kajanto 1982²: 292), is recorded. This name is followed by the formulaic aut qui sustulet (ll. 2-3; note the e for i), with which the defigens seeks to guarantee the punishment of the guilty party, whoever he or she may be (for this all-inclusive formula, see e.g., 450).

Image source: Buonopane and Croce da Villa 2002a. Courtesy of A. Buonopane. Commentary: this defixio was discovered along with the following curse (100) during the excavations carried out by Bertolini in Concordia Sagittaria in 1880. We do not have any precise information about the tablets’ archaeological context.

After being inscribed, the tablet was probably rolled up.

This curse was written on a roughly rectangular lead sheet, whose bottom and left edges are intact, while the

Translation: ‘(I curse) Secundula, or whoever has stolen.’

100. Concordia Sagittaria, Concordia Provenance: unknown. Current Location: MN Concordiese. Date: 2nd–3rd centuries CE. [Inv. No.: IG 754] Material: lead. Measurements: 6.1 × 8.6 cm. Reading:

Ae[l]ia Decimanạ ++ Iulius ++ ESINA +++VADIA(?) deperi(ant?)

Bibliography: Bertolini 1880: 420–21; Solin 1977: 149– 51; AE 1977, 306; Buonopane and Croce da Villa 2002b; AE 2005, 512; Kropp 2008: no. 1.7.3/2; Urbanová 2018: no. 39.

on an irregular lead sheet, which has four holes near some of the corners (for parallels, see 95 and 129). The text was written in capitals (measuring between 0.4 and 0.8 in height) and strongly inclines toward the left, which makes reading it more difficult. We do not know the circumstances that led to the writing of this curse, which contains at least two personal names and ends with a form of the verb depereo (l. 3), which is first attested here in the corpus of curses from the Roman West.

Image source: Buonopane and Croce da Villa 2002b. Courtesy of A. Buonopane. Commentary: this defixio was discovered along with the previous curse during the excavations carried out by Bertolini in Concordia Sagittaria in 1880. The curse’s precise archaeological context is unknown. It was written

Translation: ‘E[l]ia Decimana, (...) Iulius,... may perish (?).’ 179

Sylloge of  Defixiones from the Roman West 101. Pola, Pula Provenance: necropolis. Current Location: MNRTD. Date: second half 2nd century CE. [Inv. No.: Pola 3117] Material: lead. Measurements: unknown. Reading: 5



10



Col. I Caecilius Honoratus Mindius Donatus Minius Charmides Mindius Zoticos Mindius Hermes Mindius Maleus Mindius Narcissus Mindius Eititeus (?) Marcius Soter Decidius Hister Decidia Certa Minervius Epaphroditus

Col. II Lucifer disp(ensator) Lucifer adiutor coloni (?) Vitalis disp(ensator) Trophimus Trophimus alius Anconius qui vilicavit Tertius Amandus Viator

Bibliography: Sticotti 1905; AE 1906, 100a; Besnier 1920: no. 6; Preisendanz 1933: 157; Forlati Tamaro 1947: no. 592a; García Ruiz 1967: no. 28; Kropp 2008: no. 1.7.5/1; Bevilacqua 2012: 60–61; Urbanová 2018: no. 41: Gordon 2019a: 119–20.

from their professions: two dispensatores (II, ll. 1 and 3: two stewards), one adiutor coloni (II, l. 2, the assistant of the owner) and the man qui vilicavit (II, l. 6, i.e., the one who managed the farm). The names that make up the curse seem to have been organized with some care: the first column contains 11 individuals who seem to be liberti, whereas the second column contains slaves, 4 of whose jobs are specified. We must note that eight of the victims here are also found in 102 (ll. 1–5, 7, 10, 12 and 13): Mindius Maleus, Mindius Narcissus (I, ll. 6 amd 7), Decidius Hister, Decidia Certa, Minervius Epaphroditus (I, ll. 10–12), Lucifer, Vitalis and Anconius (II, ll. 1, 3 and 6). Four further victims, Lucifer, Trophimus, Amandus and Viator (II, ll. 2, 4, 8 and 9), may also be found in 102 (ll. 8, 9, 11 and 13, respectively), but they are referred to in different ways (compare, for instance, Lucifer alius of 102, l. 8 with Lucifer adiutor coloni in 101, II, l. 2). The Latin cognomina are Honoratus, Donatus, Hister, Certa, Lucifer, Vitalis, Tertius, Amandus and Viator (see Kajanto 1982²: 279, 298, 196, 254, 288, 274, 292, 360 and 362, respectively). Greek cognomina are Charmides, Zoticos, Hermes, Narcissus, Soter, Epaphroditus, Trophimus and Anconius (see Solin 2003: 1388, 895, 1475, 1183, 451, 343, 1047 and 714, respectively).

Commentary: this defixio was discovered 1904 during the excavation of a tomb located on the Via Flavia. The grave contained a rich array of grave goods consisting of a strigil, bronze mirror, silver cup, five lamps, a ceramic cup, several iron nails, pieces of glass and three coins dating to the reigns of Domitian, Trajan and Hadrian (for a discussion, see Sticotti 1905: 213, n. 1). Archaeologists also found two defixiones, 101 and 102. The tablet currently under consideration was written on a lead sheet that is now broken into six pieces that fit together to form what must have been a rectangular tablet. Portions of the edges are preserved, though the tablet has lost its bottom-right corner and is generally in poor condition. The text is arranged in 2 columns of 12 and 9 lines respectively, which run from left to right and are written in old Roman cursive by a trained hand. The curse tablet has been dated to the second half of the second century CE on palaeographic grounds. We do not know the circumstances that led to the writing of this curse, which targets a group of individuals who worked on a private plantation, judging

After being inscribed, the tablet was rolled up.

102. Pola, Pula Provenance: necropolis. Current Location: MNRTD. Date: second half 2nd century CE. [Inv. No.: Pola 3118] Material: lead. Measurements: unknown. Reading:

180

Italia 5



10



15



[Mind]ius Narcissus Mindius Maleus (?) Decidius Hister Decidia Certa Minervius Epaphroditus Me[nande]r (?) Lu[cifer d]ispensator Lucifer alius Amandus dispensator Vitalis dispensator Trophimus qui dispensavit Anconius qui vilicavit Viator colonus [Sept]imius Sabinianus Flavius Hedistus Annius Calvo Annius Civilis

Bibliography: Sticotti 1905; AE 1906, 100b; Besnier 1920: no. 7; Preisendanz 1933: 157; Forlati Tamaro 1947: no. 592b; García Ruiz 1967: no. 29; Kropp 2008: no. 1.7.5/2; Bevilacqua 2012: 60–61; Urbanová 2018: no. 42.

century CE on palaeographic grounds. We do not know the circumstances that led to the writing of this curse, which consists of a list of 17 personal names. Since some of these names are qualified with the victims’ professions (on this, cf. 101), we can safely say that the curse targets a group of people who worked on a private plantation. The list is made up of five liberti, eight slaves (whose jobs are specified) and then four more liberti. Of these 18, eight also appear in the previous curse: [Mind]ius Narcissus, Mindius Maleus, Decidius Hister, Decidia Certa, Minervius Epaphroditus (ll. 1–5; in 101 col. I, ll. 8, 6 and 10–12, respectively), Lu[cifer], Vitalis, Anconius and Viator (ll. 7, 10, 6 and 9; in 101 col. II ll. 1, 3, 6 and 9). The following Greek cognomina are found in the curse: Narcissus, Epaphroditus, Menander, Trophimus, Anconius and Hedistus (see Solin 2003: 1183, 343, 257, 1047, 714 and 949, respectively). The Latin cognomina are Hister, Lucifer, Amandus, Vitalis, Viator, Sabinianus, Calvo and Civilis (see Kajanto 1982²: 196, 288, 360, 274, 362, 235 and 314, respectively).

Commentary: this defixio, alongside 101, was discovered in 1904 during the excavation of a tomb located on the Via Flavia. The grave contained a rich array of grave goods consisting of a strigil, bronze mirror, silver cup, five lamps, a ceramic cup, several iron nails, pieces of glass and three coins dating to the reigns of Domitian, Trajan and Hadrian (for a discussion, see Sticotti 1905: 213, n. 1). The curse was inscribed on a lead sheet that is now broken into two pieces that fit together to form what was probably a rectangular tablet. Currently the tablet is in a sorry state, since it has lost all of its corners and a good deal of its edges; furthermore, there is considerable damage at ll. 6–7. The inscription contains 17 lines, which run from left to right and were written in old Roman cursive by a trained hand. The text has been dated to the second half of the second

After being inscribed, the tablet was rolled up.

103. Ateste, Este Provenance: necropolis. Current Location: MN Atestino. Date: 1st century BCE. [Inv. No.: 14309] Material: lead. Measurements: 11.5 × 29.3 × 0.2 cm. Reading:

Col. I + privaṭum Camidium Q(uintus) · Praesentius Albus Sec ̣unda · uxor · Prasenti T(itus) Praesentius 5 Maxsuma · T(iti) · Praesenti · uxor C(aius) · Arilius C(aius) · Arenus Polla Fabricia L(ucius) · Allius 10 L(ucius) · Vassidius Clemens

181

Col. II Prisca u`x´or Vassidi Monimus Acutius Ero[tis] Acutia C(aius) · P[r]o[-c.3-] Damio · l(ibertus) si quis [i]nimicus · inimi[ca] adve[r]sarius hostis Orce pater · [P]roserpina cuṃ tuo Plutone tibi · trado ut tu il(l)u(m) mit[t]as et deprem[as]

Sylloge of  Defixiones from the Roman West 5



Col. III tradito tuis canibus tricipitibu[s] et bicipitibus ut · eripiạ(nt) capita cogitạṭ(iones?) cor in tuom gemini[-c.3-]+ recip[i]ạ(nt) il(l)os ụṭ ++++++

Bibliography: Alfonsi 1914; AE 1915, 101; Besnier 1920: no. 39; Alfonsi 1922: 21; García Ruiz 1967: no. 61; Sup. It. 15.7; AE 1997, 584; Buchi 2001: 223–24; AE 2002, 562; Kropp 2008: no. 1.7.2/1; Urbanová 2018: no. 38; Gordon 2020a: 119–20.

The text contains three principal parts: a headline (I, l. 1), a list of the names of 13 victims (I, ll. 2–10 and II, ll. 1–4), and an invocation of the divine powers who are to act on the defigens’ behalf (II, ll. 5–10 and III). The text presents the following orthographic and phonetic features: Prasenti for Praesenti (I, l. 3), Maxsuma for Maxima (I, l. 5), mitas for mittas (II, l. 9), depremas for deprimas (II, l. 9) and tuom for tuum (III, l. 5). Following Gordon (2020a: 120), Camidium (I, l. 1), in the accusative, should be taken here as the nomen of the principal, while privatum is a substantive referring to his private property (see OCD, s.v. c). Camidius is presenting this curse tablet in the format of a legal document (compare with the headings of the Murecine tablets from Pompeii, which read vademonium factum). Despite the fact that the first name is in the accusative (I, l. 1), while the rest are in the nominative (I, ll. 2–10, II, ll. 1–4), we should interpret them all as the victims of the curse, since they are all considered enemies of the author (II, ll. 5–6). For the mixture of nominative and accusatives in lists of names, see section I.7.2. The tablet appears to belong to the group of juridical defixiones on the grounds that it has a certain all-inclusive formula: si quis (est) [i]nimicus, inimi[ca], adve[r]sarius, hostis (II, ll. 5–6), which is used to refer to the victims from the first part of the text and also those who are not mentioned explicitly (I and II, ll. 1–4; for adversarius, cf. 49, 135, 478, etc.; for inimicus, cf. 17, 137, etc. and also section I.7.5). Certain names in the curse are first attested here in Este: the nomina Camidius, Arilius and Arenus (I, ll. 1, 6 and 7, respectively) as well as the cognomina Albus (I, ll. 2; see Kajanto 19822: 226), Monimus, Erotis and Damio (II, ll. 2–4; see OPEL III, 86; II, 122 and 92, respectively). The curse is also of prosopographical interest: according to Buchi, both L(ucius) Vassidius Clemens and his wife Prisca belonged to one of the most prestigious families of Montagnana, which is also attested in Verona (2001: 224).

Image source: Alfonsi 1914: 371, fig. 1. See also App. IV.3, SD 103. Commentary: this defixio was discovered in 1908 during the excavations of the Rebato necropolis located in the socalled ‘Campo alto Cristo’ (Caldevigo, Este). The tablet was found inside one of the cinerary urns inside tomb 61, which had already been plundered in Antiquity (for the archaeological context, see Alfonsi 1914 and 1922). The curse was written on a rectangular lead sheet, which was cut out of a larger piece of lead. Despite being complete, the tablet has been broken into two parts that fit together (this breakage falls along the second fold line). The tablet is in relatively good condition, despite its surface being covered in a thin reddish patina and the presence of several cracks along the fold lines. The text contains 25 lines, which run from left to right and are arranged in three columns, the third of which has been rotated 90 degrees so that its first line runs from bottom to top alongside the right edge of column 2. This unusual layout has an easy explanation: here, the author of the text miscalculated the space needed, and so (s)he had to continue writing in the right margin, as usual in letter writing (see Sarri 2018: 112; for a similar case, see 494). The inscription is written in capitals, measuring between 0.3 and 1.1 cm in height. Here I present the drawing published in the editio princeps, together with my own reading (which is based on an autopsy carried out in May 2011). It must be stressed that my reading diverges slightly from that drawing in the following aspects: the underlined letters (present at the time of the editio princeps but which I could not see anymore), the underdotted letters (which I could see only partially) and the ending -bu(s) in tricipitibus (III, l. 2).

The defigens invokes the deities Proserpina cuṃ tuo Plutone (II, ll. 6–7) as well as tuis canibus tricipitibu[s] and bicipitibus (III, ll. 1–3). As Isidore tells us, Orce Pater 182

Italia and Pluto could be taken as the same divinity: Pluto latine est Dis Pater Orce pater, alii Orcum vocant (Orig. VIII, 11, 42; for a discussion cf. Bevilacqua and Colacicchi 2009: 310). Accordingly, it appears that the mention of Orcus Pater and tuo Plutone is somewhat repetitive. Pluto and Proserpina form an attested pair in other defixiones (e.g., 10–14 and 48), while Orcus Pater is unparalleled in the corpus of curses from the Roman West. The canes tricipites and bicipites are given the job of tearing off the victims’ capita cogitạṭ(iones?) cor (II, l. 4; for the sequence caput-cor, see 129; for cogitatio, see 339) and are identified as the famous Cerberus and his two-headed brother Orthus, who guarded the cattle of Geryon (he is only attested here in the curses from the Roman West).

After being inscribed, the tablet was folded in half, pierced with a nail (from which there is a hole with a diameter of 0.3 cm) and was then cut across the centre of the fold. Finally, it was folded in half again so that the text could not be seen. Translation (Gordon 2020a: 120, modified): ‘(In the matter of) Camidius’ property [thirteen names in nominative] (and) any person, male or female, ill-disposed (to me), (any) opponent, (any) enemy! (These) I transfer to you, Orcus Pater (and) Proserpina with your (husband) Pluto that you may take them and bring them low. I hand them over to your three-headed and two-headed dogs, that they may tear out their heads, thoughts, hearts, may carry them to your joint...’

104. Altinum, Altino Provenance: aquatic context (?). Current Location: MAN Altino. Date: end 1st century BCE–beginning 1st CE. [Inv. No.: AL 11277] Material: lead. Measurements: 10.7 × 13.1 cm. Reading: A L(ucius) Caulius · Hieronymuṣ L(ucius) Caulius Hieronymus · Stephanephoria · Secundus Onesimus · Festa · Diocles · 5 Daphnus · Proclus · Zmyrna Hieronymis · naus Trebius Severus · Maecius Carter Maịc ̣ị us · Berullus · Strenus · Muna tia Marcella Paetinụs 10 Septicianus Macrina [-c.3-]IA L(ucius) Caulịuṣ Hier[onymus]

B Ṣtephanephorịa Secundus Onesimus ·Fesṭa [Di]ocḷes Italus · Cervonịụ[s -c.3-]ONIVS O pilio Cervonia dedị de[f]ịctas

Bibliography: Scarfì 1972; Tirelli 1994: no. 20; Solin 1998: 52–53; Zampieri 2000: 131–33; Ferrarini and Cresci Marrone 2002; AE 2002, 556 a–b; Solin 2004: 128; AE 2004, 608; Solin 2005: 179–80; AE 2005, 558; Kropp 2008: no. 1.7.1/1; Urbanová 2018: no. 37.

sostegno di una grande costruzione in blocchi parallelepipedi di trachite, ubicata ai limiti meridionali dell’area urbana antica’ (Scarfì 1972: 55). The structure, probably an ancient canal, has been dated between the first century BCE and the first century CE (for a discussion, see Scarfì 1968).

Image source: Scarfì 1972, Tab. I–II.

The curse was written on a rectangular lead sheet, whose bottom left corner has been broken off. In general, the tablet is in good condition, despite some corrosion along the bottom and right edges as well as the bottom half of side B. An opisthograph, the inscription contains 15 lines, which run from left to right and were written in capitals,

Commentary: this defixio was discovered in 1965 during the excavations carried out by the Soprintendenza alle Antichità delle Venezie ad Altino; according to the editor, it was ‘a 2 metri di profondità, all’esterno delle palificazioni di 183

Sylloge of  Defixiones from the Roman West The second list, beginning again with Lucius Caulius Hieronymus at the final line of side A, repeats A, ll. 3–4 before adding Italus (B, l. 3), Cervonịụ[s] (a common cognomen in Altino, B, l. 3), [---]onius Opilio and Cervonia (B, ll. 3–4).

which measure between 0.3 and 1 cm in height. Several words are separated with interpunctuation consisting of small horizontal lines found halfway up the letters. The text consists of a long list of personal names, all of which are in the nominative. The lead character seems to be Lucius Caulius Hieronymus, who is mentioned on three occasions: twice at the beginning of the list (A, ll. 1–2) and then again at A, l. 11. The editor has taken A, l. 11 as the start of a second list, since several of the names are repeated. The first list contains several Greek names: the above-mentioned Hieronymus (A, ll. 1, 2 and 11) as well as Ṣtephanephorịa (A, l. 2), Onesimus (A, l. 3), Diocḷes (l. 3), Daphnus, Proclus, Zmyrna (all in A, l. 4) and Carter (A, l. 7). Among the Latin names, we find Satrenus (A, l. 8; perhaps for Satrenius or Satrinus?), Paetinụs (A, l. 9; see OPEL III, 120), Septicianus (A, l. 10; see OPEL IV, 69), Macrina (A, l. 10; see OPEL III, 43) and Berullus (l. 8; see OPEL I, 289). The reading NAVS (A, l. 6) has proven a bit more difficult: following Solin (1998 and 2004 contra Scarfì), we should take this as the praenomen naus, and hence read the tria nomina Gnaeus Trebius Severus.

The final phrase, dedị de[f]ịctas (B, l. 4), is used to consign these victims to a deity that was presumably invoked orally (on this, see section I.4.1). See 345 (l. 4: Tacita deficta) and 71 (IIA, l. 7: dificdos) for parallel uses of the participle. While the circumstances that led to the writing of this curse are unknown, Scarfì has conjectured that the victims probably belonged to an enterprise (either agricultural or commercial), of which Lucius Caulius Hieronymus was the boss. While this is an attractive hypothesis that would find some parallels in 101–02, the text’s lack of detail precludes a definitive answer. After being inscribed, the tablet was pierced with a nail, folded and later closed with another fold along the bottomright corner.

105. Cremona Provenance: unknown. Current Location: MA San Lorenzo. Date: 1st century CE. [Inv. No.: 347] Material: lead. Measurements: 12 × 15 cm. Reading: 5



Q(uintus) · Domatius · C(ai) · f(ilius) · bonum · tempus mihi · meaque · aetati · id · ego · mando · remandata quo · is · apud · deos ·iferos · ut · pereant · et [·] defigantur · quo · ego · heres · sim pupil(l)us · Corani(us) · C(ai) · f(ilii) · A(ulus) · Poblici(us) · populi · l(ibertus) Aprod[i]sị(us) · L(ucius) · Corneliu(s) · meo · sumtu defigo · illos · quo · pereant

Bibliography: Gualazzini 1966; Pontiroli 1974: 221–24; AE 1975, 449; Solin 1987; AE 1987, 455; Solin 1995: 575– 76; Solin 1998: 289–92; Solin 2004: 123–26; AE 2004, 616; Kropp 2008: no. 1.7.4/1; Urbanová 2018: no. 40.

the letters. The present reading generally follows the editio princeps, though I have made some changes based on an autopsy I carried out in November 2010. The text presents the following orthographic and phonetic features: meaque for meaeque (l. 2), ifero for infero (l. 4) pupilus for pupillus (l. 6), sumtu for sumptu (l. 7) and Aprodisius for Aphrodisius (l. 7). Solin (1987: 133) has proposed reading demandata for remandata (l. 3). Nevertheless, according to Lewis and Short, re-mando can also serve to repeat a command, and thus the sequence mando remandata (l. 3) would make sense here. Based on palaeographic and onomastic criteria, the curse dates to the Julio-Claudian epoch.

Image source: See also App. IV.3, SD 105. Commentary: this defixio was discovered in July 1964 ‘a circa m. 1’50 sotto il piano cantina di uno stabile in demolizione fra via Garibotti e via Alfeno Varo’ (Gualazzini 1966: 317). The curse was written on a rectangular lead sheet, which is in fair condition, despite some areas being seriously affected by corrosion. There are two areas that were damaged when the sheet was unfolded as well as a crack along the fold line.

Based on its content, this curse should be grouped among the juridical defixiones. The text has a somewhat anomalous structure in that it begins by requesting good fortune for the defigens (ll. 1–2). This sort of beginning is without parallel in the corpus of curses from the Roman West. Next, the text presents the curse proper, in which the author asks the infernal deities to deal with his enemies (ll.

The inscription contains eight lines, which run from left to right and were written in capitals (measuring between 0.2 and 0.6 cm in height). The words are separated with interpunctuation consisting of a single dot found halfway up 184

Italia 3–4). The victims are, in order of appearance, the pupil(l)us Corani(us), Aulus Poblici(us), the populi l(ibertus) Aprod[i]sị(us) and L(ucius) Corneliu(s). With respect to the freedman, Solin has pointed out, ‘les esclaves municipaux affranchis étaient souvent nommés coloniae ou colonorum ou bien municipii ou municipium libertus. A côté de ces désignations, un populi libertus serait facile à admettre’ (1995: 576). The nomen Domatius is documented (l. 1; see OPEL II, 104) as is the cognomen Coranius (l. 6; see OPEL II, 74). The inscription closes with the final cursing formula defigo illos quo pereant (l. 8; for other curses with the verb defigo, see 51, 108, 339 and 530).

After being inscribed, the tablet was folded so that the written surface was closed inside. Translation (Urbanová 2018: 246, slightly modified): ‘(I), Quintus Domatius, son of Gaius, (I wish that) I and my age(?) are happy. Therefore I convey the order/ message [i.e., the tablet with the curse] by which they may be accursed by infernal gods and perish, by which I shall become an heir: boy Coranius, son of Gaius, Gaius Publicius, the freedman, Aphrodisius, Lucius Cornelius. At my own cost, I accurse them so that they may die.’

106. Verona Provenance: necropolis. Current Location: SBAVV. Date: 2nd–3rd centuries CE. [Inv. No.: IG 124412] Material: lead. Measurements: 5.2 × 10.2 × 0.3 cm. Reading: 5

Trophimen Zosimen Chariten vindictam de illis fas

Bibliography: Cavalieri Manasse and Bolla 1998: 138–39; Buonopane 2000; AE 2000, 618; Buonopane 2002a; AE 2002, 512; Kropp 2008: no. 1.7.6/1; Urbanová 2018: no. 211.

larger sheet with irregular edges (one of the corners is missing). The sheet has an uneven surface and contains limestone incrustations, as I was able to confirm during an autopsy (carried out in September 2013). The inscription contains five lines, which run from left to right and were written in capitals (measuring between 0.5 and 0.7 cm in height). Palaeographically, we must note that the author’s handwriting suggests a familiarity with papyrus, as can be seen in the s of l. 2, which it must be noted is different from the s in l. 5. Given these palaeographic features, the text can be dated to either the second or third century CE.

Image source: Buonopane 2000: 135. Courtesy of A. Buonopane. Commentary: this defixio was discovered in 1990 during the excavations carried out in the area of Verona’s Porta Palio along the ancient Via Postumia, where a necropolis containing over 500 graves had been located. Said necropolis was in use between the first century BCE until either the third or fourth century CE. The curse was uncovered among the grave goods of cremation burial 142, which was covered with three tegulae. The grave goods have been dated to the second century CE and consisted of a glass bottle, a lamp (with the stamp CRESCES), a ceramic cup decorated with horizontal bands, ceramic sherds, nails for making shoes as well as larger iron nails. Among these items, archaeologists discovered the defixio, which in all likelihood the defigens deposited at a later date, taking advantage of the gaps between the tegulae (for the archaeological context, see Cavalieri Manasse and Bolla).

We do not know the circumstances that led to the writing of this curse, which mentions three slaves with Greek names (for Trophime, see Solin 2003: 1052; for Zosime, see Solin 2003: 890, 898 and 1477 and for Charite, see Solin 2003: 489 and 1481) and then contains a cursing formula. Buonopane has proposed the following analysis of the final formula: vindicta (l. 4) here means ‘punished’ and fas (l. 5) is a shortened version of either fas est or fas sit, a phrase which is frequent in religious contexts. The text, accordingly, requests the punishment of these three women (for a parallel expression, see 482, A, ll. 2–4: commendo deabus iniurium fas ut me vindicetis). Translation: ‘(I curse?) Trofime, Zosime, Charis. They should be punished.’

The curse was written on a lead sheet with the shape of an isosceles triangle, which had been cut out of a 185

Sylloge of  Defixiones from the Roman West 107. Verona Provenance: necropolis. Current Location: SBAVV. Date: 1st century CE. [Inv. No.: VR 19925] Material: lead. Measurements: 5.2 × 8.4 × 0.2 cm. Reading:

Secunda Sentia

a lead tag (cf. 99). Given the archaeological context, however, it seems more likely that this lead tablet is indeed a curse. The inscription contains two lines, which run from left to right and were written in capitals (measuring between 1.6 and 2.2 cm in height. We do not know the circumstances that led to the writing of this curse, which contains the name Secunda Sentia. Buonopane is correct to note the use of the cognomen Secunda as a praenomen, a phenomenon that is extremely well documented in the eastern part of Cisalpina between the first century BCE and first century CE.

Bibliography: Cavalieri Manasse and Bolla 1998: 138–39; Buonopane 2002b; AE 2002, 566; Kropp 2008: no. 1.7.6/2. Image source: Buonopane 2002b. Courtesy of A. Buonopane. Commentary: this defixio was discovered in 1990 along with 108 during the excavations carried out in the area of Verona’s Porta Palio along the ancient Via Postumia, where a large necropolis had been located. Said necropolis was in use between the first century BCE until either the third or fourth century CE. The curse was uncovered in cremation burial 1281, whose grave goods consisted of a broken lamp, an as from the reign of Antoninus Pius (dated between 140 and 144 CE; see RIC 3, 691(a) and 706 (a)) and a small knife, which should not be connected to the deposition of the defixiones, since this is a frequent good grave in the area (contra Cavalieri Manasse and Bolla 1998: 139; for a fuller discussion, see Sánchez Natalías 2016: 119).

Due to the name as well as the text’s palaeography, Buonopane has suggested that the tablet dates to the first century CE. This date, however, is clearly undercut by the dating of the burial’s grave goods, which are securely dated to the middle of the second century CE. There is a solution to this apparent contradiction: as G. Pelucchini has confirmed (per litt.), given the rushed nature of the emergency excavation grave 1281 was partially contaminated with grave 1318, which dates to the end of the first century CE. The accidental mixing of the two graves’ contents during the excavation could easily explain the inconsistency in the dates. If this explanation is correct, 107 actually came from grave 1318.

This curse was written on a rectangular lead sheet, with irregular edges and an uneven surface, as I was able to confirm during an autopsy (carried out in September 2013). The tablet has a small hole located near the lowerright corner, which suspiciously makes the tablet resemble

108. Verona Provenance: necropolis. Current Location: SBAVV. Date: second half of the 2nd century CE. [Inv. No.: VR 19924] Material: lead. Measurements: 10.6 × 12.1 × 0.1 cm. Reading: 5

Auguŝtum Carsadịa Secundum Caupunum def[ig]o

186

Italia Bibliography: Cavalieri Manasse and Bolla 1998: 138–39; Sánchez Natalías 2016a; AE 2016, 537; Urbanová 2018: no. 45.

the tablet is in good condition, though its surface is slightly corroded, as I was able to confirm during an autopsy (carried out in September 2013). The inscription contains five lines, which run from left to right and were written in capitals (measuring between 0.5 and 0.9 cm in height). The text, written for unknown reasons, contains a list of personal names in the accusative followed by the verb defigo.

Image source: Sánchez Natalías 2016a: 120, fig. 2. Commentary: this defixio was discovered in 1990 alongside the previous curse during the excavations of the necropolis found along the ancient Via Postumia, in the area of Porta Palio (Verona). The tablet was deposited in a cinerary grave 1281, which was covered with tegulae. The grave goods date to the second century CE and consisted of a broken lamp, an as from the reign of Antoninus Pius (dated between 140 and 144 CE; see RIC 3, 691(a) and 706 (a)) and a small knife. It is worth noting that around Alto Adige knives are frequently found in graves and accordingly the knife in burial 1281 should not be connected to the deposition of the defixiones (contra Cavalieri Manasse and Bolla 1998: 139; for a fuller discussion, see Sánchez Natalías 2016a: 119).

We must note the absence of the final -m in the second name (l. 2), which is probably explained as the author’s error, who included this letter in the rest of the names. Carsadịa (l. 2), can be interpreted as either an unattested Latin cognomen or more reasonably as an error for Carsidia, which is attested in Rome (see CIL VI, 14440). Likewise, the cognomen Caupunum (l. 4) can be understood in two ways: either as a new name derived from the attested Caupo or as an error for Cauponium. Despite implying both having u for o and the omission of an i, Cauponius is documented as a nomen (see Solin-Salomies 19942: 54). Finally, though rare in Latin curses, the verb defigo (l. 5) finds several parallels in the corpus of curses from the Roman West (e.g., 51, 105 and 339).

The curse was written on an irregularly shaped lead sheet, which was cut out of a bigger sheet. In general,

Sicilia 109. Lilybaeum, Marsala Provenance: necropolis. Current Location: M. Baglio Anselmi. Date: 3rd century BCE. [Inv. No.: 1647] Material: lead. Measurements: 3.8 × 10.9 cm. Reading: A καταδέω Ζωπυρίωνα τᾶς Μυμβυρ παρὰ Φερσε φόναι καὶ παρὰ Τιτάνεσσι καταχθονίοις καὶ παρ´ἀ π[ε]υχομένοισι νεκροις (vacat) {ἐς τοὺς ἀτελέστους} καὶ παρ ὰ [ἱ]αρίαις Δάματρος < καὶ > παρ’ ἀπευχομέ[ν]α[ισ]ιν (vacat) 5 καταδέω δέ νιν ἐμ βολίμωι α[ὐτον καὶ νοῦ]ν αὐτοῦ καὶ ψυχὴν αὐτοῦ ὡς μὴ δύνα[ται ἀντία] λαλῖν (vacat) καταδέω δέ νιν ἐμ βολίμωι σ[-ca. 5-] [-c.1-] ΥΝ [α]ὐτὰν καὶ νοῦν καὶ ψυ[χήν αὐτᾶς]

187

Sylloge of  Defixiones from the Roman West B [καταδέω] δὲ ὅπως [μὴ δύνανται] ἀντία [λέγειν] μτε πο[ιεῖν] [-] Iunius 5 [-] Septumius C(aius) Acinus M(arcus) An(n)ius L(ucius) Umbonius M(arcus) Nautius 10 M(arcus) Rustius L(ucius) Nautius Umbonia

Bibliography: Bechtold and Brugnone 1997; AE 1997, 737; Brugnone 1999: 467, 470–72; Jordan 1997: 391–96; Jordan 2001: no. 79; Kropp 2008: no. 1.11.1/1; Bernabé and Martín Hernández 2020.

third century BCE and its content places it in the group of juridical defixiones. The reading presented here follows both Jordan 1997 (side A) and the editio princeps (side B). The curse begins with the phrase καταδέω Ζωπυρίωνα τᾶς Μυμβυρ (A, l. 1), in which the first victim, qualified by a matronymic is cursed (Ζωπυρίων is Greek, whereas Μυμβυρ could be Phoenician-Punic in origin or Libyan). Next, the text invokes deities Persephone (Φερσεφόνα with an initial Φ just like in a defixio from Morgantina [SEG XXIX, 929]), the Titans (whose invocation is so far unparalleled in the corpus of Latin curse tablets from the Roman West, on this see Bernabé and Martín Hernández 2020), the atelestoi and Demeter’s priestesses. After that, the defigens asks for the victim’s mental capacities, including the ability to speak (see λαλῖν, A, l. 7), to be paralysed. Such a purpose may suggest that this curse should be classified among the juridical defixiones. In this case, the goal would be to prevent Zopyrion from testifying in court. This curse is then repeated on side B, where it is then followed by a list of additional victims. This list comprises eight citizens and a woman. The first group is identified with their praenomina and nomina. The lack of any cognomina lends support to the editors’ dating of the curse.

Image source: Bechtold and Brugnone 1997, Tab. XV, fig. 2 and Tab. XVI, fig. 2. Courtesy of A. Brugnone. Commentary: this defixio was discovered in 1991 during the excavation of the necropolis of Via Berta (ancient Lilybaeum, Sicily) in grave no. 186. This burial contained two inhumed individuals with their respective grave goods, among which two curse tablets were deposited: one written in Greek, the other in Greek and Latin. The latter, which is the subject of the present entry, was found among the earth used to fill the grave between the two inhumations. The archaeological context gives us a terminus post quem in the first quarter of the third century BCE (on this, see Bechtold 1997: 114). Upon its discovery, the curse was rolled up and stuck to an iron nail, contact with which led to the corrosion of the tablet’s surface. After restoration, the sheet was left in two parts that fit together and originally formed a strip of lead. The top and right edges have nearly been fully preserved, while the bottom-left corner has been lost. An opisthograph, the inscription contains both a Greek and Latin text: the Greek portion contains 11 lines, which run from right to left and occupy side A and part of side B (ll. 1–3). The Latin portion contains nine lines of capitals, which run from left to right and are perpendicular to the Greek text (B, ll. 4–12; for this layout, see 166). The text is dated on palaeographic grounds to the end of the

After being inscribed, the tablet was rolled up around a nail (for a parallel for this act, see 53). Translation (Jordan 1997: 396): ‘I bind down Zopyrion son of Mymbyr before Persephone and before the underground Titans and before the abominating (male) dead. (And I bind him down) also before the priestesses of 188

Italia Demeter and before (the) abominating (sc. female dead). And I bind him down in lead, him and his mind and soul, so that he will be unable to speak in opposition. And I bind

her down in lead, (i.e.) S[--]`yn, her and her mind and her soul. I bind them down so that they will be unable to speak or act against (me) [list of names follow].’

Sardinia 110. Orosei Provenance: unknown. Current Location: Museum of Nuoro. Date: 1st century CE. Material: lead. Measurements: 4 × 5.8 cm. Reading: 5

Ṃ(---?) · S(---?) · Petius · neni gostali Fausti Arsati Ṃosan · nurgo onu tono · nurgo benis s[---] Coturissonis ẹltibi nec Conatini Coturis Capito ite bulsu ite ụltibi nec Mosan nurgo · onu tono

Bibliography: Caprara 1978: 152–53; Sotgiu 1998: B 128; Mastino and Pinna 2008: 67–69.

a roughly rectangular lead sheet and contains five lines, which run from left to right and were written in old Roman cursive dating to the first century CE.

Image source: Mastino and Pinna 2008: 69, fig. 26. Courtesy of A. Mastino.

When it comes to diction, the editor has stressed that some of the words appear to recall ancient Sardinian, such as nurgo (ll. 2 and 5) whose root nur could be refer to the underworld, ‘nella sua accezione di “voragine”, e quindi probabilmente di luogo infero’ (Caprara 1978: 152).

Commentary: this fragment, which forms part of the so-called ‘Collezione Cabras’, probably comes from Orosei and was discovered on an unknown date and under unknown circumstances. The curse was written on

111. Orosei Provenance: unknown. Current Location: Museum of Nuoro. Date: 1st century CE. Material: lead. Measurements: 4.5 × (3.6) cm. Reading:

a[---] ito[---] Aeli Lesi[---] ASVPSA[---]

189

Sylloge of  Defixiones from the Roman West Bibliography: Caprara 1978: 154, Tav. LV, 2; Sotgiu 1998: B 129; Kropp 2008: no. 1.10.1/1; Mastino and Pinna 2008: 68–69.

Orosei and was discovered on an unknown date and under unknown circumstances. The tablet has an irregular shape and contains the remains of four lines of text, which run from left to right. Based on palaeographic grounds, the text has been dated to the first century CE.

Image source: Mastino and Pinna 2008: 69, fig. 28. Courtesy of A. Mastino. Commentary: this fragment, which forms part of the so-called ‘Collezione Cabras’, probably comes from

After being inscribed, the tablet was pierced.

112. Locus Incertus Provenance: unknown. Current Location: unknown. Date: unknown. Material: lead. Measurements: 16 × 21 cm. Reading: 5



10

[------] [------] [------] [---]numerum venerunt ra [---r]oco dom(i)ne ut[---] [------] rogo subruptus ṣịṭ Urvanus[---] [------] [---]denari[---] [---]rogo [---] cum pessim[---]

Bibliography: Gasperini 1992: 323; AE 1992, 911; Kropp 2008: no. 1.10.2/1; Mastino and Pinna 2008: 68; Urbanová 2018: no. 213.

to as dom(i)ne (note the syncopated form domne, l. 5). Despite Gasperini’s attempt to identify this deity as Dis Pater, we must note that domine is extremely generic and is used in other curses to refer to other deities, such as Attis (see 121, 486 and 490). It is worth noting the term subruptus (l. 7), a compound of the verb rumpo (to break, burst, tear) with the prefix sub (under), with which the author of the text describes the punishment of his/ her victim. While we cannot say with any certainty what kind of curse this is, the presence of the word denari (l. 9) points towards the category of the defixiones in fures. Urvanus (for Urbanus, l. 7) is a cognomen well attested on Sardinia.

Commentary: this defixio was discovered on Sardinia on an unknown date and under unknown circumstances. The curse was written on an irregularly shaped lead sheet, which contains 12 lines that run from left to right. The letters, some of which were written backwards, measure between 0.5 and 1.5 cm in height. Given the fragmentary nature of the text, it is no longer possible to discern the circumstances that led to the writing of this curse. Using the verb rogo (ll. 7, 10 and perhaps 5), the defigens invokes a deity, who is referred

After being inscribed, the tablet was pierced twice.

113. Neapolis, Terralba Provenance: domestic context. Current Location: unknown. Date: 3rd century CE. Material: ostrakon. Measurements: 7 × 2.3–2.6 × 0.7 cm. Reading:

Marsuas a Neapoli Dec(imum) Ostiliu- m Donatum `mis[er]um ´mutum surdum reddas quantu(m) homini respondes

190

Italia Bibliography: Zucca 2005a: 255; Zucca and Cossu 2005: 212–18; Zucca 2007: 1531–34; AE 2007, 690; Mastino and Pinna 2008: 68–69.

height. The text has been dated to the third century CE on palaeographic grounds. Although the expressed desire to make the victim mutum surdum (ll. 2–3; cf. 125, 477–78) could suggest that this curse ought to be grouped among the juridical curses, Zucca and Cossu have provided a different analysis in which they argue that respondes (a second person singular) can only be referring to Marsyas and refers to the oracular response. In this case, the defigens’ goal is to impede communication between the victim and deity, which would be quite unusual. The invocation of the satyr Marsyas (who discovered the flute and challenged Apollo to a musical contest) has no parallels among the Latin curses of the Roman West.

Image source: Zucca and Cossu 2005: 215, fig. 5.12. Courtesy of R. Zucca. Commentary: this defixio was discovered in 2000 during the excavations of the monumental area of Sardinian Neapolis’ northern sector (modern Terralba). The curse was written on a rectangular ostrakon, which had originally been part of a polished amphora. The use of an ostrakon as the surface on which to write a curse is not attested in either the literary or archaeological record in the Roman West.

Translation: ‘Marsyas of Neapolis, make wretched, dumb and deaf Decimus (?) Hostilius Donatus, as far as you can answer the human being.’

The text, etched with a metal-tipped stilus, contains four lines of capitals, which measure between 0.3 and 0.6 cm in

114. Giuncalzu Provenance: on the ground. Current Location: SBASN. Date: 5th century CE. Material: lead. Measurements: vid. reading. Reading: Frag. 1 ((3.7) × (7) × 0.05 cm) A [------] [---]omṇi scabie [---]++ [---lig?]ạte damṇịa[te?---] [---]T omṇib[us?---] ++IOM ita[---][---]ÊNO sị (h)abe[ant or -at?---] ++++S [---]OSMPACCOSA totidie BÊN[---] 5 [fa]ṃula (traces) ÊN (traces) TR[---] [pr]oteg[ite] iurạ ligo demones[---] [-c.4-]patrọ[no---] ligo aqua[m?] TILASA[---] B [------] [---]NVR de faṃula dedọ[---]imudes ̣ [---]neque ma[l]os (h)omines ṇe[que---] data A[---] [---n]eque notem dormi[re---]D[-c.1-]LA 5 +[-c.3-]+T[---] (traces) [---]VITE cubạte RNAT (vacat) [---]E (vacat) Frag. 2 ((1.6) × (0.6) × 0.05 cm) A [---]++[---] [---]at[---] (vacat) B (vacat?)

191

Sylloge of  Defixiones from the Roman West Bibliography: Bevilacqua 2010a; Del Monaco 2017.

and phonetic features: totidie for cottidie (frag. 1, A, l. 4), imudes for immutes (frag. 1, B, l. 2), omines for homines, notem for noctem (both in frag. 1, B, ll. 3–4). It is worth noting the use of the terms scabies (frag. 1, A, l. 2, maybe related with the punishment of the victims), famula (frag. 1, A, l. 5 and B, l. 2 instead of the most common serva) and the verb cubo (frag. 1, B, l. 6). The targets of the text, whose names are unknown, were probably the female slave (frag. 1, A, l. 5 and B, l. 2) and her master (frag. 1, A, l. 7). The defigens calls upon the deities invoked (frag. 1, A, l. 6) with a series of commands, whose endings -ate (frag. 1, A, ll. 2 and 6) have been preserved. The presence of [n]eque notem dormi[re] (frag. 1, B, l. 2) opens up the possibility that the curse contained a longer version of a non permittas formula used to harm the health and well-being of the victims (for some parallels, see 352, 358, 443, etc.). Probably, the expression [pr]oteg[ite] iurạ ligo demones (frag. 1, A, l. 6) is an oath formula, perhaps related to the causa defigendi (on this, see Bevilacqua 2010: 1958).

Commentary: a resident of Giuncalzu (near Olbia) chanced upon this curse tablet, which was broken into several fragments, eight of which were found on the ground. They originally belonged to a tablet that was in all likelihood rectangular. After being inscribed, it was folded (which led the tablet being broken). An opisthograph, the inscription contains a text of at least seven lines that run from left to right and were written in new Roman cursive (on this, see Bevilacqua 2010: 1945), whose letters measure between 0.2 and 0.3 cm in height. Based on palaeographic considerations, the text has been dated to the fifth century CE. Following the proposal of Del Monaco (2017), seven of these fragments join together (here, they are referred to as ‘Fragment 1’), while an eighth remains unplaced (‘Fragment 2’). Given that the text is incomplete, it is no longer possible to discern the circumstances that led to the writing of this curse. The text presents the following orthographic

After being inscribed, the tablet was folded several times.

Corsica 115. Mariana, Lucciana Provenance: necropolis. Current Location: unknown. Date: 1st century CE. Material: lead. Measurements: 7 × (20.5) cm. Reading: 5

[---]VLE · vindica · te · qui · tibi · male [·] f[ecit] qụị [-c.4-5-]+++ ̣ [vi]ndica · te · `et ·´ si · C(aius) · Statius · tibi [·] nocuit · ạ[b ·] ẹọ · vind[ica · te]+++++ [persequa]ris · eum · ut · male · contabescat · usque · dum · morị[et]ur+++ [qui]cumque · ali(u)s · et · si · Pollio conscius est · et · illum · persequaris n[e] annum · ducat

Bibliography: Moracchini-Mazel 1974: 18–19; Solin 1981: 120–21; AE 1982, 448; Versnel 1991: 82; Solin 1995: 575, n. 20; Kropp 2008: no. 1.9.1/1; Urbanová 2018: no. 212.

Commentary: this defixio was discovered in the I Ponti necropolis (ancient Mariana, Corsica) during the excavation carried out in 1965. The curse was found in burial no. 12, a cremation grave, whose grave goods (dated to the first century CE) consisted of two lamps, two ceramic cups and the remains of a glass jar.

Image source: Moracchini-Mazel 1974: 19, fig. 46. © fonds FAGEC. See also App. IV.3, SD 115. 192

Italia see 350 and 460). Next, the defigens employs another allinclusive formula ([qui]cumque a(l)ius, l. 4), which is immediately followed by another named victim, Pollio (si Pollio conscius est, l. 4). If this individual is guilty of any wrongdoing, he ought to die within the year (for a parallel, see 92; for magical time frames in general, see Sánchez Natalías 2019a).

The curse was written on a strip of lead, which is now broken in three fragments that fit together. Only part of the upper edge of the tablet has survived. The inscription, written in capitals (but note the use of old Roman cursive b, d and e), contains five lines, which run from left to right. The words are separated with interpunctuation consisting of a single dot found halfway up the letters. The height of the letters ranges from 0.4 to 0.8 cm. Versnel has proposed the reading male f[ecit] (l. 1) and [qui]cumque ali(u)s (l. ̣ 4), both of which are accepted here.

Several words are notable and deserve brief mention: the imperative vindica (cf. 482), the phrase si conscius est (l. 3), which puts a possible limit on Pollius’ punishment (cf. 451 and 479), and the use of the verb persequor (for parallels, see 127, 493 and 523). Also, note the use of variation between male fecit (l. 1) and nocuit (l. 2) within a repetitive text.

We do not know the circumstances that led to the writing of this curse, in which the defigens asks for the deity to take vengeance on his behalf. While many practitioners from Britannia seek to make the deity into an injured party and hence guarantee his/her intervention (e.g., 205, 349, etc.), this curse is unique in making the deity into the sole victim of a crime. The text is made up of a list of (potential) wrongdoers who deserve punishment. In l. 1, there is a sort of all-inclusive formula, before Caius Statius is specifically named: if he did the god any wrong (si C(aius) Statius tibi nocuit, l. 2), the deity ought to hound him until he dies (usque dum morị[et]ur, l. 3; for this magical time frame

Translation (Versnel 1991: 82): ‘-ule (probably the name of a god), avenge yourself. Whoever has done you harm... avenge yourself on him, and if C. Statius has injured you, avenge yourself on him [… persecute] him in order that he may waste away horribly until he dies. And whoever else —for instance, if Pollio— is an accomplice, persecute him as well, so that he won’t live out the year.’

Locus incertus 116. Locus incertus 1 Provenance: unknown. Current Location: unknown. Date: 1st century CE. Material: lead. Measurements: 6.5 × 10 cm. Reading:

M(arcus) Afarius Ampliatus

Bibliography: Garrucci 1866: 27–28; CIL X, 504; Wünsch 1900: no. 6; DT 211; Jeanneret 1917: no. 211; Kropp 2008: no. 1.6.1/1; Urbanová 2018: no. 37.

this curse, which contains only the victim’s name (for the cognomen Ampliatus, see Kajanto 1982²: 349). It appears that after being inscribed, the tablet was pierced three times in its upper section, which, for Garrucci, indicates that ‘dovette essere infissa a qualche nicchia di cinerario’ (1866: 28). There are other examples of nails used to mount a tablet (e.g., 95 and 129).

Commentary: this defixio, which is held in a private collection, probably comes from the region of Lucania. The curse was written on a lead sheet of unknown shape. We do not know the circumstances that led to the writing of

193

Sylloge of  Defixiones from the Roman West 117. Locus incertus 2 Provenance: unknown. Current Location: MACB. Date: 4th–5th centuries CE. [Inv. No.: ROM 2238-2241] Material: lead. Measurements: vid. reading. Reading: Frag. I ((13.9) × 9.9 cm). φωρβη charaktêr ψιαο τιωρ φωρβεθ βραι βαριω φωρβεν ω βαθακαρ φωρβι (crux) καμφι 5 φωρρω (magical ρηο sign) οεφ[ρνβ] φωρβι Por- ιαγαακ[ερβε] φωρρα cel- Cηθ· O lus · 10 Porcellu mo- Porcel[lus] lom[e]dicu · medicụ[s] molomedicu · inṭerficite · omṇ[e] c ̣orpus · caput · tente · oculuṣ +ạ(or ụ (?))+TAS · facite · Porcellu(m) · et · [Mau-] 15 rilla(m) · usure(m) · ịpsius · DITE [---] EM · corpus · omnịṣ · menbra · [---] bisc[e]d[a] · Porcelli · qui · ICED[---] [cada]t · languạt · et ru[at ?] ̣ (traces) 20 Porcelluṣ (crux) +[Po]rcellu[s et] [m-]   [au-] rilḷ-   [am] ipsi-   us molo-   medicus





Porcelluṣ

mulo · Porce[llus mu-] lo · molomedico interficite · eụm ·occidite eni[ca]25 te profucate · Porcellu(m) · et Mau rilla(m) · usure(m) · ipṣ[i]us · anima · cor ·+ ṇata · (h)epar · ISI [-c.4-]Ẹ+ṂR(or S (?)) (traces) VṢ (?) -----

Frag. II ((3.6) × (5.3) cm).



[---] (traces) [febres(?)]+ṭẹrc ̣ianas · quarṭana[s---] [pal(?)]luris · frigora · morbu · EM[---] [---P]orcellus · molomedicus ·[---] [---a(?)]rdor · AVTVS · ATON · ad+[---] [---]+ASEIMI · AMOR+AB ·ACE[---] [---]·GYMNI · AMORFVS · PANT[---] [---]MERAS · CEORAS · MACAS+[---] [---]CTRIAS · CATAPOMAS · SI+[---] [---]AS · ENPRACTIAS LESON+[---] [---]EISOPEISOS[-c.1-]T[---] [---]E (or C (?)) S[---]

5



10



194

Italia Bibliography: Olivieri 1899: 194–97; Besnier 1920: nos. 1, 3 and 4; García Ruiz 1967: nos. 5, 7 and 8; Kropp 2008: nos. 1.1.2/1, 1.1.2/3 and 1.1.2/4; Sánchez Natalías 2011a; AE 2011, 377; Urbanová 2018: nos. 3 and 5.

have a fairly regular ductus. The words are separated with interpunctuation consisting of a single dot found halfway up the letters. In fragment 1, the text presents the following orthographic, morphological and phonetic features: loss of final -m (Porcellu(m) ll. 14 and 25; Maurilla(m) ll. 14–15 and 25–26), molomedicus for mulomedicus (ll. 10–12, 23 and the fifth ascending line), tentes for dentes (l. 13), oculus for oculos (l. 13), usure for uxorem (ll. 15 and 26), omnis for omne (l. 16), menbra for membra (l. 16), languat for langueat (l. 19) and profucate for praefocate (l. 25). In fragment 2, we find [pal]luris for [pal]lores (l. 2), morbu for morbos (l. 2) and again, molomedicus for mulomedicus (l. 3). The curse uses several verbs that are rare or unparalleled in the corpus of defixiones from the Roman West: langueo (languạt, l. 18; cf. 118 and 96), ̣ interficio (interficite, l. 24) and praefoco (profucate, l. 25). Furthermore, the sequence occidite eni[ca]te (ll. 24– 25) and the word biscida (l. 17; perhaps a vulgarization of viscera) are only paralleled in 118. Porcellus is a diminutive from porculus and Maurilla is derived from the cognomen Maurus (see Kajanto 1982²: 328 and OPEL III, 68, respectively).

Image source: Sánchez Natalías 2011a: 204, fig. 1. See also App. IV.3, SD 117. Commentary: the archaeological context of this defixio, which was written on a lead sheet that was probably rectangular, is unknown. Nevertheless, it is very likely that it was found together with 118, with which this curse shares several iconographic and textual features. That said, and given the palaeographic differences, it is worth noting that 117 and 118 were written by different hands. As I was able to confirm during my autopsies (carried out in 2009 and 2010), tablet 118 is broken into 12 pieces, 11 of which fit together and form ‘fragment 1’. Although the other part, which does not preserve any of the tablet’s original edges, does not fit with fragment 1, it surely comes from the same defixio. Despite having been broken into so many pieces, the tablet is in good condition. Iconographically, this defixio is of great interest and importance: it not only represents the victim, who is depicted as a mummy in the lower part of the tablet, but also the deity invoked. The deity is shown standing with crossed (or tied) hands and bare feet. On his/her chest, there is a magical symbol and a series of onómata; in the genital regions we find an eight-pointed star. From each side of his/her crowned head, three sinuous snakes emerge (reconstructed in part on analogy with 118). This recalls a passage from PGM in which Hekate-Selene is described as ἡ φοβερῶν ὀφίων χαίτην σείουςα μετώποις (IV, 2800–02, that is ‘[you] who shake your locks of fearful serpents on your brow’, translation by E.N. O’Neil apud Betz 19922). Although we cannot firmly identify the god in this tablet with Hekate-Selene, since the iconography does not represent all of her known characteristics (like being trimorphic or bearing a torch), it is possible that this is an idiosyncratic depiction of said deity. The position of the deity’s arms, crossed or tied, is surprising, since this pose is normally reserved for victims. As I see it, the posture can be interpreted in two ways that are not mutually exclusive. In the first case, the position of the hands should be understood as a divine gesture causing the immobility of the victim and related to the unleashing of the snakes, who will carry out the curse. In a different vein, we could see this as a sign that the deity is in a sense bound: by inscribing the curse, the defigens (or a professional acting on his/her behalf; see section I.2) is forcing the god to take action.

We do not know the circumstances that led to the writing of this curse, which targets the veterinarian Porcellus and his wife Maurilla. Fragment 1 begins with an invocation of the deity, which includes a series of Greek onomata barbarika (this section is almost identical to 118, ll. 1–9). In the remainder of fragment 1, the victim is introduced and cursed. After being identified by his profession, various parts of Porcellus’ body are cursed (his head, teeth, eyes, limbs, organs, spirit, heart, buttocks (?) and liver) before the text breaks off. This section of the text uses tricolons of verbs ([cada]t lạngụạt et ru[at], l. 18 and occidite eni[ca]te praefocate, ll. 24–25). Fragment 2 begins with the unleashing of a series of diseases meant to harm the victims (the classic fevers as well as [pal]luris (?), frigora and morbu, l. 3). From the fifth line onwards, however, the defigens (or person working on his/her behalf) transitioned to writing in a sort of coded language (perhaps his/her own idiosyncratic magical language?). While there are some transcribed Greek words into the Latin alphabet in this section (e.g., gymni, fragment II, l. 7), ll. 5–12 are mostly filled with what phonetically appears to be Greek but does not actually make any sense. Translation (Sánchez Natalías 2011a: 207 and 215, modified): (fragment I) ‘Porcellus. Porcellus the veterinarian. Porcellus the physician. The veterinarian. Destroy his entire body, his head, teeth, eyes… Let Porcellus and his wife, Maurilla, be… (?) May all Porcello’s body, limbs, entrails… disintegrate, languish and collapse. Porcellus the veterinarian and Maurilla his (wife). Porcellus. Porcellus. Veterinarian (?) Porcellus the veterinarian. Destroy, crush, kill, strangle Porcellus and his wife Maurilla. Their soul, heart, buttocks, liver…’

The inscription, written in Latin and Greek, contains more than 40 lines that run from left to right as well as an additional five lines that run from bottom to top. The palaeographic characteristics of the new Roman cursive text date it to either the fourth or fifth century CE (cf. 22 and 31). The letters (measuring between 0.5 and 0.8 cm in height) are neatly separated from one another and

(fragment II): ‘…tertian, quartan, (fevers?) …pallor, colds, diseases …Porcellus the veterinarian …Fire (?) …’

195

Sylloge of  Defixiones from the Roman West 118. Locus Incertus 3 Provenance: unknown. Current Location: MACB. Date: 4th–5th centuries CE. [Inv. No.: ROM 2238-2241] Material: lead. Measurements: (11.5) × 6.5 cm. Reading:

φωρβη charaktêr για τιωρ φωρβεν βιρα βαριω φωρβεο ω βαθαcωρ φωρβι κανφι 5 ρηο φωρβω (crux) οεβφηβ φωρβι ιαγαακερβε Fis φωρρο tu LHO Fistụ[(m)] 10 Fistu(m) sina occi- sinator(m) D+ dit̂e CAEQEṂ[---] [---]ṭure(m) occi- inic- TIV occi[dite] [di]t̂e ini- at̂e QAN[---] [ca]t̂e (traces) ++ 15 [occid]it̂e inca[te] Fi[stum] F̣istus diffloiscat lânguat [m- (?)] ergat et disulvit̂e omnị[a] menbra omnis viscida ip sius ·Fisṭi disolbit̂e ṃen20 bra biscida la[ng]ụạ[t] runpit̂e binas ipsiu[s] runpit̂e omnis menb[ra] F̣ịṣti ṣinat[o]rịṣ[---]

Bibliography: Olivieri 1899: 195–96; Besnier 1920: 7, no. 2; García Ruiz 1967: no. 6; Kropp 2008: no. 1.1.2/2; Sánchez Natalías 2012a; Urbanová 2018: no. 4.

Greek, contains 23 lines that run from left to right and were written in new Roman cursive and also in scriptio continua (with letters measuring between 0.15 and 0.6 cm in height). Given the presence of ligatures and the same letter written in more than one way (e.g., the e in et and disolvite in l. 17), it is clear that a trained hand wrote the text. The curse’s palaeographic features date it to the fourth or fifth century (for similar texts, see 5, 16 and 22). For the iconography, see the commentary to 117.

Image source: Sánchez Natalías 2012a: 141, fig. 1. See also App. IV.3, SD 118. Commentary: the archaeological context of this defixio, which was written on a roughly rectangular lead sheet, is unknown. Nevertheless, it is very likely that it was found together with 117, with which this curse shares several iconographic and textual features. That said, and given the palaeographic differences, it is worth noting that 117 and 118 were written by different hands. Today, the tablet is broken in two fragments that fit together. Although some sections are corroded and some areas along the edges are fairly uneven, the curse is generally in good condition, as I was able to confirm during my autopsies (carried out in 2009 and 2010). The inscription, written in Latin and

The text presents the following orthographic, morphological and phonetic features: senator and sinatur for senator (ll. 10–12 and 23; cf. App. Probus 64: senatus, non sinatus), inicate for enicate (ll. 11–14), difloiscat for diffluat (ll. 16), disulvite and disolbite for dissolvite (ll. 17 and 19), languat for langueat (ll. 16 and 20), omnis for omnia (l. 18), menbra for membra (ll. 18 and 22), viscida and biscida for viscera (ll. 18 and 20) and runpite for rumpite (ll. 21 and 22). This curse draws on a similar lexicon as 117, since 196

Italia have not identified a Late-Antique Fistus. That said, if we correct the name to Festus, PLRE contains several possible candidates. Though proposing specific identifications can be tempting, the lack of a full onomastic formula and the absence of specific information about this curse’s find spot make identifying this sinator both difficult and necessarily speculative.

both use forms of the verbs occido-enico, langueo and the noun viscida. The use of the verb diffluo (diffloiscat, l. 16) is only paralleled in a curse from Mainz (497), a site where several tablets use the image of the victim’s body dissolving or flowing like salt in water (e.g., 488 and 499). The Latin cognomen Fistus is attested in the fasti (where the consul P. Curiatius Fistus Trigeminus is mentioned in the year 453 BCE; on this, see Kajanto 1982²: 210).

Translation (Sánchez Natalías 2012: 144): ‘Crush, kill Fistus the senator. Crush, kill Fistus. Crush (…) Fistus the senator… Crush, kill Fistus the senator. May Fistus dilute, languish, sink and may all his limbs dissolve, all his entrails (?) of Fistus. Dissolve his limbs and entrails (?), may he languish. Burst his veins (?), break all his limbs. Fistus the senator…’

We do not know the circumstances that led to the writing of this curse, which targets the senator Fistus. This is the only case of a curse clearly taking aim at an individual from the senatorial order (there are few curses in general that attack illustrious victims; see 105, 134–36, 149 and 478). Though clearly identified as a senator in the text, prosopographers

119. Locus Incertus 4 Provenance: unknown. Current Location: M. Paul Getty. Date: 1st century CE. [Inv. No.: 77.AC.97] Material: lead. Measurements: 9.3 × 12.6 cm. Reading:

Col. I Philocomus Antioc(h)us P(h)arnace(s) Sosus Erato Epidia



Col. II tabesca(n)t domnis non pla`ceat´ eide(m) his quorum nom[ina] hic sunt perea[nt] quo e(t) placean(t) peculio il(l)orum dic˹at˺ facta ad inferos + [---]

5

Bibliography: Jentoft-Nilsen 1980; Solin 1988; AE 1988, 1146; Solin 1991; Solin 1995: 574–75; Solin 2004: 119– 22; AE 2004, 1895; Kropp 2008: no. 1.5.6/1; Urbanová 2018: no. 35.

written in capitals, which date to the end of the Republican period. We do not know the circumstances that led to the writing of this curse, which contains a list of victims (col. I) and a cursing formula (col. II). The text presents the following orthographic and phonetic features: Antiocus for Antiochus, Parnace for Pharnaces (I, ll. 2 and 3, respectively), tabescat for tabescant, eide for eidem (II, ll. 1–3), placeat/placean for placeant (II, ll. 2 and 5–6), ilorum for illorum and dicta por dicat (II, ll. 7–8). Note the use of ther verb tabesco (II, l. 1; for parallels, see 56 and 67). Solin has argued that his (II, l. 3) is an archaic form of hi (1991: 493).

Image source: R. Kotansky, apud Jentoft-Nilsen 1980: fig. 2. See also App. IV.3, SD 119 and prolegomena, fig. 5.4. Commentary: this defixio, which was supposedly found in an aquatic context, was donated to the Paul Getty Museum at an unknown date. The curse was written on a rectangular lead sheet, which is nearly whole, though quite corroded. There is damage on the bottom-left corner and along the bottom edge.

The names of the victims, Philocomus, Antioc(h)us, P(h) arnace(s) and Sosus (I, ll. 1–4) suggest that they were a group of slaves of Greek origin (for these names, see Solin 2003: 168, 209, 243, 1450 and 1348, respectively). There was another victim, Epidia Erato (I, ll. 5–6), a liberta

The inscription was arranged in two columns (of six and eight lines), whose text runs from left to right and was 197

Sylloge of  Defixiones from the Roman West probably of Campanian origin, given that the Oscan name Epidius/a is well attested in that area. The second part of the curse contains the cursing formula, which is directed against the above-mentioned individuals, quorum nom[ina] hic sunt (II, ll. 3–4). Significantly, there is a nail that pierces the word nomina in this phrase. This could be understood as an attempt to forge an unbreakable bond between the victims and the spell, since in ancient magic, names and individuals were inextricably linked (cf. section I.7.2). After being inscribed, the tablet was pierced twice and then folded. Translation (Urbanová 2018: 120, slightly modified): ‘Philocomus, Antiochus, Pharnaces, Sosus, Erato, Epidia, may they languish, may they fall into disgrace of their masters, and may those, whose names are (inscribed) here, die... (I commend?) their words and deeds to the infernal gods...’

198

2 Hispania Hispania, which constitutes the westernmost provinces of the Roman Empire, is one of the earliest provincial regions to adopt the practice of writing curses on lead tablets. The earliest examples date to the Republican period and hail from three important urban centres: Emporion, Carmo/ Carmona and Corduba/Cordoba. During the High Empire and coinciding with the rise of the epigraphic habit, there was a noticeable increase in the number of curses produced, most of which come from the southern portion of the Iberian Peninsula, especially from Saguntum. In Late Antiquity, the practice of writing defixiones slowed in Hispania, just as it did in the rest of the Empire. Accordingly, we only have a few isolated Late-Antique curses from sites like Braga (see 146–47) and Villanueva del Cauche (see 130).

second century and mid-third century CE, was made of grey clay and is decorated with vertical lines on its body. It also has a graffito that reads Olla Saturn[---] habebit dua gena irata (see Hervés Raigoso 1995: 122–24, HEp 13, 432 and Marco Simón 2011a). [Hispania Citerior] Saguntum, Sagunto: Corell (1994: 285) mentions an uninscribed defixio that was found together with 142 around 1980, from the mountain El Castel. This rectangular lead sheet measures 4.8 × 17.8 × 0.1 cm and dates, according to its editor, between the first and third centuries CE. As I was able to confirm during an autopsy of the tablet in April 2010, it is very similar to 141 in shape and size. Before being deposited, it was rolled up. This tablet would be the first uninscribed curse from Hispania, though others have been found in other sites of the Roman West (e.g., 159). On this, see HEp 1, 824; Corell, Gómez and Ferragut 1999: no. 8; IRPV 18; Corell 2009: 68.

In addition to the items discussed in this chapter, we must add several others that are suspecta (doubtful), still being studied or were found long ago and subsequently lost before being published properly (e.g., the tablets from Bolonia and Enguídanos). Here, I provide a brief discussion of such tablets organized by provincia and the date of publication.

[Hispania Citerior] Emporion, Ampurias: after being discovered, two defixiones are currently being prepared for publication. The first was found in 2002 during the excavations of La Granada necropolis. The tablet was deposited in a child’s grave (for preliminary information, see Aquiluè et al. 2004: 43 and 46). The second curse, which apparently was not inscribed, was found next to a grave. We must also note that a medicinal label had been wrongly identified as an eleventh Latin curse from Emporion (see IRC III, 178).

[Hispania Citerior] Remeseiros (near Aquae Flaviae, Chaves): in 1729 an inscribed piece of granite (measuring 125 × 260 cm) was discovered. The inscription contains five lines, which run from left to right and were written in capitals. The words are separated with interpunctuation consisting of a single dot found halfway up the letters. Hübner included the inscription in CIL (II, 2476), where it is described as imprecationis vivae rupi inscriptae, similis eius quam supra dedimus inter Emeritenses ad deam Ataecinam directae.

[Hispania Citerior] Turiaso, Tarazona: during the excavation of a sanctuary, where there was an aquatic cult (perhaps dedicated to the deity Silvis-Minerva), a piece of terra sigillata hispanica, measuring (10) × (11) × (0.8), was found. It was inscribed with a brief graffito, for which F. Beltrán Lloris (2004: 256–57) has cautiously given the reading [---]PIS taceat. If this is correct, the fragment could be a curse meant to silence its victim (whose name would be Greek given the ending –pis). The fragmentary nature of the text along with the unusual object in which it was inscribed (terra sigillata) prevent us from determining whether this is truly a defixio.

Given the difficulties involved reading and interpreting the text, I here include the text published in CIL: in · ac · conducta · conservanda/ oi · in · ac · conducta · P · mici/ · invola++icquuaequnque res · ae · MII/ A+++S· si · L · siqui · ea · S · V · S · E · V · IANCE · CI (see AE 1980, 580; AE 1981, 527; AE 1985, 577; AE 1987, 562q; HEp 2, 875, AE 1993, 1027; HEp 5, 1058; HEp 7, 1245). [Hispania Citerior] Enguídanos: M. López (1949: 40) mentioned two inscribed lead tablets whose whereabouts were unknown (according to this author, it is possible that they were ‘melted for making bullets’).

[Hispania Citerior] Tarraco, Tarragona: during a 2006 intervention at a plot located at Antoni Company i Fernández de Córdoba Street, no. 7, part of a necropolis belonging to the western part of the city was uncovered. It contained at least 17 burials and an ustrinum dating to the High Empire. A defixio was found in grave 8, an inhumation burial that had already been ransacked in Antiquity. Nevertheless, the following grave goods were

[Hispania Citerior] Lucus Augusti, Lugo: during the 1986 excavations carried out in Ferrol Square, archaeologists discovered a cinerary urn measuring 18.5 × 21 cm (height × diameter). The urn, which dates between the end of the

199

Sylloge of Defixiones from the Roman West found: a lamp, an iron nail, three bronze appliqués and a small wooden box. An opisthograph, the curse has been broken into eight fragments and has a Latin text written in capitals (for the preliminary notice, see Ciruana 2009: 221–24). [Hispania Ulterior] Baelo Claudia, Bolonia: P. Paris (et al., 1926: 89) mentioned the discovery of ‘une tablette magique en plomb, avec un trou de suspension’, which has apparently not been published. Its current whereabouts are unknown. [Hispania Ulterior] Castulo, Linares: B. Díaz, F. Arias and A. Guzmán are preparing the edition of an opisthographic lead tabula ansata that contains a curse. I thank B. Díaz for this information.

200

 Hispania Hispania Ulterior Lusitania 120. Emerita Augusta, Merida Provenance: unknown. Current Location: MNAR Merida. Date: second half of the 2nd century CE. [Inv. No.: 10302] Material: marble. Measurements: (24) × 22 cm. Reading:   5 10

Dea · Ataecina Turibrig(ensis) · Proserpina per tuam maiestatem te rogo oro obsecro uti vindices quot mihi furti factum est quisquis mihi imudavit involavit minusve fecit [e]a[s res] q(uae) · i(nfra) · s(criptae) · s(unt) tunicas · VI [pa]enula lintea · II · in[dus]ium · cuịus · I+C+V[-c.3-]M ignoro ỊA[---]ius VI[---]

Bibliography: CIL II, 462; Wünsch 1897: 25; DT 122; ILS 4515; Jeanneret 1917: no. 122; Preisendanz 1933: 160; Rubio and Bejarano 1955: no. 183; AE 1959, 30; Álvarez Sáenz de Buruaga 1957; HAEp 1957–960, 11; D’Ors 1960: 329; AE 1961, 102; ILER 736; García Iglesias 1973: 69–74; Versnel 1991: 91; Astori 2000: 63; Kropp 2008: no. 2.3.1/1; Tomlin 2010: 247–53; AE 2010, 108; Marco Simón 2011b; AE 2011, 457; Velaza Frías 2016; Urbanová 2018: no. 219.

The inscription, surrounded by a border that is 6 cm wide, contains (at least) 13 lines of capitals, which run from left to right and whose letters measure between 1.5 and 2.5 cm in height in the heading and between 1.2 and 1.5 cm in the following lines. Given the palaeographic features of the text, García Iglesias (1973: 72) has proposed a date in the mid-second century CE. Based on its content, this curse is a classic example of a defixio in fures. Here, the defigens asks the goddess Proserpina Ataecina of Turibriga to punish a thief (for the theonym, see Marco Simón 2011b). In an invocation that recalls 240 (ll. 1–6: dea Sul[i] Minervae rogo [s]anctissimaṃ maiestatẹṃ tuam u[t] vindices ab hiṣ [q]ui [fraude]m fecerụnt), the defigens calls upon the deity’s maiestas (l. 2) and then reinforces the invocation with a tricolon of verbs in l. 4 (rogo, oro, obsecro). With this formula, the defigens requests divine intervention to secure vengeance (uti vindices, l. 5; for the verb, cf. 240, 340 and 365). The curse first refers to the thief as qui mihi furti factum est (ll. 5–6; for this common formula, see e.g., 240 and 449), which is then underscored with quisquis mihi imudavit involavit minusve fecit (ll. 6–8). The verb involavit, it must be said, is extremely common in the corpus of British curses (e.g., 210, 220 and 451), whereas the verb immuto (‘to change,’ in l. 7: imudavit for immutavit) is not (see Velaza Frías 2016). The phrase minus fecit is a reference to the wasting away of the thief’s possessions (for a parallel, see 483, ll. 4–5: cuius Veronius res minus fecit…). The sequence [e]a[s res] q(uae) i(nfra) s(criptae) s(unt) (l. 8) is quite bureaucratic and almost legalistic (for the abbreviation see 441). After this abbreviation, there follows a list of the stolen objects, which are mostly articles of clothing. The last part of the text has been lost but could very well have ended with a final invocation of the deity,

Image source: see App. IV.3, SD 120. Commentary: this defixio was discovered around the year 1826 attached to the wall at La Albuera marsh, where it had been used for sharpening knives (hence the cuts between ll. 6 and 9). The inscription then disappeared until Álvarez Sáenz de Buruaga located it in a private collection in Mérida in 1955. The curse was written in a small white marble plaque, which has been broken along the bottom edge and at the top right corner. The backside of the plaque was not polished according to García Iglesias (1973: 70). The use of a marble plaque for inscribing a curse is extremely uncommon in the corpus of defixiones. In fact, this curse has only one parallel in the entire corpus of curse tablets from the Roman West, 72, which was written on a slate plaque and attached to a funerary monument in the Pompeiian necropolis of Porta Nocera. It seems clear that in both cases the choice of a material common in monumental epigraphy was related to the desire to publicly display a defixio. This cannot be denied in the Pompeiian case, since to this day it remains in situ (for a fuller discussion of these two curses, see Sánchez Natalías 2016b: 74ff.; also see Tomlin 1988: 62, 2002: 166 and 2010: 249). 201

Sylloge of  Defixiones from the Roman West as is the case in similar curses (e.g., 365, 353, among others).

majesty, I beg you to avenge the theft which has been done me, whoever has changed, stolen, diminished the things which are written below. 6 tunics, 2 linen cloaks, a shift (?)... I do not know…’

Translation (Tomlin 2010: 249; modified): ‘Goddess Ataecina of Turibriga, Proserpina, I ask you by your

121. Salacia, Alcácer do Sal Provenance: sanctuary. Current Location: MNA Lisbon. Date: second half of the 1st century CE. Material: lead. Measurements: 15.3 × 9.2 × 0.15 cm. Reading: A Domine · Megare invicte · tu · qui · Attidis corpus · accepisti · accipias · corpus · eius · qui · meas · sarcinas · 5  supstulit · qui · me · compilavit de · domo · Hispani · illius · corpus tibi · et · anima(m) · do · dono · ut · meas res · invenia(m) · tunc · tibi · (h)ostia(m)

B quadripede(m) · done · Attis · voveo si · eu(m) · fure(m) · invenero · dom(i)ne Attis · te · rogo · per · tu(u)m · Nocturnum ut · me · quam · primu(m) · compote(m) · facias

Bibliography: Faria 2000: 105–11; D’Encarnaçao 2000: 1503–05; D’Encarnaçao 2001: 244–46; AE 2001, 1135; D’Encarnaçao and Faria 2002; Guerra 2003; Marco Simón 2004; HEp 11, 705; Ribeiro 2006: 242–43; Kropp 2008: no. 2.3.2/1; Tomlin 2010: 260–64; Versnel 2010: 297–300; HEp 19, 550; AE 2010, 108; Urbanová 2018: no. 220.

The curse was written on a rectangular lead sheet, which is fully preserved. It has a small hole between the words Domine and Megare (A, l. 1), which Faria has suggested could have served as a way of attaching the tablet to another object. That said, given the use of interpunctuation consisting of dots found halfway up the letters, it seems more likely that this hole was in origin an interpunct that was inscribed too deeply and hence pierced the tablet. An opisthograph, the inscription contains 12 lines, which run from left to right and were written in capitals (but note the cursive shape of the e made of two vertical strokes as well as the f made of a vertical stroke and a small oblique one). Based on its content, this curse should be grouped among the defixiones in fures. The text presents the following orthographic and phonetic features: supstulit for sustulit (A, l. 5), quadripede for quadrupedum (B, l. 1), done for dono (B, l. 1), tum for tuum (B, l. 3), unc for hunc (A, l. 8), ostia for hostia (A, l. 8) and the omission of final -m in anima (A, l. 7), invenia (A, l. 8), eu fure (B, l. 2), compote (B, l. 4) and primu (B, l. 4).

Image source: D’Encarnaçao and Faria 2002: 261, fig. 2. Courtesy of J. D’Encarnaçao. Commentary: this defixio was discovered during the 1995 excavations carried out in the Roman sanctuary of Salacia, which is located in the north-east part of the forum. The sanctuary’s temple has a rectangular plan and consists of two cellae connected by passageways. In the entrance to the smaller chamber, there is a small, compartmentalized space that was used to receive votive offerings (lamps, terracotta vases, etc.). Inside this cella, there is a small square tank, which Faria claims was used for sacrifices. A defixio was discovered at the bottom of this tank together with a copper coin, whose imprint can be seen on the back of the curse tablet.

This curse conforms to the general ‘template’ of a defixio in fures: the defigens, the victim of a theft offers the 202

 Hispania deity the thief’s body and soul (A, ll. 6–7; for this type of ‘offering’ see 349 and 308) and asks that the stolen goods be located as soon as possible (quam primu, B, l. 4; for a similar magical time frame see quanpr[imu]ṃ in 452, l. 6; for magical time frames in general, see Sánchez Natalías 2019a). In an attempt to increase the likelihood that the god will help, the defigens also offers a sacrificial victim (quadripede(m), B, l. 1; an unparalleled offering in the corpus of curses from the Roman West). The word Hispani (A, l. 6) has been interpreted as either a reference to the thief’s origin (Guerra 2003 and Tomlin 2010), as a reference to the origin of the owner of the house where the theft took place (Urbanová 2018: 261), or as a personal name (Faria 2000 and Marco Simón 2004). This third possibility seems the most likely, since the author would be referring to the place where the theft took place: the house of Hispanus (for a parallel, see 518; for this wellattested personal name in the Iberian Peninsula, see Abascal 1994: 387).

Faria identified the deity invoked, Domine Megare (A, l. 1), as Megara, daughter of Creon and wife Hercules. That said, Marco Simón (and subsequently Tomlin) have objected based on the gender of Domine and rather have identified the god as Hades, king of the underworld, who received the body of Attis. Indeed, in addition to invoking Domine Megare, the defigens also calls upon Attis, who is asked to help recover the purloined objects with the help of Nocturnus (on this deity, see Buonopane 2016). Translation (Versnel 2010: 297, slightly modified): ‘Lord, great and invincible, you who have received the body of Attis, please receive the body of the one who took away my baggage, who robbed me from the house of Hispanus. His body (viz. the thief ’s) and his soul I give (you), I confer on you in order that I may find my possessions. I vow to give you a four-footed sacrificial victim, Lord Attis, if I find that thief. Lord Attis, I implore you by your Nocturnus, that you make me obtain my wish as soon as possible.’

Hispania Ulterior Baetica 122. Corduba, Cordoba Provenance: necropolis. Current Location: MAN Madrid. Date: 1st century BCE–1st century CE. [Inv. No.: 1955/29/1] Material: lead. Measurements: 7.4 × 5.4 × 0.68 cm. Reading: Dionisia Denatiai ancilla · rogat deibus (!) ego rogo bono(m) bono(m) deibus (!) rogo oro bono(m) 5 einfereis · bono(m) · Salpina rogo oro et bonis · inferis ut DIOSO quod fit deibus (!) inferabus ut hoc quo(d) sit causa et ecquod votum 10 feci ut solva(s) rogo ut illam ducas rogo oro

Image source: Navascués 1934: Plate I. See also App. IV.3, SD 122.

Bibliography: Navascués 1934: 52–55; AE 1934, 23; Rubio and Bejarano 1955: no. 183*; Vázquez de Parga 1955–57: 59–60; HAEp 12–16, 2051; García Ruiz 1967: no. 58; Solin 1968: no. 22; ILER 5913; HEp 10, 163; Museros Ortiz 2000; Marina Sáez 2001: 57–58; Vaquerizo Gil 2001: 193; Stylow 2005: 255, 261; Ribeiro 2006: 253; Díaz Ariño 2008: no. U33; Kropp 2008: no. 2.2.3/1; CIL II²/7, 250; Marco Simón 2013; Sánchez Natalías 2014; AE 2014, 648; Urbanová 2018: no. 53; Marco Simón 2019: 387.

Commentary: this defixio was discovered together with 123 and 124 in 1932 during the excavation of a necropolis near Camino Viejo de Almodóvar, in which both Roman and indigenous Iberian artefacts were found. The curse was written on an irregularly shaped lead sheet. As I was able to confirm during an autopsy 203

Sylloge of  Defixiones from the Roman West in June 2013, the tablet is in excellent condition. It is worth mentioning, though, that the surface of side A is smooth, while that of side B is not. As Navascués has suggested, the imperfections of side B appear to be a result of the manufacturing process, in which the molten lead may have been poured onto a stone mould, which left its imprint on side B. This hypothesis would explain the tablet’s thickness (0.68 cm at the thickest point), which presents an outlier when compared to other tablets (generally between 0.1 and 0.2 cm).

Luzón (1975: 125; contra Navascués 1934: 54) have reconstructed deorsum from DIOSO. Scholars have tended to identify Salpina as a divinity, who was either indigenous (Navascués 1934: 54; Vázquez de Parga 1955–57: 60; Marco Simón 2013: 581ff.) or to be identified with Proserpina (e.g., Museros Ortiz 2000: 87; Marina Sáez 2001: 57–58; Díaz Ariño, Stylow apud CIL). That said, it seems much more reasonable that Salpina is the curse’s victim (Sánchez Natalías 2014). Indeed, Salpina would be the feminine form of the cognomen Salpinus, which is attested in a funerary inscription from Cannae (CIL IX, 319). Marina Sáez (2001: 58, esp. n. 64) has rightly pointed out that the ending –a is often used for the dative singular in Hispanic inscriptions of a similar date; accordingly, Salpina could be taken as a dative (also see Villar 1986). In this case, Salpina is the recipient of the final bonum, for which Dionysia ironically and repeatedly asks the infernal deities. By the end, Salpina becomes an offering for the infernal gods: ut illam ducas rogo oro (l. 11).

The inscription contains 12 lines, which run from left to right and were written in capitals (but note the cursive shape of e with two vertical strokes, f containing a long and a short vertical strokes and o formed with two curved strokes). The text has been dated to the first century BCE on palaeographic grounds. Some words are separated with interpuncts consisting of a single dot found halfway up the letters, though these are not used consistently (in fact, they are used only in ll. 2, 5 and 6). We cannot conclusively determine what kind of curse this is (contra Solin 1968 no. 22 and Urbanová 2018: 258, who believe that the tablet is an erotic curse). The text presents the following orthographic and phonetic features: Dionisia for Dionysia (l. 1), Denatiai for Dentatiae (l. 1), deibus for dibus (ll. 2, 4 and 7), bono for bonum (ll. 3–5), einfereis for inferis (l. 5), inferabus for inferis (l. 8), quo for quod (l. 8) and solva for solvat (l. 10; see Marina Sáez 2001). Gil and

Translation (Marco Simón 2019: 387, modified): ‘Dionysia, slave of Dentatia, asks the gods: I ask a good, a good of the gods, I ask and beg a good of the gods of the underworld for Salpina; I ask that she be offered down to the gods of the underworld; she is the reason and for whom I have made this vow. I ask that you fulfil it. I ask and beg that you take her.’

123. Corduba, Cordoba Provenance: necropolis. Current Location: MAN Madrid. Date: 1st century BCE– 1st century CE. [Inv. No.: 1955/29/2] Material: lead. Measurements: 4.7 × 5.72 × 0.25 cm. Reading: A T(itus) · Noster Fausta · Faustus Pollio · filius

B Casius Clipius Mun̂nitia

204

 Hispania Bibliography: Navascués 1934: 55–58; AE 1934, 24; Vázquez de Parga 1955–57: 60; HAEp 12–16, 2052; García Ruiz 1967: no. 59; Solin 1968: no. 23; ILER 5914; Vaquerizo Gil 2001: 193; Stylow 2005: 255, 261; Díaz Ariño 2008: no. U35; Kropp 2008: no. 2.2.3/2; CIL II²/7, 252; Urbanová 2018: no. 54.

and were written in capitals (but note the cursive shape of e consisting of two vertical strokes, o formed with two curved vertical strokes, etc.). Navascués has proposed a dating near the end of the first century BCE or beginning of the first century CE, whereas Stylow (2005: 255) has opted for a later date in the mid-first century CE. On side A, the words are separated with interpunctuation consisting of a single dot.

Image source: Navascués 1934: Plate II. See also App. IV.3, SD 123.

The circumstances that led to the writing of this curse are unknown since it only consists of a list of personal names. On side A, they are referred to with cognomina (perhaps since they were members of the same family), whereas on side B victims are named with nomina (this is uncommon in Hispania).

Commentary: this defixio was discovered together with 122 and 124 in the necropolis near Camino Viejo de Almodóvar during the excavation of 1932. It is a small, roughly rectangular lead sheet that was cut from a larger piece of metal. The curse tablet is in excellent condition, as I was able to confirm during an autopsy in June 2013. An opisthograph, the inscription contains six lines, which run from left to right

After being inscribed, the tablet was folded in half so that side B was on the inside.

124. Corduba, Cordoba Provenance: necropolis. Current Location: MAN Madrid. Date: 1st century BCE–1st century CE. [Inv. No.: 1955/29/3] Material: lead. Measurements: (8.1) × 0.17 (Ø × thickness) cm. Reading: C(aius) · Nu(misius) Sex[to?] C(aius) · Num(isius) · P(h)ileṃ[on] Num(isia) · (H)era[cli]a Calịpṣo · Num(isiorum) 5  C(aia) · Avilia · Ir[e]na C(aius) · Num(isius) · Epa[p(h)]rodi[tu]ṣ C(aius) · Num(isius) · Ae[s]c(h)inus Scinti[ll]a · Num(isiorum)

Almodóvar together with 122 and 123. According to Navascués, the curse tablet was still folded and broken into two fragments when it was found. Furthermore, one of the fragments was pierced with an iron nail. When unfolded, the tablet broke into a total of five fragments, all of which fit together. Only the two pieces of the left half of the tablet, however, are conserved today: the other three fragments were never catalogued at the Museo Arqueológico Nacional in Madrid (Vázquez de Parga, 1955: 61) and can only be studied from the photographs and sketches that Navascués made. Today, the two remaining fragments are in good condition, as I was able to

Bibliography: Navascués 1934: 58–60; AE 1934, 25; Vázquez de Parga 1955–57: 60–61; HAEp 12–16, 2053; García Ruiz 1967: no. 60; Solin 1968: no. 24; ILER 5915; Vaquerizo 2001: 193; Stylow 2005: 255, 261; Díaz Ariño 2008: no. U34; Kropp 2008: no. 2.2.3/3; CIL II²/7, 251; Urbanová 2018: no. 55. Image source: Navascués 1934: Plate III. See also App. IV.3, SD 124. Commentary: this defixio was discovered in 1932 during the excavation of the necropolis near Camino Viejo de 205

Sylloge of  Defixiones from the Roman West confirm during an autopsy in June 2013. It is worth noting that two concentric circles can be found on the back of the curse. While the editor has taken these as a decorative motif, it is more likely that they are a by-product from the manufacturing process, as is the case with 122.

CE. We do not know the circumstances that led to the writing of this curse, which contains the names of the victims, all of whom were members of the gens Numisia. Many of the cognomina are Greek (P(h)ileṃo, (H)era[cli]a, Epa[p(h)]rodi[tu]s and Ae[s]c(h)inus, ll. 2–3 and 6–7, respectively), which suggests that they were liberti. Two slaves (Calịpṣo and Scinti[ll]a, ll. 4 and 8, respectively) are also among the victims, as well as a woman (C(aia) Avilia Ir[e]na, l. 5), who likely married into the gens Numisia.

The inscription contains eight lines, which run from left to right and were written in capitals (but note the cursive shape of e made of two vertical strokes). The words are separated with interpunctuation consisting of a single dot found halfway up the letters. Based on palaeographic characteristics, Navascués has dated the text to the end of the first century BCE or the beginning of the first century

After being inscribed, the tablet was folded and pierced with an iron nail measuring 20.8 cm in length.

125. Corduba, Cordoba Provenance: necropolis. Current Location: MA Cordoba. Date: second half of the 1st century BCE. [Inv. No.: 32.632 and 32.632/1] Material: lead. Measurements: 2.4 × 10.5 × 0.1 cm. Reading: A Priamus · l(ibertus) · mutus · sit omnibus · modis B ni · qis · pos(s)it · de reditate `silient´ quet (!) · `ha´ṇnue verbum facere · oṃnes · omụṭ[e]sq[ua]ṇt Bibliography: Ventura Villanueva 1996: 143–49; Marina Sáez 2001: 58–60; Ventura Villanueva 2001: 193; Stylow 2005: 255 and 261; Ribeiro 2006: 252; Díaz Ariño 2008: no. U36; Kropp 2008: no. 2.2.3/4; CIL II²/7, 251a; Urbanová 2018: no. 56; Marco Simón 2019: 386.

in the first century BCE (note how e, o and f are made of two vertical strokes, the m of four and how the short stroke of a is diagonal or vertical). Both tablets were inscribed by the same hand. With the exception of B, l. 3, the author systematically used interpunctuation consisting of a single dot found halfway up the letters.

Commentary: this defixio was discovered by chance with 127 in a dump called El Arenal, which had been filled with earth taken from a plot of land located on Abéjar Street. This had been the site of an ancient necropolis located along the Via Augusta at the edge of Corduba. Both curses were found inside of a small cinerary urn dated during the second and first centuries BCE (for a discussion, see Vaquerizo 2001: 192). Analysis of the ashes determined that the remains belonged to a child, that is to say an aoros. For that reason, this urn proved to be an attractive place to deposit these two curses (see 96 for a parallel).

Based on its content, this curse should be grouped among the juridical defixiones. Indeed, in this type of inscription, the desire to silence victims is often expressed explicitly (e.g., 131, 522, etc. See section I.7.5). The text presents the following orthographic and phonetic features: qis for quis (B, l. 1), posit for possit (B, l. 1), dereditate for de hereditate (B, l. 1), ni for ne (B, l. 1), silient for sileant (B, l. 2), quet for quid (B, l. 2), hannue for annue (B, l. 3) and omutesquant for obmutescant (B, l. 4). After being inscribed, the tablet was rolled up together with 126 in such a way that the present curse was on the inside (and hence better protected), while the other was on the outside.

The present curse is broken into four pieces that fit together, as I was able to confirm during an autopsy in February 2011. An opisthograph, the inscription contains six lines, which run from right to left with the letters ‘mirrored’ (i.e., written from right to left). The text was written in capitals (measuring between 0.4–1 cm on side A and 0.3–0.9 cm on side B), whose features point to a date

Translation (Marco Simón 2019: 386): ‘May Priamus the freedman hold his tongue in all circumstances. Do not allow anyone to utter a word about the inheritance; let them all remain silent; let them shut up!’ 206

 Hispania 126. Corduba, Cordoba Provenance: necropolis. Current Location: MA Cordoba. Date: second half of the 1st century BCE. [Inv. No.: 32.633] Material: lead. Measurements: 5.5 × (7) × 0.1 cm. Reading: A [--- de]mentia [---]ORE · sin [--- om]utesquant

B [---] c ̣onari[---]NOT[a---] [---]CVO · sib[i -c.3?-] +EN [--- m]alevoluṣ · [-c.2-]OM[-c.1?-] [--- omut]esquant · DSVE 5 [---] ạnue · hered[es?] mutuis IN[-c.1?-] [---]SI [---]RE

Bibliography: Ventura Villanueva 1996: 143–49; Ventura Villanueva 2001: 194; Ribeiro 2006: 253; Díaz Ariño 2008: no. U37; Kropp 2008: no. 2.2.3/5; CIL II²/7, 251a; Marco Simón 2019: 386–87.

written from right to left). The text was written in capitals (note how e and o are made of two vertical strokes, the m of four and how the short stroke of a is diagonal or vertical) measuring between 0.5–1.4 cm on side A and 0.2–0.7 cm on side B. Based on palaeographic characteristics, the text should be dated to the first century BCE. Some words (in A, l. 2 and B, ll. 2–5) are separated with interpunctuation consisting of a single dot found halfway up the letters. Based on its content, this curse should be grouped among the juridical defixiones. Accordingly, it likely complements the previous defixio (125), with which it shares several lexemes, such as [om]utesquant (for obmutescant, A, l. 3 and B, l. 4) and hered[itate] (B, l. 5). If we accept Ventura Villanueva’s reconstruction for B, ll. 2–3 (1996: 147–48), the deity invoked is a [g]en[ius] malevolus (a ‘malignant spirit’). We must note, however, that there is no parallel for such an epithet in the corpus of curses from the Roman West.

Image source: see App. IV.3, SD 126. Commentary: this defixio was discovered by chance with 125 inside a cinerary urn containing the remains of a child. The urn was found in a dump where earth taken from a taken a plot of land located on Abéjar Street was relocated (for the archaeological context, see 125). As I was able to verify during an autopsy in February 2011, the curse was written on an irregularly shaped lead sheet, which is now broken into three pieces that fit together. An opisthograph, the inscription contains 11 lines, which run from right to left with most of the letters ‘mirrored’ (i.e.,

127. Italica, Santiponce Provenance: domestic space. Current Location: MA Seville. Date: first half of the 2nd century CE. [Inv. No.: I-G1215] Material: lead. Measurements: 10 × 9.2 × 0.1cm. Reading: A Dom(i)na fons fonṭ[i] ut tu persequaris duas res demando quiscunque caligas meas tel5 luit et solias tibi dea demando ut ˹tu˺ illas abce(p)tor si quis puel(l)a si mulier sive [ho]mo involavit B [in] illos persequaris

207

Sylloge of  Defixiones from the Roman West Bibliography: Gil and Luzón 1975; AE 1975, 497; Castillo 1979: 58, n. 1; Canto 1985: no. 20; HEp 1, 530; HEp 4, 725; Marina Sáez 2001: 103–04; Ribeiro 2006: 244; Kropp 2008: no. 2.2.4/1; Tomlin 2010: 253–58; HEp 19, 296; AE 2010, 108; Urbanová 2018: no. 218.

l. 1; cf. 337), quiscunque for quicumque (A, ll. 3–4), telluit for tolluit (A, ll. 4–5), solias for soleas (A, l. 5) and puela for puella (l. 8). The verbs tollo (A, ll. 4–5) and involo (A, l. 9) are quite common in defixiones of this genre (e.g., 120, 337, 367, among others). Given the text’s palaeographic characteristics, the reconstruction fonṭ[i] (A, l. 1, Tomlin: 2010: 254) is more convincing than the proposals Fori[nae] (Gil and Luzón 1975: 127) or Fovens (Canto 1985: 161).

Image source: Gil and Luzón 1975: 124. Commentary: this defixio was discovered in 1972 during the excavations of the northern sector of ancient Italica in the so-called ‘Casa de la Cañada Honda’ (for a discussion see, see Gil and Luzón 1975: 122). Inside the house, which was built during the reign of Hadrian, archaeologists uncovered objects dating between the mid-second century and the third century CE. The curse was found at floor level.

After the initial invocation of dom(i)na fons (l. 1), the defigens asks the deity to find the purloined objects with the phrase persequaris duas res (A, ll. 2–3; Gil and Luzón read tuas). Here the duas refers to each of the pairs of shoes that had been stolen. With this phrase, the defigens offers the goddess a reward for finding the thief. By transferring the stolen goods to the deity (A, ll. 6–7: ut tu illas abce(p)tor), the defigens converts a normal theft into a sacrilegious act and hopes to prompt divine intervention (for a similar idea, see 231, 237, 239, 337, etc.). Here we acknowledge the difficulties of interpreting the word abce(p)tor (l. 7): Gil and Luzón (1975: 129–30) read aboitor, while Canto (1985: 161) reads ad˹cep˺tor. However, a palaeographic analysis of the curvature of the vertical stroke in the b (different from that of the d) and the long curve of the c supports the reading *abce(p)tor, which can be taken as a word for ‘thief,’ which would fit well in the context (Tomlin 2010: 256). The text concludes with the allinclusive formula si quis puel(l)a, si mulier sive [ho]mo (A, ll. 7–9), which is extremely common in this genre of curse and widely paralleled in Britannia, though found here for the first time in Hispania (for Britannia, see e.g., 459, l. 4, which reads: si baro si mụlier si pueḷ[l]a…). In the same sentence, the defigens also reiterates the desire that the divinity track down whoever is responsible for the theft.

The defixio was inscribed on a sheet with the shape of a tabula cum capitulo (contra Urbanová 2018: 250), a common shape found in writing practices, which has here been rendered in lead and hence adapted for a magical purpose (for parallels, see 128 and 163). As I was able to confirm during an autopsy in February 2011, the tablet is in good condition, though it has lost its top-right corner and has been damaged near the bottom-left corner. The surface of side A is uneven and somewhat corroded, especially near the top-left corner. The capitulum is 1.5 cm high and has a slight break (which, contra Canto, is not a hole). At an unknown date, the capitulum was broken off from the rest of the tablet, but by the time I had examined the curse, it had been restored. An opisthograph, the inscription contains 10 lines, which run from left to right and were written in old Roman cursive (note the close similarity between b and d). The text has been dated to the first half of the second century CE on palaeographic grounds. The letters measure between 0.3 and 1.3 cm in height on side A and between 0.4 and 0.9 cm on side B.

Translation (Tomlin 2010: 258): ‘Lady Spring, I entrust two things to your spring that you exact them, whoever stole my boots and sandals. I entrust them to you, Goddess, that you exact them, whoever the thief that stole them, whether boy or girl, woman or man.’

Based on its content, this curse should be grouped among the defixiones in fures. The text presents the following orthographic and phonetic features: domna for domina (A,

128. Baelo Claudia, Bolonia Provenance: sanctuary. Current Location: MA Cadiz. Date: 2nd century CE. [Inv. No.: 16551] Material: lead. Measurements: 9.5 × 5.9 × 0.1 cm. Reading: Isis Muromem tibi c ̣ọnmendọ furtu(m) meu(m) mi(hi) fac tu{t}ọ numini maẹṣ5 tatị exsempḷariạ ut tu evide(s) immedio qui fecit autụlịṭ aụṭe(m) res operṭoru(m) albụ(m) nọṿ(um) ṣṭragulu(m) 10 nov(um) lodịc ̣es dụaṣ ṃe[o?] uso rogọ domina per maieṣṭạṭẹ(m) ṭua(m) uṭ (h)ọc ̣ furṭu(m) reprị ṇdas

208

 Hispania Bibliography: Bonneville, Dardaine and Le Roux 1988; AE 1988, 727; HEp 2, 227; Alvar 1994: 21 and 26 (no. 19); RICIS 602/0101; Ribeiro 2006: 246–47; Dardaine et al. 2008: 118–19, 155–57, 206 and 215; AE 2008, 668; Kropp 2008: no. 2.2.1/1; Tomlin 2010: 258–60; Versnel 2010: 283–86; HEp 19, 114a; AE 2010, 108; Urbanová 2018: no. 217.

exemplaria (l. 5), evide for evite(s) (l. 6), immedio for in medio (ll. 6–7), autulit for abstulit (l. 7), aute for autem (l. 8), opertoru for opertorium (l. 8), albu for album (l. 9), nov for novum (ll. 9–10), stragulu for stragulum (l. 9), maiestate for maiestatem (l. 12), tua for tuam (l. 12) y oc for hoc (l. 13). In a characteristic attempt to encourage the deity’s intervention, the defigens dedicates the stolen goods to the goddess (ll. 2–3). This transforms a simple theft into a sacrilegious act. The curse’s language recalls that of 237 (ll. 1–4): dẹae Ṣuli Minerv(a)e Sọlinus dono ṇumini ṭuo maiesṭaṭị pạx̣sa(m) ba(ln)earẹṃ et [pa-] l[l]eum, where the defigens also invokes the deity’s numen and maiestas. Next, the practitioner asks that the thief be executed publically as a punishment: ut tu evide(s) immedio qui fecit autụlịṭ aụṭe(m) res (ll. 6–8). The curse closes with rogọ domina per maieṣṭạṭẹ(m) ṭua(m) uṭ (h)ọc ̣ furtu(m) repriṇdas, which is used to reiterate the desire for Isis to intervene on the defigens’ behalf and punish the wrongdoer (cf. 121 and especially 365).

Commentary: this defixio was discovered in 1970 during the excavation carried out in the temple of Isis at Baelo Claudia, where it was found next to the well located in the south-eastern corner of the temple’s courtyard (for a discussion, see Dardaine et al. 2008: 206, 118–19, fig. 59 and 60). The curse was written on a tabula cum capitulo (contra Urbanová 2018: 250), a form frequently used in writing practices, which has here been rendered in lead and adapted for a magical purpose (for parallels, see 127 and 163). At the base of the projecting tab, there are two visible groves, which the editors have argued were used to attach the curse to a wall. Needless to say, this theory cannot be verified. The inscription contains 14 lines, which run from left to right and were written in old Roman cursive (note the cursive shape of b, q, e, m, etc., while d resembles a capital). The letters measure between 0.1 and 0.9 cm in height. The text has been dated to the second century CE on palaeographic grounds. Although the curse tablet is fully preserved, it is in a sorry state, since, as I confirmed during an autopsy in February 2011, a portion of the text has been lost after its last restoration.

The invocation of Isis is notable, since it is unparalleled in the rest of the curses from Hispania. That said, the cultic worship of Isis in Baetica between the first and third centuries CE (especially in Italica and Baelo Claudia, where sanctuaries have been found) explain the invocation (see Alvar for a discussion). The epithet muromem (l. 1) should be connected to the well attested Myrionyma (see Bonneville et al. 1988: 22, n. 4). Translation (Tomlin 2010: 260): ‘Isis Myrionyma, I entrust you with what has been stolen from me. Make me proofs of your divinity and majesty, so that you publicly take away the life of the man who did this theft, indeed who stole my property: a new white coverlet, a new rug, two used blankets. I ask you, Lady, by your majesty, that you punish this theft.’

Based on its content, this curse should be grouped among the defixiones in fures. Here, the theft of various articles of clothing is denounced before Isis (ll. 8–11). The text presents the following orthographic and phonetic features: furtu for furtum (ll. 3 and 13), meu for meum (l. 3), mi for mihi (l. 3), tuto for tuo (l. 4), exsemplaria for

129. Carmo, Carmona Provenance: unknown. Current Location: unknown. Date: second half of the 1st century BCE. Material: lead. Measurements: 5 × 32 × 0.2 cm. Reading: Dis · Imferis · vos · rogo · utei · recipiates · nomen Luxsia · Ạ(uli) Antesti · filia · caput · cor · co(n)silio(m) · valetudine(m) vita(m) · membra · omnia · accedat morbo · cot(t)idea · et sei · faciatis · votum · quod · faccio · solva(m) · vostris [·] meritis

Bibliography: Corell 1993; AE 1993, 1008; Castillo 1993: 308–09; HEp 5, 695; Maltomini 1995; AE 1995, 770; CILA II, 1249; Marina Sáez 1999; AE 1999, 894; Marina Sáez

2001: 52–53; HEp 9, 503; Stylow 2005: 255, 261; Díaz Ariño 2008: no. U29; Kropp 2008: no. 2.2.2/1; Versnel 2010: 295–97; Urbanová 2018: no. 52; Marco Simón 2019: 388. 209

Sylloge of  Defixiones from the Roman West Image source: Corell 1993: 262.

cotidea for cottidia (l. 3), vita for vitam (l. 3), sei for si (l. 4), faccio for facio (l. 4), solva for solvam (l. 4) and meritis for meretis (l. 4; see Marina Sáez 1999: 297). The dative phrase dis imferis (l. 1) is interesting, since syntactically we would expect a vocative (this is likely explained by analogy with fixed phrases like dis manibus). Also note the use of adjectival cot(t)idea (l. 3; Castillo and Marina Sáez 1999: 297), the expression nomen recipere (‘a standard term for a magistrate agreeing to hear a (criminal) case’, according to Marco Simón 2019: 388) and the lack of a preposition in the expression solva(m) vostris meritis (l. 4), which CILA (II, 1249) corrects as solva(m) pro vostris meretis or solva(m) vobis ob vostris meretis. This final phrase is especially noteworthy, since the defigens makes an offering (without specifying the details) to the deities in exchange for their assistance (ll. 3–4).

Commentary: the circumstances in which this defixio were discovered around 1990 are unknown. The curse’s editor has claimed that it was found nearby Carmona, while Civantes Moral (apud Styllow 2012) has insisted that it was found near ‘Los Chorrillos’ (Mengíbar), where the curse is currently kept in a private collection. The defixio was written on a thin strip of lead that is now broken into three pieces that fit together. The text contains four lines, which run from left to right and were written in capitals (note the e and f with two vertical strokes, the diagonal short strokes in l and a and o consisting of two strokes). The words are systematically separated with interpunctuation consisting of a single dot found halfway up the letters. We do not know the circumstances that led to the writing of this curse, which targets Luxsia, who is handed over to the infernal deities and ought to suffer a terrible illness. Given that the text lists various body parts, it is possible, though far from certain, that this curse belongs among the erotic defixiones (for possible parallels, see 18 and 51). The text presents the following orthographic and phonetic features: Imferis for Inferis (l. 1), utei for uti (l. 1), recipiates for recipiatis (l. 1), Luxsia for Luxia (l. 2), filia for filiae (l. 2), cosilio for consiliom (l. 2), valetudine for valetudinem (l. 2),

After being inscribed, the tablet was pierced in its four corners (for parallels, see 95 and 133, among others). Translation (Marco Simón 2019: 388, modified): ‘To the gods of the underworld: I ask that you consent to hear the case against Luxsia, daughter of Aulus Antestius. May illness afflict her head, heart, thoughts, health, vitality, her entire body day after day. And if you attend to the request that I make, I will reward you for your help.’

130. Aratispi, Villanueva del Cauche Provenance: unknown. Current Location: Patrimonio Histórico Andaluz. Date: 1st–2nd centuries CE. Material: lead. Measurements: 9 × 7.2 × 0.1 cm. Reading: (eight lines of charaktêres) hutos apoleson M(arcum) · Domitiu(m) Nigrum 5 Firmanae filium Bibliography: CIL II²/5, 729; HEp 8, 356; Schmidt 2004: 49; Gordon 2020b: 197–98.

charaktêres arranged into 8 lines found throughout the tablet’s surface. At the tablet’s bottom-right corner there are also five lines written in the Latin alphabet, which run from right to left. The text was written in old Roman cursive dated to the second half of the first century CE, a dating that is quite early for the use of charaktêres, which are usually found from the second century CE onwards.

Image source: see App. IV.3, SD 130. Commentary: this defixio was discovered in the surroundings of Casabermeja (Malaga) and therefore could come from Aratispi or another nearby place. The curse tablet was part of the so-called Fundación Alhornoz (on this, see the note in HEp 8: p. 7), which was a private collection kept in a country estate of the same name near Écija. Around the year 2000, the Fundación came under the control of the Patrimonio Histórico Andaluz.

We do not know the circumstances that led to the writing of this curse, in which lines 1–2 were written in the Greek language (οὕτος ἀπόλεσον, though transcribed with the Latin alphabet), perhaps as a way to reinforce the magical power of the text (for parallels, see the third part of 117 or DT 271 from Carthage, and I.4.2.1.2). Gordon (2020b: 198), for his part, has suggested that a professional

The defixio, which is intact, was written on a small rectangular lead sheet. The inscription contains 36 210

 Hispania first found here in Hispania (for examples, see e.g., 230, 360 and 454). Niger is a well documented cognomen in the Iberian Peninsula (see Abascal 1994: 439).

practitioner first drew the charaktêres and then told the defigens to include this expression. This theory suggests that there were two people involved in the manufacture of this curse. In anycase, this ‘coded’, the identification of the victim with a matronymic and the use of charaktêres are all conventional elements of aggressive magic, which are

Translation: ‘Destroy Marcus Domitius Niger, son of Firmana!’

131. Celti, Peñaflor Provenance: necropolis. Current Location: CLPC. Date: second half of the 1st century BCE. Material: lead. Measurements: 4 × 13 × 0.5 cm. Reading: Marcel(l)us · Valerius · mutus · tacitus · siet adversus · C(aio) · Licinio · Gallo · quadmodum rana · sene (!) · lingua muta · tacita · est · sic · Marcellus · mutus · tacitus · debilitatus · siet 5 advrsus · L[i]cinio · Gallo

Bibliography: Stylow 2012; AE 2012, 740; HEp 18, 438; Stylow 2014; Kruschwitz 2016.

Based on its content, this curse should be grouped among the juridical defixiones. This is made clear by the repeated desire that the victim be silenced: mutus tacitus siet (l. 1) and mutus tacitus debilitatus siet (l. 4). These phrases strongly reference the name of the ancient Roman deity Muta Tacita, who is invoked in two other Latin curses that appear to be juridical in nature (522 and 529; for a discussion, see Stylow 2012: 152). The attempt to silence the victim is reiterated in a similia similibus formula (ll. 2–5), in which an analogy is drawn between a mute, tongueless frog and the victim. The above-mentioned debilitatus adds to the picture, suggesting that the victim should be rendered incapable of testifying in court or of taking action against the victim. In the larger corpus of defixiones, there are other texts that use similia similibus formulae to draw a persuasive analogy between the cursed and an animal used in a magical ritual (such as the puppy in 160 or the supposed cock from Cartago in DT 222). That said, this is the first curse to mention a frog, even though this animal appears in different (and much later) magical practices described in PGM (such as XXXVI, 250–55: ‘just as this frog drips with blood and dries up, so also will the body of him, NN, whom NN bore’ [trans. R.F. Hock apud Betz 19922]). Kruschwitz, for his part, has proposed two new interpretations for the word rana, which could either be taken either as technical term alluding to an illness (Kruschwitz 2016: 163, n. 5; a hypothesis that picks up on debilitatus) or an allusion to a ritual described by Pliny (HN XXXII, 49) to get the truth out of a woman. Despite these new interpretations

Image source: J. Borrego, apud Stylow 2012: 150, fig. 2. Commentary: this defixio was discovered in c. 1979 in the area called El Camello situated to the west of ancient Celti (Peñaflor, Seville), where there was a Roman and early Christian necropolis. According to the editor’s report, the tablet was found buried some five metres away from two graves covered with tegulae. The curse was written on a rectangular lead sheet. Though it broke into three pieces during the excavation, the fragments all fit together, and the curse has been wholly preserved. It contains five lines, which were written in rustic capitals, which date to the second half of the first century BCE based on palaeographic analysis. The words are separated with interpunctuation consisting of a single dot and the letters measure between 0.3 and 0.4 cm in height. The tablet’s layout deserves mention, since the lines run from right to left and bottom to top (for parallels, see e.g., 134–36, 149 and 303). Despite this atypical layout, the letters are oriented normally (i.e., from left to right). The text presents the following orthographic and phonetic features: Marcelus for Marcellus (l. 1), advrsus for adversus (l. 5), quadmodum for quemadmodum (l. 2) and sene for sine (l. 3). The archaic form siet (ll. 1 and 4) is found up until the Augustan period, which matches the palaeographic dating of the text. 211

Sylloge of  Defixiones from the Roman West Translation (Marco Simón 2019: 389, slightly modified): ‘May Valerius Marcellus fall silent and unable to speak against Caius Licinius Gallus. Just as a frog without a tongue is mute and silent, so may Marcellus be mute, silent, enfeebled against Licinius Gallus.’

of rana, we cannot forget that the curse’s main objective was to silence its victim, a commonplace theme in juridical defixiones. After being inscribed, the tablet was rolled up.

132. Astigi, Écija Provenance: necropolis. Current Location: MHM Écija. Date: second half of the 1st century CE. [Inv. No.: P2/2005 (sigla excavación BE 01)] Material: lead. Measurements: 11.9 × 14.7 × 0.2 cm. Reading: gregs · An[t]oniani · veneta et russea · quadriga Lascivi Veri · quadriga Lascivi Vetii · qua[d]riga Margaritei · qua[d]riga Margaritei · quadriga Gelotis · quadriga Urbici · quadriga (H)ila5 ri · quadriga · (H)eleni · quadriga Basilisci · quadriga Nomantini · quadriga · Barbarionis · quariga Cal(l)idromi · quadriga Lụpi · agitatores · Piramus · agitator[e]ṣ · et · quadrigas · Antonia10 ni · Patricium Martialem Successum Atiarionem · Ṿaicus · Narcis(s)us · Atsertor tota grẹx Antonianị

Bibliography: Tinoco Muñoz 2002: 478; Sáez Fernández et al. 2004: 85; García-Dils de la Vega and de la Hoz Montoya 2013; AE 2013, 830; Gordon 2019a: 120–21; Gordon 2020a: 123–24.

was outside of the walled portion of the colonia and was part of a Roman necropolis. More specifically, the tablet was found in a cinerary burial in a ditch (no. 36/ U.E. 67, 145 × 130 cm, with a north–south orientation and dated between the final third of the first century BCE and the first half of the first century CE), in which there was a layer of organic material that was mixed with charred bones. On top of this layer, archaeologists found grave goods (a cup, plate and an as minted at the colony Iulia Traducta and dated between 33 and 27 BCE). According to the editors, the tablet was deposited in the tomb after it had

Image source: García-Dils de la Vega and de la Hoz 2013: 246, fig. 2. Courtesy of S. García Dils. See also App. IV.3, SD 132. Commentary: this defixio was discovered in 2001 during the excavation at Calle Bellidos, no. 18 in Écija. This area

212

 Hispania originally been sealed, but within the same century. This deposit, which is found in a necropolis near the circus, recalls the deposition of the agonistic curse tablets from Susa and Carthage.

found as a cognomen. The remaining six horse names are all attested in other inscriptions. After the list of quadrigae, the enumeration of victims continues and turns to the so-called agitatores (i.e., the drivers). That said, in this part of the text, the specific list of individuals is momentarily interrupted after the first agitator (Piramus), where the text has an all-inclusive phrase, in which all the drivers and chariots belonging to Antonianus are all cursed again (ll. 8–10). After this formula, the list of personal names continues with four cognomina in the accusative and three in the nominative before concluding with a final formula that encompasses all the defixio’s victims (l. 14). It must be stressed that of the nine names found in the second part of the curse, seven of them are used not only as anthroponyms but also as horse names. Their position, however, strongly suggests that they refer to the agitatores in this context. All the names are attested in Latin onomastics, though it is worth mentioning Vaicus (l. 12), which is an indigenous name otherwise found only in one inscription from Lusitania.

The defixio was written on a roughly triangular lead tablet, which appears to have been cut out from a larger sheet and then flattened. The inscription contains 14 lines, which run from left to right and were written in old Roman cursive (measuring between 0.47 and 0.68 cm in height). Several words are separated with interpunctuation consisting of a single dot found halfway up the letters. The curse is agonistic (currently the only such defixio from all of Hispania) and targets the grex Antoniani and two factiones circenses (the blues and reds). The text presents the following orthographic and phonetic features: gregs for grex (l. 1; cf. l. 13), Ilarus for Hilarus (ll. 4–5), Eleni for Heleni (l. 5), Nomantinus for Numantinus (l. 6), Calidromi for Callidromi (l. 7) and Piramus for Pyramus (l. 8). The text, which does not have any verbs, consists of a simple list of the victims’ names found in both the nominative and the accusative (see section I.7.2 for a discussion). The list begins with the phrase gregs Antoniani (l. 1), which is a reference to the factions supported by Antonianus. Next, we find a list of the 13 quadrigae, the curse’s targets, each of which is followed by a name in the genitive. As happens in circus mosaics, each chariot is identified with the name of its most important horse (or funalis). Within the list, we find quadriga Margaritei (ll. 2–3) repeated, which is perhaps an accidental repetition on the author’s part. Among the horses’ names, four are variations of other nomina (Lascivus for Lascovus, Margariteus for Margarita, Urbicus for Urbanus and Barbario for Barbarus, in ll. 2–4 and 6–7), while Nomantinus (in l. 6) is first attested here as a horse name, though it is elsewhere

After being inscribed, the tablet was folded in half and closed with an iron pin. Translation (Gordon 2020a: 123, slightly modified): ‘Stable of Antonianus! Blues and Reds! Quadriga of Verus Lascivus! Quadriga of Vettius Lascivus! Quadriga of Margariteus! Quadriga of Margariteus! Quadriga of Gelos! Quadriga of Urbicus! Quadriga of (H)ilarus! Quadriga of (H)elenus! Quadriga of Basiliscus! Quadriga of Nomantinus! Quadriga of Barbarion! Quadriga of Callidromus! Quadriga of Lupus! Drivers: Piramus! (I curse:) drivers and quadrigas (accus.) of Antonianus: Patricius, Martialis, Successus, ‘Attiaro’. (nom:) Vaicus, Narcissus, Adsertor, the entire stable of Antonianus!’

133. El Portal Provenance: unknown. Current Location: unknown. Date: first half of the 1st century BCE. Material: lead. Measurements: (10) × (14.5) × 0.3 cm. Reading: Col. I Antronice Frontaca Pusellio Cimisex 5 Felix Rustica Storge Karis Philonia 10  Mena Helene [------]

Col. II +++ Polio N[i]colave Stabilio Diocare [T]alame Sum[mo]ni [---]ASIA [---]ARIS So[ri?]ce G[-c.4-] Princeps (vacat)

213

Col. III [---]TANA Anus Rustica Optate Aucta Fabulla Iulia Celido Prime Traxe Dio

Sylloge of  Defixiones from the Roman West Bibliography: González Fernández 2015; AE 2015, 587.

The text is written in capitals that are ‘mirrored’ (i.e., written from right to left). Due to its palaeographic features as well as some phonetic features (like the transcription of Greek aspirates and the use of k for c), González Fernández has dated the curse tablet between the second half of the second century BCE and the beginning of the first century BCE. In my view, the palaeographic features of the text, which are similar to those of 129 and 131) suggest that the curse tablet should be dated to the first century BCE. The circumstances that led to the writing of this curse are unknown, since it consists of a list of 33 Greek and Latin personal names. Of these names, 22 are undoubtedly feminine and 5 masculine. The fact that the names here typically belong to slaves opens up the possibility that the curse’s victims were enslaved labourers on one of the region’s farms (for a parallel, see 104).

Commentary: this defixio was discovered in the 1980s under unknown circumstances on the grounds of the Frías agricultural estate, which is located between the villages El Portal and Puerto Real. The curse was written on a lead sheet, which was originally rectangular. The tablet has been broken along its upper edge and has lost the whole top-left corner, as well as part of the lower edge. On its surface there are three big cracks and many little ones. The text is arranged in three columns of 11–12 lines. According to González Fernández, the first column is the one on the left-hand side, and the lines throughout the text were written from right to left starting at the bottom of the tablet (for a similar layout, see 131, 134–36 and 149). Nevertheless, it seems more likely that the text was written from right to left and from top to bottom, starting by the right column (contra González Fernández). In favour of this hypothesis there are several factors. First, the top line of the right column (here, column I, which is the only column whose top line is completely intact), is perfectly aligned with the upper edge of the tablet. This contrasts with the rest of the lines from the column (I, ll. 3–11) in which the letters progressively float up to the left. Furthermore, the lines in the upper portion of the tablet are generally straighter than those from the bottom portion. Additionally, there is an empty space left at the bottom of the central column. These observations suggest that the lines were written from top to bottom and not bottom up (contra González Fernández). Furthermore, the right column is aligned with the right edge of the tablet. This suggests that the right column was actually written first, followed by the central and left columns, respectively (contra González Fernández). In short, to correctly interpret the curse, the reading ought begin with the top of the right column and end with the bottom of the left column. In other words, González Fernández’s organization of the lines is backwards.

Within the list, several Greek names are attested for the first time in the Iberian Peninsula: N[i]colave (II, l. 3), Diocare (II, l. 5; for both of the first two names, the masculine form is elsewhere attested in Hispania, but the feminine version is first attested here; see Solin 2003: 45, 46), Antronice (col. I, l. 1; despite the editor’s suggestion that this name is derived from anthropos, it is better to read it as a form of the wellattested Andronice), Cimisex (I, l. 4) Storge (I, l. 7; see Solin 2003: 1344) and Philonia (I, l. 9; probably the feminine form of Philon, which is attested in Hispania; see Solin 2003: 800). The following Latin names are attested here for the first time in the Iberian Peninsula: Frontaca (I, l. 2, there are no parallels for this name, which is the feminine form of the well-known Frontaccus) and Pusellio (I, l. 3; without parallels, unless it is a mistake for Pusillio [cf. OPEL III, 172]?). It is worth noting the personal name Sum[mo]ni (II, l. 7), which the editor has connected to the indigenous name Summoi (see González Fernández: 2015: 111 and n. 33). After being inscribed, it appears that the tablet was nailed up (perhaps inside a tomb? See 95 and 129 for parallels), since there are three holes in the lower corners as well as the top left one.

Hispania Citerior 134. Emporion, Ampurias Provenance: necropolis. Current Location: MAC Barcelona. Date: 75–85 CE. [Inv. No.: 28032] Material: lead. Measurements: 3.2 × 5.1 × 0.2 cm. Reading: A Fulvus legatus Augusti Rufus legatus Augusti Maturus proqurator Augusti 5 legati atvocati Ind[i-] cetanoruṃ



214

 Hispania B consilium Fulvi legati Olossitani Campanus Fidentinus Augus5 (ti) [---]

Bibliography: Almagro Basch 1947: 125; Gómez Moreno 1949: 332–35; HAEp 1950–52: 442–43; Almagro Basch 1952: 163–68; AE 1952, 122; Almagro Basch 1955: 23– 24, 60; AE 1955, 222; Syme 1958: 7; Lamboglia 1959: 147–72; Pflaum 1960: 95–98; Balil 1964: 198; García Ruiz 1967: no. 55; Solin 1968: no. 26; Alföldy 1969: 19– 21; ILER 5919; Ripoll-Perelló 1980: 413–16; IRC III, 172; HEp 4, 446a; Gager 1992: no. 52; Plana and Pena 1995– 96: 99–101; IRC V ad IRC, III: 172; Pi Vázquez 2003; Pi Vázquez 2005; AE 2005, 882a–b; Wilburn 2013: 219–53, no. 1; Kropp 2008: no. 2.1.1/3; Marco Simón 2010a: no. 1; Urbanová 2018: no. 48.

Its letters are ‘mirrored’ (i.e., written from right to left), with the exception of c, d, g and q. The letters measured between 0.2 and 0.5 cm in height. While we do not know the details that led to the writing of this curse, most scholars (Pflaum, Balil, Solin, Ripoll Perelló and Gager) believe that these curses give evidence of a legal conflict between the Indicetani and Olossitani with the latter being the defigentes. That said, several scholars (Almagro (1951: 59; Fabré et al. 1991: 162; Marco Simón 2010a: 407, Urbanová 2018: 257) have correctly pointed out that all the names and groups mentioned in the text are in the nominative case. Therefore, it is more likely that all those mentioned are the victims of the curse, which belongs to the group of juridical ones.

Image source: Almagro Basch 1947: Plate 85, 2A and 2B. Commentary: in 1994, this defixio was discovered together with 135 and 136 at the necropolis of Ballesta inside of a mausoleum containing a total of eight cremation graves. The grave goods consisted of perfume jars of glass or ceramic, some pottery as well as some bronze and iron nails and two coins, all of which point to a date of 25 BCE (see Almagro 1955: 23). Three curses were found inside of three urns (nos 21, 22 and 23), which were still filled with the cremation remains (Almagro 1947: 124). This important detail makes it clear that the urns were not only used for a funerary rite, but also for magical practices. They served as a deposit for the curses, since they offered a means of ensuring that communication with the infernal deities would be successful thanks to assistance of the spirits of the deceased (contra Lamboglia 1959: 147, n. 1 and 155, who thought that these urns were used exclusively as a place to hide and deposit the tablets).

Based on the comparison with an honorific inscription from Aquae Flaviae in which the same ranks are mentioned as in the present curse (see AE 1952), Pflaum (1960) has identified the Roman officials T. Aurelius Fulvus, Rufus and Maturus in the present curse. The first was the governor of Hispania Tarraconensis and had the title of legatus Augusti pro praetore. Pflaum (AE 1952, 122) has identified this individual as the T. Aurelius Fulvius who was the grandfather of Antoninus Pius and cos. II 85 CE (PIR2 A 1509 = RE Aurelius no. 136). The Rufus, who is mentioned in the curse as a legatus Augusti, was in reality a legatus iuridicus (see Syme 1958: 7; Alföldy 1969: 21; Fabré, Mayer and Rodà (IRC); Pi Vázquez 2005: 170; Marco Simón 2010a: 403–04). As Alföldy (1969: 21) has argued, this legatus should be identified as Q. Pomponius P. f. Rufus, cos. suff in 95 (PIR² P749 = RE Pomponius no. 68), who was in Hispania Tarraconensis between either 74 or 75 and 78 CE. Pflaum has identified the procurator Augusti Maturus as ‘Marius Maturus, le seul procurateur à avoir porté ce cognomen très rare, qui était en 69 procurator Alpium maritimarum (Tac. Hist. II, 12; III, 42–43)’ (AE 1952, 122).

The present curse was discovered in cinerary urn no. 21, which did not have any grave goods (see Almagro Basch 1955: 60). The curse was written on an irregularly shaped lead sheet, which is in rather poor condition. As I was able to verify during an autopsy in March 2010, side A is corroded in several distinct spots, whereas side B is corroded throughout to the extent that several inscribed areas have been completely lost. An opisthograph, the inscription contains 11 lines of capitals, which run from right to left and from bottom to top (contra Urbanová 2018: 256; for this unusual layout, see 131, 135, 136, 149 and 303). The text has been dated to the last quarter of the first century CE based on palaeographic grounds.

Translation (Marco Simón 2010a: 400): ‘Fulvus, the legate of the emperor, Rufus, the legate of the emperor, Maturus, the procurator of the emperor, the legate’s (consilium) and the defenders of the Indicetani. The council of the legate Fulvus, the Olossitani legates, Campanus Fidentinus Aug[usti].’ 215

Sylloge of  Defixiones from the Roman West 135. Emporion, Ampurias Provenance: necropolis. Current Location: MAC Barcelona. Date: 75–85 CE. [Inv. No.: 28036] Material: lead. Measurements: 4.6 × 5.3 × 0.1 cm. Reading: A [Ful]vus legatus [Aug]usti Rufus legạ[tus Aug]usti Matu[ru]s procurator 5 A[ug]ụsti consilium legati atvocaṭi [Indice-] t[ano-] rụ[m] B [Oloss]itani Sempronius Campanus Fidentinus atver5 {ver}sari mẹi inique ne int[-c.2-]sịṇṭ

Bibliography: Almagro Basch 1947: 125; Gómez Moreno 1949: 332–35; HAEp 1950–52: 442–43; AE 1952, 122; Almagro Basch 1952: 163–68; Almagro Basch 1955: 24– 25 and 62; AE 1955, 222; Syme 1958: 7; Lamboglia 1959: 147–72; Pflaum 1960: 95–98; Balil 1964: 198; García Ruiz 1967: no. 57; Solin 1968: no. 28; Alföldy 1969: 19–21; ILER 5917; Ripoll-Perelló 1980: 413–16; IRC III, 174; Gager 1992: no. 52; HEp 4, 446c; Plana and Pena 1995–96: 99–101; IRC V ad IRC III, 174; Comes and Rodá 2002: 239–40; Pi Vázquez 2003; Pi Vázquez 2005; AE 2005, 883 a–b; Wilburn 2013: 219–53, no. 3; Kropp 2008: no. 2.1.1/4; Marco Simón 2010a: no. 3; Urbanová 2018: no. 49.

The present curse was written on a small irregularly shaped lead sheet. As I was able to verify during an autopsy in March 2010, the tablet is quite corroded on side A. An opisthograph, the inscription contains 16 lines, which run from right to left and bottom to top (for a similar layout, see 131, 134, 136, 149 and 303). The text was written in capitals that measure between 0.2 and 0.5 cm in height and have been dated to the first century CE on palaeographic grounds. The letters are ‘mirrored’ (i.e., written from right to left), with the exception of c, d and q. As was the case with the previous curse, this tablet targets the Indicetani, Olossitani and local Roman officials (see the commentary for 134), all of whom are called atve{ve}rsari mẹi (B, ll. 4–6). Based on this and the lexeme atvocati (A, ll. 6–7, for advocati), the curse can be grouped among the juridical texts.

Image source: Almagro Basch 1947: Plate 85, 3A and 3B. Commentary: in 1994, this defixio was discovered together with 134 and 136 at the necropolis of Ballesta inside of a mausoleum dated to c. 25 BCE (for the archaeological context, see the commentary for 134). This item was found inside of cinerary urn no. 22, which also contained the remains of two bronze nails which measure 3.8 cm in length and a rectangular bronze plaque (measuring 1.2 × 0.4 cm)

Translation (Marco Simón 2010a: 401–02, modified): ‘Fulvus, legate of the emperor, Rufus, legate of the emperor, Maturus, procurator of the emperor, the legate’s council, the legal representatives of the Indicetani. May the Olossitani, Sempronius Campanus Fidentinus, my adversaries not (...) me wickedly.’

216

 Hispania 136. Emporion, Ampurias Provenance: necropolis. Current Location: MAC Barcelona. Date: 75–85 CE. [Inv. No.: 28033] Material: lead. Measurements: 3.2 × 5.7 × 0.1 cm. Reading: A Olossita[ni] Titus Aurelius Fulvus legatus Augusti 5  Rufus legatus Augus[ti] B Maturus proqurator Augusti consilium : legati ḷegati Indiceta5 norum Indiceta[norum]

Bibliography: Almagro Basch 1947: 124; Gómez Moreno 1949: 332–35; HAEp 1950-1952: 442–43; AE 1952, 122; Almagro Basch 1952: 163–68; Almagro Basch 1955: 24; AE 1955, 222; Syme 1958; Lamboglia 1959: 147–72; Pflaum 1960: 95–98; Balil 1964: 198; García Ruiz 1967: no. 56; Solin 1968: no. 27; Alföldy 1969: 19–21; ILER 5918; Ripoll-Perelló 1980: 413–16; IRC III, 173; HEp 4, 446b; Gager 1992: no. 52; Plana and Pena 1995–96: 99– 101; IRC V ad IRC III, 173; Pi Vázquez 2003; Pi Vázquez 2005; AE 2005, 881 a–b; Wilburn 2013: 219–53, no. 2; Ribeiro 2006: 250–51; Kropp 2008: no. 2.1.1/2; Marco Simón 2010a: no. 2; Urbanová 2018: no. 47.

The present curse was written on an irregularly shaped lead sheet that today is in good condition, as I was able to verify during an autopsy in March 2010. An opisthograph, the inscription contains 12 lines, which run from right to left and bottom to top (for a similar layout, see 131, 134, 135, 149 and 303). The text is in capitals that measure between 0.2 and 0.5 cm in height. Its letters are ‘mirrored’ (i.e., written from right to left), with the exception of c, d, g and q. Here it is worth noting the repetition of the word Indicetanorum (B, l. 6), probably the result of a copying error. Like the two previous tablets, the present curse targets the Indicetani, Olossitani and local Roman officials for unknown reasons, although the use of certain lexemes suggests that the text can be grouped among the juridical ones (see the commentary for 134).

Image source: Almagro Basch 1947: Plate 85, 1A and 1B. Commentary: in 1994, this defixio was discovered together with 134 and 135 at the necropolis of Ballesta inside of a mausoleum dated to c. 25 BCE (for the archaeological context, see the commentary for 134). This curse was found inside cinerary urn no. 23, which contained the burnt remains of the deceased and the remains of four nails measuring 4.5 cm in length (Almagro Basch 1955: 61).

Translation (Marco Simón 2010a: 400–01): ‘The Olossitani, Titus Aurelius Fulvus, legate of the emperor, Rufus, legate of the emperor. Maturus, procurator of the emperor, the legate’s council, the legates of the Indicetani, of the Indicetani.’

217

Sylloge of  Defixiones from the Roman West 137. Emporion, Ampurias Provenance: aquatic context (?) Current Location: MAC Ampurias. Date: end of the 1st century BCE. [Inv. No.: 2670] Material: lead. Measurements: 12.5 × 8.9 × 0.1 cm. Reading: Ṿeraṇio Pupilius · Stabilio Ạpọḷinidorus · Philargurus · Scapi 5 Surisca · Alexaẹ Papus Amphio · Parnaci[s] Zodiana · omnes quei 10 inimeici Senecae

Bibliography: Almagro Basch 1947: 123; Gómez Moreno 1949: 331–32; HAEp 1950–52: 441; Almagro Basch 1952: 161–62; Mallon 1952: 172–73; García Ruiz 1967: no. 54; Solin 1968: no. 25; ILER 5916 and 6846; IRC III, 175; HEp 4, 447; Curbera 1996; Marina Sáez 2001: 55–56; IRC V ad IRC III, 175; Solin 2004: 126– 27; AE 2004, 834; Ribeiro 2006: 248–49; Díaz Ariño 2008: no. C87; Kropp 2008: no. 2.1.1/1; Urbanová 2018: no. 46.

corner. The top-left corner had already been broken by the time that I conducted an autopsy in April 2011. The inscription contains 10 lines, which run from left to right and were written in capitals (measuring between 0.3 and 0.7 cm in height). The text has been dated to the end of the Republican period on palaeographic grounds (for a dissenting view, see Gómez Moreno, who dates the tablet to the first century CE). Based on its content and the use of the word inimeici (l. 10), this curse can be identified as a juridical defixio. The text presents the following orthographic and phonetic features: inimeici for inimici and quei for qui (ll. 8–9; see Marina Sáez 2001: 56). The curse’s final phrase, omnes quei inimieici Senecae (ll. 9–10), can be compared to the all-inclusive formulae common in British curse tablets (e.g., 342, l. 4: et ṣị qui ạfuerẹ). In l. 2, Cubrera (1996: 293) reads Pupilius Stabilio A- (contra IRC). The final Acan then be read with l. 3 to give the theophoric name A-/ polinidorus. The text also contains a series of onomastic formulae with a name in the nominative followed by a

Image source: Almagro Basch 1947: 161, no. 113. Commentary: this defixio was discovered in Emporion in 1946. The archaeological context, nevertheless, has been discussed: according to the author of the editio princeps (Almagro Basch 1947), it was found by chance by a treasure hunter, while for the editors of IRC (and so for Curbera 1996: 292, Díaz Ariño 2008: 172 and Kropp 2008 a: 2.1.1/1), the tablet was found at the beach in Emporion. No matter what, the curse was written on a roughly rectangular lead sheet missing its lower-left 218

 Hispania name in the genitive (perhaps a list of slaves and their masters): Philargurus Scapi (l. 4; according to Abascal 1994: 498, the second name is Latin), Surisca Alexaẹ (l. 5) and Amphio Parnaci[s] (l. 7; P(h)arnacis is here attested for the first time in the Iberian Peninsula). There are also some uncommon Latin cognomina in the curse: Veranius (l. 1; see OPEL IV, 156), Papus (l. 6, see OPEL III, 124)

and Zodiana (ll. 8–9; without known parallels; contra Curbera 1993: 294). Translation: ‘Veranio, Pupilius Stabilio, Apolinidorus, Philargurus slave(s) of Scapus, Surisca slave of Alexa, Papus, Amphio slave(s) of P(h)arnaces, Zodiana, all the enemies of Seneca.’

138. Emporion, Ampurias Provenance: unknown. Current Location: MAC Ampurias. Date: 1st–2nd centuries CE. [Inv. No.: 4474] Material: lead. Measurements: (5.8) × (9.5) × 0.1 cm. Reading: -----[---] ei [---] [---] c ̣um ese pelven[i]unt eos [---] cum qui mi facinus inposuit 5 [---] cossit paupertati meam [---] hodie podui me inopia fuit [---] c ̣um putet eo modo facio tibi [---] parturiens P[-c.2-]donis turnavit

Bibliography: Almagro Basch 1952: 168; HAEp 1950– 952: 444; García Ruiz 1967: no. 73; Solin 1968: no. 29; ILER 5920; IRC III, 176; Comes and Rodà 2002: no. 78; IRC V ad IRC III, 176; Kropp 2008: no. 2.1.1/5; Urbanová 2018: no. 214.

written in capitals (measure between 0.3 and 0.5 cm in height). The curse has been dated between the first and second century CE on palaeographic grounds. Though the reading is secure, it has defied interpretation due to the high quantity of ‘Vulgarisms’ and orthographic problems.

Image source: Almagro Basch 1952: 169, no. 117.

That said, Fabré, Mayer and Rodà (IRC) have argued that the defigens sought to dedicate to the infernal deities an individual who had made an accusation against the defigens and ruined him economically. As revenge, the defigens hopes to seduce his victim’s wife with presents. The editors have also proposed the following reconstructions: perveni[u]nt (l. 3), fa(s)cinus (l. 4) and po˹t˺ui (l. 6).

Commentary: in October 1951, this defixio was discovered on the street that connects the Museo Monográfico and the ancient port. It was written on a lead sheet, which originally must have been roughly rectangular. As I was able to confirm during an autopsy in April 2011, a portion of the right and bottom edges are intact. The inscription contains nine lines, which run from left to right and were

After being inscribed, the tablet was rolled up.

139. Emporion, Ampurias Provenance: domestic space. Current Location: MAC Ampurias. Date: end of the 1st century BCE–1st century CE. [Inv. No.: 4473] Material: lead. Measurements: (2.4) × (5) × 0.2–0.1 cm. Reading: A -----[---]ḷịṭis nequ[e---] [--- a]beant · PLV[---] [---]VIT · TVCMV [---] 5 [---]COVGNA or COVCINA· [---] ------

B -----[---] A or N or M [---] [---]Min(icius?) Feli[x ---] [---]Itali[cus? ---] ------

219

Sylloge of  Defixiones from the Roman West Bibliography: Almagro Basch 1952: 169; HAEp 1950–52: 445; García Ruiz 1967: no. 74; Solin 1968: no. 30; IRC III, 177; HEp 4, 448; Kropp 2008: no. 2.1.1/6.

which have been dated to the end of the Republic or the first century CE. We do not know the circumstances that led to the writing of this curse. Fabré et al. (apud IRC) have reconstructed the phrase “nequ[e---a]beant Plu[tone---]” (A, ll. 2–3). Even though this hypothesis is very suggestive, the fragmentary nature of the inscription precludes us form accepting the reconstruction.

Commentary: this fragment of a curse tablet was discovered in 1952 within the area surrounding house no. 2 in Emporion. It was written on an irregularly shaped lead sheet, which today is in good condition, as I was able to confirm during an autopsy in April 2011. An opisthograph, the inscription contains seven lines, which run from left to right on side A and from right to left on side B. The text is in capitals, measuring between 0.2 and 0.7 cm in height,

After being inscribed, the tablet was cut.

140. Saguntum, Sagunto Provenance: unknown. Current Location: MA Sagunto. Date: 70 CE. Material: lead. Measurements: 6.3 × 23 × 0.1 cm. Reading:

Quintula cum Fortunali sit semel et numquam

Bibliography: Corell 1994: 280–82; AE 1994, 1072; HEp 5, 822; Corell, Gómez et al. 1999: no. 6; IRPV 16; Ribeiro 2006: 252; AE 2007, 156; Corell 2009: 68; Museros Ortiz 2007: 989; AE 2007, 156; Kropp 2008: no. 2.1.3/1; CIL II²/14, 757a; Urbanová 2018: no. 51.

background and says nothing about the shape that the lead tablet should take. The inscription contains two lines, which run from left to right and were written in old Roman cursive. The letters measure between 0.4 and 2.2 cm in height and have been dated on palaeographic grounds around 70 CE. Based on its content, this curse can be considered the first (and currently the only) erotic curse from the Iberian Peninsula. This interpretation is based on the use of the euphemistic phrase esse cum, meaning ‘to be in a relationship with’ (see Lewis and Short, s.v. sum, I A, 3, γ). The phrase semel et nunquam (ll. 1–2) is echoed by Pliny’s semel et unquam (HN II, 100). Both Fortunalis and Quintula are Latin cognomina that are rarely attested in Hispania (see Abascal 1994: 371 and 478).

Image source: Corell 1991: 281, fig. 1. Commentary: according to this curse’s previous owners (Ripollés and Adelantado), this defixio was discovered c. 1980 on the southern slope of the Castell hill, between the plaza de Armes and the cemetery. As I was able to confirm during an autopsy in April 2010, the tablet is in good condition. There are a few concretions and a thin whitish patina on its surface, which has protected the inscription.

After being inscribed, the tablet was folded in half and then a coin was slid inside (the coin dates the the reign of Vespasian and its legend reads Iudaea capta). Next, it was folded in half again. The insertion of a coin in a defixio is an unusual practice that finds few parallels (see the lead sheet with a coin from Britannia discussed in the introduction to section II.4).

The curse was written on a lead sheet that has the shape of a footprint, something that is unparalleled in the corpus of curses from the Roman West. Corell has connected the curse’s shape to a formula for subjugating a victim that is found in PGM VII, 925–29. This recipe recommends that after inscribing a curse, the defigens should place it under his/her left foot so that the spell can take effect. While the comparison is undoubtedly interesting, the papyrus recipe is much later than the curse under discussion, comes from a very different

Translation: ‘May Quintula never be in a relationship with Fortunalis again.’ 220

 Hispania 141. Saguntum, Sagunto Provenance: unknown. Current Location: MA Sagunto. Date: 2nd century CE. Material: lead. Measurements: 4.4 × 17.5 × 0.1 cm. Reading: quisquis (?) tunica(m) tol(l)it (?) {II} Livia obi eam vel eum (?) ite(m)+ qui{s} questo {h}habeat trata

Bibliography: Corell 1994: 282–85; AE 1994, 1073; HEp 5, 823; IRPV 15; Corell, Gómez et al. 1999: no. 7; Marina Sáez 2001: 102–03; HEp 12, 469; Kropp 2008: no. 2.1.3/2; Corell 2009: 68; Tomlin 2010: 268–69; CIL II²/14, 757b; AE 2010, 108; Urbanová 2018: no. 215.

(measuring between 0.3 and 1 cm in height). The text has been dated to the second century CE based on palaeographic grounds (for a discussion, see Corell 2002: 74). Since the defigens denounces the theft of a tunic, this curse should be grouped among the defixiones in fures. The text presents the following orthographic and phonetic features: tolit for tollit (l. 1), tunica for tunicam (l. 1), ite for item (l. 2), questo for quaestu(m) (l. 2), trata for tracta (l. 3; see Marina Sáez’s commentary on this point). The verb tollo is commonly found in British defixiones in fures (e.g., 337 and 367).

Image source: Corell 1994: 283, fig. 2. Commentary: according to this curse’s previous owners (Ripollés and Adelantado), this defixio was discovered c. 1980 on the ground on the Castell hill. The curse, which was inscribed on a lead strip, is well preserved, despite having been broken along the lower-left corner. The damaged area along the centre of the tablet (between the words tolit and Lidia) has affected the two first lines.

After being inscribed, the tablet was likely rolled up. Translation (Urbanová 2018: 259): ‘Whoever has stolen the tunic from Livia, pursue her or him, as well as the one who could profit from it, catch (him/her).’

The inscription contains three lines, which run from left to right and were written in old Roman cursive

142. Saguntum, Sagunto Provenance: unknown. Current Location: MA Sagunto. Date: second half of the 1st century–beginning of the 2nd century CE. Material: lead. Measurements: 11 × 19 × 0.5 cm. Reading: Felicio Aure[lian]i rogat et mandat pequnia(m) quae a me accepit Heracla conservus meus ut inst{t}etur (h)uius senus o[c]elus et 5 [v]ires quicumqui sunt ari demando pequniam (h)onori sacricola(m) Bibliography: Corell 2000: 241–47; AE 2000, 795; IRPV 13; HEp 10, 622; Ribeiro 2006: 241; Kropp 2008: no. 2.1.3/3; Corell 2009: 68; Tomlin 2010: 264–68; Versnel 2010: 292– 95; HEp 19, 394; AE 2010, 108; Urbanová 2018: no. 216.

information about the archaeological context for this defixio. The curse was written on an irregularly shaped lead sheet, which is in good condition, though, as I was able to confirm during an autopsy in April 2010, there is a whitish patina on the tablet’s surface as well as several corroded areas along the bottom edge. The inscription contains seven lines, which run from left to right and were written in old Roman cursive (measuring between 0.3 and

Commentary: according to this curse’s previous owners (Ripollés and Adelantado), this defixio was discovered c. 1980 on the ground on the Castell hill. There is no further 221

Sylloge of  Defixiones from the Roman West 1.1 cm in height). Based on palaeographic grounds, the text has been dated between the second half of the first century CE and the second century CE.

new reading fits well with rogat et mandat (l. 2; see Tomlin 2010: 264 and 266 contra Corell). The phrase demando pequniam (ll. 5–6; contra Corell) reiterates the initial mandat pequnia(m) (l. 2), both of which are used to dedicate the stolen money to the deity’s cult. It is worth noting the change from the third to the first person. The use of the third person could suggest that someone was writing the curse on behalf of the defigens, as is the case in 350 (whose first line reads: nomine Camụlorigi(s) et Titocun(a)e molam quam perdederunt...). That said, this would not explain the later change to the first-person singular verb.

Based on its content, this curse should be grouped among the defixiones in fures. In the text, the theft of a sum of money is denounced. The defigens first seeks to guarantee divine intervention by setting aside some money for the deity’s cult, before turning to curse various parts of the thief’s body. The text presents the following orthographic and phonetic features: pequnia for pecuniam (ll. 2 and 6), quae for quam (l. 2), insttetur for instetur (l. 4), uius for huius (l. 4), senus for sinus (l. 4), o[c]elus for oculus (l. 4) and quicumqui for quaecumque (l. 5).

After being inscribed, the tablet was folded four times. Translation (Tomlin 2010: 267): ‘Felicio the slave of Aurelianus (?) asks and entrusts the money which Heracla my fellow-slave received from me, that his bosom (?) be attacked, his eye (?) and strength (?), whoever they are […] I entrust the money (?) to the honour of the priest.’

The curse opens with the defigens’ name, Felicio Aure[lia-] [n]i. To properly read the name the first two strokes in Corell’s drawing must be ignored (I could find no trace of these strokes during my autopsy in April 2010). This

143. Saguntum, Sagunto Provenance: unknown. Current Location: private collection. Date: end of the 1st century–beginning of the 2nd century CE. Material: lead. Measurements: (7.5) × 16.5 × 0.2 cm. Reading: A Ịạọ (vacat) rogat ụṭi man{u}datum q(a)ẹ[que m]enra a tibi commendo aṇị[ma] vụl[tus] venter Eterionis Auḷịạ[ni?] fịḷị ịṇ omni ira da dis ịṇfẹ̣ 5 [ris] rọg[at ̣ ora?]t et querit ut pẹcụs [-c.12-]QVRES qụ[o]d [-c.5-] [-c.13-]ịṭạ [-c.9-]

B Fụlvilla rogat quod os manus digitus aures vox [------] [------] [------] [---]ITEVS[---] [------]

Bibliography: Corell 2000: 241; IRPV 14; HEp 10, 623.

seems somehow illogical (HEp 10, 623). Furthermore, the invocation of Iao is unparalleled in the corpus of Latin curse tablets. At this point in the text, it would be more normal to find the name of the defigens (though see Fulvilla rogat, B, l. 1) instead of this strange theonym. At a future date, then, the reading should be confirmed during an autopsy.

Commentary: according to this curse’s owners (Ripollés and Adelantado), this defixio was discovered c. 1980 on the Castell hill. We do not know the exact archaeological context in which it was found. The curse was written on an irregular lead sheet, which has lost its bottom-left corner and part of its bottom edge. The curse’s reverse is in quite bad condition and hardly legible. An opisthograph, the inscription contains 14 lines, which run from left to right and were written in capitals. The letters measure between 0.3 and 1 cm in height. Based on its palaeographic characteristics, the text has been dated to the end of the first or beginning of the second century CE.

After the (apparent) invocation of Iao, we find the figurative dismemberment of the first victim, whose aṇị[ma] vụl[tus] venter are all cursed (‘soul, appearance, belly’ A, l. 3; for anima, see 211 and 242; for vultus and venter, see 56). Eterio, the victim, is identified with a patronymic and given over to the infernal deities (A, ll. 4–5, dis ịṇfẹ[ris]), ̣ who are invoked with a tricolon of verbs (rọg[at, ora?]t ̣ et querit, A. l. 5; for a parallel, see 120). Side B contains the name of the defigens, Fulvilla (B, l. 1; a cognomen first attested here in Hispania) and an anatomical sequence: os, manu, digitus, aures, vox (‘mouth, hands, fingers, ears, voice’ B, l. 2; for os, see 51; for manus/digitus, see 56; for aures see 14, where the diminutive oriculas [for auriculas] is attested. There are no known parallels for vox).

We do not know the circumstances that led to the writing of this curse, in which two victims are symbolically dismembered and handed over to the deities invoked. The text presents the following orthographic and phonetic features: manudatum for mandatum (A, l. 1), queque menra for quaeque membra (A, l. 2), querit for quaesit (A, l. 5). Since I have not had the chance to conduct an autopsy, I give Corell’s reading here. However, Castillo is correct to point out that the inscription

After being inscribed, the tablet was folded six times. 222

 Hispania 144. Saguntum, Sagunto Provenance: unknown. Current Location: private collection. Date: 1st century CE. Material: lead. Measurements: (1.7) × (1.8) × 0.15 cm. Reading: [---] Tropae (!) [---] Argus · Pạ[---] [---Sym?]peruṣạ ------? Bibliography: Corell 2000: 241; IRPV 17; HEp 10, 624.

The curse was written on a roughly square lead sheet, which has lost its bottom-right corner. It was pierced at its top-left corner and has been damaged along its bottom edge. The inscription contains four lines, which run from left to right and it constists of a list of personal names in the nominative case. Of these names, Tropae (for Trope, l. 1) is here first attested for the first time in Hispania. Perusa (l. 3) could be for Pherusa or Sympherusa, which is a common personal name in Hispania (see Abascal 1994: 519). Given that the names are Greek, the victims were likely to have been slaves.

Commentary: according to this curse’s owners (Ripollés and Adelantado), this defixio was discovered in 1998 outside of its original archaeological context on the southern slope of the Castell hill (in the same area that 140 was found). Due to its material and shape as well as the presence of a hole, Corell has suggested that this is, indeed, a defixio.

145. Barchín del Hoyo Provenance: domestic space. Current Location: MA Cuenca. Date: 1st century BCE–1st century CE. [Inv. No.: AA87/1/1] Material: lead. Measurements: Ø 5.9–6.2 cm. Reading: A ὑπὲρ ἐμοῦ κα[ὶ] ὑπὲρ τῶν ἐμῶν τοῖς κατὰ ῞ᾼδην δίδωμι, παραδίδωμι Νεικίαν καὶ Τειμὴν καὶ τοὺς ἄ[λ]λους οἷς δικαίως κατηρασάμην B pro me pro meis devotos defixos inferis devotos defixos inferis Timen et Niciam et ceteros quoṣ merito devovi supr[a pro] me pro mei[s] Timeṇ Nician Ṇịciạ[n]

Image source: Curbera, Sierra Delage and Velázquez 1999: 280, fig. 1. See also App. IV.3, SD 145.

Bibliography: Velázquez and Curbera 1998; Curbera, Sierra Delage and Velázquez 1999; AE 1999, 954 a–b; Jordan 2001: no. 89; Curbera and Velázquez 2002; Ribeiro 2006: 256; Kropp 2008: no. 2.1.2/1; Arbabzadah 2009; Versnel 2010: 290–92; Urbanová 2018: no. 50; Sarullo 2020.

Commentary: this defixio was discovered during the excavations carried out in 1987 at the archaeological site 223

Sylloge of  Defixiones from the Roman West Fuente de la Mota (Barchín del Hoyo, Cuenca). This Iberian settlement reached its apogee in the fourth century BCE and was abandoned around 210 BCE. The curse was found near the city’s southeastern gate. Given that the curse tablet has a much later date than the abandonment of the settlement (see below), the depositional context can be explained in one of two ways: first, ruined or abandoned places were thought to be suitable for communicating with the underworld (since these places were thought to be plagued with ghosts [see section I.6.5.2]); second, perhaps this abandoned settlement was used for depositing the curse tablet since the area was associated with the victim (there were several active mines in the vicinity of Fuente de la Mota).

in Greek as well because the curses’ main victims were Greek. It is hard to know. The text has been dated to the end of the first century CE or beginning of the first century CE based on its palaeographic and linguistic features (e.g., the distinction between short and long vowels on side A). The circumstances that led to the writing of this curse are unknown. It begins with an unparalleled phrase: (A) ὑπὲρ ἐμοῦ κα[ὶ] ὑπὲρ τῶν ἐμῶν and (B) pro me pro meis. Also quite unusual is the defigens’ choice to morally justify his/her choice to curse the victims with the use of the adverbs δικαίως (A) and merito (B). Besides naming two individuals directly, Time and Nicias, the curse also contains the words ἄ[λ]λους (A) and ceteros (B), which serve as a sort of all-inclusive formula, similar to that found in 137 (ll. 9–10: omnes quei inimieici Senecae). It is worth noting that between the two sides of the tablet, the named victims appear in a different order resulting in a chiastic structure (for a discussion, see Arbabzadah). On the expression devotos defixos and its parallels see Sarullo 2020.

The curse was written on a small lead disc (for parallels, see 64, 472 and 514). As I confirmed during an autopsy in April 2010, it is intact and in excellent condition with hardly any signs of corrosion. An opisthograph, the inscription is organized in a spiral that begins on the outside and moves inward. The size of the letters decreases as the text moves in toward the centre (ranging from 0.6 to 0.35 cm on side A and from 0.6 to 0.2 cm on side B). The text is written in Greek and Latin. Since the same information is found in both languages, the text is bilingual sensu stricto (for another bilingual curse, see DT 249). The choice to communicate the same message in two different languages could constitute an attempt to facilitate communication with the deities invoked and/or to increase the spell’s potency (see section I.4.2.1.2). Based on palaeographic features, the use of Greek δ for Latin d and the presence of the Greek accusative in -n in the Latin part, the editors have compellingly argued that a single scribe wrote the text and that (s)he was a native Greek speaker. Perhaps (s)he was a professional practitioner working on behalf of a Latin-speaking defigens, or (s)he was writing

After being inscribed, the tablet was pierced twice in its central area coinciding with the names of the two main victims (cf. 338). Translation (Arbabzadah 2016: 195, slightly modified): ‘On behalf of me and on behalf of my house, to those below in Hades I give, I hand over Nicias and Time and the others whom rightly I have cursed. On my behalf, on behalf of my house, (I handle) those cursed (and) bound to those below, those cursed (and) bound to those below, Time and Nicias and the others, (all of) whom rightly I have cursed above. On my behalf, on behalf of my family, Time, Nicias, Nicias.’

146. Bracara Augusta, Braga Provenance: Hill of Maximinos. Current Location: Pio XII Museum. Date: 5th century–beginning of the 6th century CE. [Inv. No.: BRA MAXX 0841] Material: slate. Measurements: unknown. Reading: A ------? [---omi]ne cancer braca«r»rice persequiris [--- S]erpentiu servu Rufi5  ne deunde istu ibi non ducis per policis verticis manos intra annum summ[um?---] [v]indict[u(m)?---] 10 ------?

B [---] deus sante persequiris

Bibliography: Barroca 1991: 146; Barroca 2000: 12; Velázquez 2000: no. 151; Velázquez 2001: 158; Velázquez 2004: no.151; HEp 10: 715.

slate plaque (a common material used in Visigothic epigraphy), whose dimensions and formal characteristics are unknown. An opisthograph, the inscription contains 10 lines of text written in new Roman cursive. The text runs from left to right and has been dated to the end of the fourth or beginning of the fifth century CE based on palaeographic grounds.

Commentary: this defixio was discovered together with 147 on the Maximinos hill during the emergency excavations of Bracara Augusta. It was written on a 224

 Hispania annum summ[um] (A, l. 8; for a parallel, see 92, B, ll. 12– 13; for magical time frames, see Sánchez Natalías 2019a).

We do not know the circumstances that led to the writing of this curse, in which the following orthographic and phonetic features are found: [omi]ne for hominem (A, l. 2), bracarrice for bracaricense (?, A, ll. 2-–), persequiris for persequeris (A, l. 3 and B, ll. 1–2), Serpentiu servu for Serpentium servum (A, l. 4), Rufine for Rufinae (A, ll. 4–5), istu for istum (A, l. 5), policis for pollicis (A, l. 6) and sante for sancte (B, l. 1). In the curse, the defigens expresses the desire that the victim suffer from a cancer (A, l. 2; for a parallel, see 483), which should kill him within a year,

In addition to the use of the cognomen Serpentius, the victim is identified by the name of his owner, Rufina (servu Rufine, A, ll. 2–3), and more unusually by place of origin, Bracara Augusta (braca«r»ice, A, ll. 2–3). Also, the invocation deus sante (B, l. 1) is notable since the editor has connected it to Christian cult (Velázquez 2001: 158). The name Serpentius is derived from serpens (cf. Kajanto 1982²: 116, 359).

147. Bracara Augusta, Braga Provenance: Hill of Maximinos. Current Location: Pio XII Museum. Date: 5th century–beginning of the 6th century CE. [Inv. No.: BRA MAXX 0841] Material: slate. Measurements: 3.9 × 5.6 cm. Reading: A ------? [---h?]omine {can}cancer bracarice persequi[ris?] 5  FICHCEO + intro annum ------?

B [---]sum vindicti vi via de police vique vertice non ductis

Bibliography: Barroca 1991: 146; Barroca 2000: 12; Velázquez 2000: no. 152; Velázquez 2001: 158; Velázquez 2004: no. 152; HEp 10, 715.

The circumstances that led to the writing of this curse are unknown. Although it is not an exact copy of the other curse from Braga, this text is closely related to 146: in fact, the two tablets share linguistic features as well as common phrases and themes. For example, we find omine for homine (A, l. 2), bracarice for bracaricense (A, ll. 3–4), persequi[ris] for perseque[ris] (A, l. 4) and police for pollice (B, l. 2). As far as phrasal and thematic similarities are concerned, note omine {can} cancer bracarice persequi[ris?] (A, ll. 2–4), the magical deadline of one year (A, ll. 5–6) and the repeated path that the cancer ought to follow (B, ll. 2–3: via de police vique vertice…). There are no convincing reconstructions for the sequence FICHCEO (A, l. 5).

Commentary: this defixio was discovered during the emergency excavations of Bracara Augusta together with 146. It was written on a slate plaque (a common material used in Visigothic epigraphy) that already had a crack running from the bottom to top before it was inscribed. An opisthograph, the inscription contains nine lines written in new Roman cursive, which run from left to right and have been dated to the fourth or fifth century CE.

148. Cabrera de Mar Provenance: domestic space. Current Location: M. Mataró. Date: unknown. [Inv. No.: 50.490] Material: lead. Measurements: 9.5 × 11.6 × 0.2 cm. Reading: B (traces)

A dat (traces) [-c.6-] POM (or PCN) Anête A (or L) EO invidiu[s?] (trace) Anio fari fuder IIIII (traces) ++MM+ IIIIII negari(?) pat[i]at INA+V [-c.2-] ̣ 5 miṇimum ut ṣper[---] ++noci vicit NOC[-c.3-]+ Firmani (traces) [---] + IIC+M (trace) [---] ++(trace)++AS++AD [-c.2-3-] amici II[.]A[---] (in the left margin and from top to bottom) iniṃico Pravo+++[---] 225

Sylloge of  Defixiones from the Roman West Bibliography: Bonamusa, Garì et al. 1997: 50, n. 4; IRC I, 209; Bonamusa, Clariana et al. 2000: 169, 175 and 183, n. 8; IRC V, p. 44; AE 2002, 864; Bonamusa, Garì et al. 2004: 92, n. 4; HEp 12, 44.

confirm during an autopsy (March 2010), the tablet is in quite bad condition. An opisthograph, side A contains 10 lines, which run from left to right, except for the final line, which was written along the left edge of the tablet and runs from top to bottom. The text was written in old Roman cursive. Although side B preserves some traces of writing, its current state precludes any reading of the text. Despite the fragmentary nature of the text, Fabré, Mayer and Rodà (editors of IRC) have read the following words: dat or data (A, l. 1), invidia or invidiu[s?] (which would be for invidus, A, l. 2), necari? (A, l. 4), nocuit (A, ll. 5 and 6), F̣irmani or [G]ermani (A, l. 6) and inimico (A, ll. 9 and 10). Furthermore, they reconstruct the phrases nec fari nec dici, nec dicere nec fari (l. 2) and minimum, noli meos spernere voces (l. 5). Even if the current state of the curse tablet precludes us from definitively accepting all of these reconstructions, the presence of terms such as inimico could point towards a juridical dispute behind the writing of this curse (see section I.7.5).

Commentary: this defixio was discovered during the 1978 excavations carried out at the site of Can Modolell (Cabrera del Mar), which has a large Roman villa that was inhabited between the Augustan period and the sixth century CE (for a discussion, see TIR, K/J-31, s. v. Modolell, Can). More specifically, the curse was found among the fill material from the so-called trapezoidal area (see Bonamusa, Clariana et al., 2000: 169). This space separated walls P-14 and P-13 of the cryptoporticus, one serving as a buttress for the other. The curse, which was broken into two fragments that fit together (that measure 9.5 × (7.2) × 0.2 cm and 8.3 × (4.4) × 0.2 cm), was written on a lead sheet that was originally rectangular. The tablet is rusted on both sides, damaged in several areas (especially the top and bottom edges) and has various concretions on side B. As I was able to

After being inscribed, the tablet was folded in half and perhaps cut with a sharp object.

149. Locus Incertus Provenance: unknown. Current Location: private collection. Date: last third of the 1st century–beginning of the 2nd century CE. Material: lead. Measurements: see commentary. Reading: Col. I C(aius) · Ligustius · Maximus Metlius · Spratus Laberius primiplus Q[uad]ratianus · centurio 5 L(ucius) · Anenius · centurio Apuselius · cnturio Cassius Bas(s)us [---?] [-c.2-3-]++++VS Retecius 10 [-c.3-]+++NVS trbunus [-c.2-3-]VS tbunus Rufus legatus Col. II L(ucius) · Enṇ[ius?] SESOS++[---] Q(u)intu+[---] signifer 5 Istanius Rufinu[s---] [---?] Viator Valerius Victor 10 Vessius Fort[tunatus?] ATE++[---] (vacat)

226

 Hispania Bibliography: Rothenhöfer 2016; AE 2016, 2012; Sánchez Natalías 2020a.

Based on the text’s date, the editor has identified the Rufus legatus mentioned in the curse as Q. Pomponius Rufus, coss. suff 95, who commanded the V Legio in the Danube area around the end of the first century CE. This would be the same legate who appears alongside T. Aurelius Fulvius (the governor of Hispania Tarraconensis) and Maturus (procurator augusti) in 134–36 from Emporion. As Rothenhöfer notes, having the same victim mentioned here would be extremely unusual. That said, we need to stress the striking similarities between the curses from Emporion (134–36) and the present curse of unknown provenance: the dating, palaeographic characteristics, layout (left to right and bottom to top) and the same victim (Rufus legatus). Given the many similarities and the lack of any concrete evidence that the defixio is actually from the Danube area (it was in a private collection is Spain at the time of publication), it seems much more reasonable to posit that this curse is from the area of Emporion (for a discussion, see Sánchez Natalías 2020a).

Image source: Rothenhöfer 2016: 237. Courtesy of P. Rothenhöfer. See also App. IV.3, SD 149. Commentary: we do not know when or where this defixio was discovered. According to Rothenhöfer (per litt.): ‘it was bought on the Antiquities market, and as many metal objects came from the Danubian areas within the last decades, I assumed a similar provenance. But indeed, it remains unclear.’ Whatever the case may be, the tablet was found broken in two pieces that fit together at l. 5. All of the original edges, except for the left one, are preserved. The first fragment measures 6 × (6.2–8.5) × 0.2 cm, while the second measures 5.6 × (6.5–8.5) × 0.2 cm. The text, which is organized in two columns, runs from right to left. Given both the space (3.5 cm) that has not been inscribed at the top of fragment I and also the careful layout of the final line in column I, it seems that the text begins with the last line and then ascends up the tablet (Sánchez Natalías 2020a: 91, contra Rothenhöfer 2016: 237; for the same layout, see 131, 134–36, 303, 461, 465 and 525). In general terms, given that the curse contains a list of names, the new order of the reading does not substantially change our understanding of the defixio’s content, except for the signifer mentioned at col. II, l. 5: he ought to be identified as the Quintus that comes before, and not as the Istanius that we found in the next line (contra Rothenhöfer 2016: 245).

After being inscribed, the tablet was pierced (perhaps) twice and folded.

The text was written in capitals that measure between 0.3 and 0.9 cm in height. Some words are separated with interpunctuation consisting of a single dot found halfway up the letters (see I, ll. 6–10 and II, l. 10), but interpuncts are not employed systematically. The text’s palaeographic characteristics suggest a date between the final third of the first century CE and the beginning of the second century CE. We do not know the circumstances that led to the writing of this curse, which targets a group of legionaries and Roman military officials, whose names are all given in the nominative case. The text presents the following orthographic and phonetic features: tbunus and trbunus for tribunus (I, ll. 10–11), Basus for Bassus (I, l. 7), cnturio for centurio (I, l. 6), primiplus for primipilus (I, l. 3), Metlius for Metilius (I, 2) and Qintu- for Quintu- (II, l. 3). With the exception of Basus, most of these errors can be attributed to the difficulties inherent in the text’s layout. As far as the victims go, eight are qualified with their rank within the army: legatus (I, l. 12), tribunus (I, ll. 10–11), centurio (I, ll. 4–6), primipilus (I, l. 3, i.e., ‘the senior centurion of a legion’, OLD s.v.) and signifer (II, l. 4, i.e., ‘a military standard-bearer’, OLD s.v.). The cognomen Retecius is Celtic and already attested (I, l. 9). The text contains various Latin cognomina: Apuselius (I, l. 6; without known parallels), Spratus (I, l. 2; without known parallels), Vessius (II, l. 10; attested as anomen) and Istanius (II, l. 5; attested as a nomen). 227

3 Galliae incisions that comprise a ‘text’ of four or five lines, which apparently cannot be deciphered. The second curse is also irregularly shaped and measures 6.1 × 11.8 × 0.5–0.7 cm. It contains markings that are arranged in four lines of ‘text.’ Judging from the published photographs, these tablets are reminiscent of the pseudo-inscriptions from Bath (317–21). Even though they are illegible, their archaeological context and material lead us to think that these are two new defixiones.

There are 55 known defixiones that come from the Galliae. These can be divided into two main groups: those written in Gaulish and those written in Latin, though some of the Latin texts undoubtedly contain reflections of the Gaulish language (e.g., 157 and 160). Unlike other provinces such as Britain, the curses from Gaul tend to be found individually. In fact, there are only two caches that have been discovered: Amélie-les-Bains with six curses (cf. 150–155; it is difficult to determine in which language these were written) and Trier with 32 defixiones (28 of which are included in the catalogue, cf. 175–202). To this list, perhaps Rom should be added (see 159 below). Curses from Gaul have generally been found in one of two contexts: either in necropoleis or in aquatic contexts (which are essentially water sanctuaries and wells; for a critical re-examination of the traditional concept of aquatic contexts, see Sánchez Natalías 2019b). Chronologically, the defixiones from the Galliae mostly date to the High Empire: the oldest curse tablet comes from Montfo (cf. 157, first century CE), while the the latest curses hail from Augusta Treverorum/Trier and Aquae Tarbelliacae/Dax (cf. 175–202 and 164, respectively).

[Gallia Lugdunensis] Blanzy (Ardennes): this lead tablet (measuring 6 × 11 × 0.1–0.2 cm) was found in a small sanctuary. It was discovered rolled up and was apparently fixed to a wooden object with a nail (cf. 55 for a parallel). Furthermore, a small uninscribed strip of lead was wrapped around the nail. On both sides of this opisthograph there are unidentifiable signs made up of dots as well as two stamps (one depicting a female head, the other a leaf) that are found various times on both sides of the tablet. This type of inscription with dots and stamps is typically found on ingots of all types of metal, but not on curses. That said, the characteristics of the lead sheet, nail and lead strip as well as the archaeological context do indeed suggest that this could be a defixio. Despite my uncertainty and given these similarities, the item at least deserves to be mentioned in this preliminary section (for further information, see Feugère and Lambot 1998: 21 as well as Bailliot and Lambot 2012).

In addition to the defixiones studied in this chapter, we must add several more, which are either currently being studied (e.g., the curse from Bastendorf, see below) or were found during old excavations and were subsequently lost without being published (cf. 159). In this section I have also included basic information and bibliographic references for artefacts that may or may not be defixiones. In the subsequent paragraphs I provide the information that we have about these unpublished defixiones.

[Gallia Lugdunensis] Vindinum, Le Mans: between 2010 and 2011 an emergency excavation was carried out in the middle of Le Mans. This work uncovered the remains of various buildings and a sacred area consisting of a small temple and pond, which were in use from the first century BCE until the third century CE. Inside the pond, excavators uncovered an extensive ritual deposit, which contained all types of offerings: ceramics, organic materials, coins, objects made of gold, silver, iron, copper and lead. The deposit contained a total of 93 lead objects, 34 of which are folded, pierced or rolled up. An additional nine show signs of having been inscribed. This accumulation of lead tablets deposited in a sacred aquatic space strongly suggests that these tablets could be defixiones. That said, the inscribed tablets contain vertical incisions or traces with no clear meaning (see Loiseau apud Chevet et al. 2014: 137–38 and figs. 15–17). These tablets then, like those from Allonnes, recall the pseudo-inscriptions from the sanctuary of Sulis Minerva at Bath (317–21). Until the closed tablets, which perhaps have legible inscriptions, are opened and published, I have opted to include information about these items in this preliminary section of the chapter (for a fuller discussion, see Loiseau apud Chevet et al. 2014: 136–45).

[Gallia Aquitania] Mediolanum, Châteaumeillant: in the summer of 2012 during the excavations of a well located in the oppidum Mediolanum, various folded lead tablets were found at the depth of 13 m. These tablets are part of an immense ritual deposit of some 6,000 items that includes organic material, ceramics and metal sculptures, among other things. Although the lead tablets do not appear to have been inscribed, given their formal characteristics and the archaeological context, it is quite possible that they are indeed defixiones. I thank L. Rossetti for sharing information with me about this collection (for an introductory discussion, see Krausz and Coulon 2015: 303 and 319). [Gallia Lugdunensis] Allonnes: Loiseau (apud Chevet et al. 2014: 144) mentions the discovery of two lead tablets found in 1979 during the excavation of a well in Argenton-le-Marin. The first has an irregular shape and measures 5.2 × 7.8 × 0.6–1.3 cm. The tablet has a series of

229

Sylloge of  Defixiones from the Roman West [Gallia Lugdunensis] Neuville-sur-Sarthe: Louis (apud Chevet et al. 2014: 145) has provided information about an emergency excavation carried out in 2010, during which the remains of a sanctuary were found. On this site, a lead tablet, which contains a series of incissions without any deciphered meaning, was discovered. Given the archaeological contest and the medium, this item bears ressemblance to the pseudo-inscriptions from the sanctuary of Sulis Minerva at Bath (317–21). [Gallia Belgica] Bastendorf: during the excavations of a sanctuary between 1991 and 1994, a defixio was discovered. The archaeological context points to a fourthcentury CE date (see the note in Reinert 2000: 378–79). The curse, which was written on a lead sheet, is currently being studied by Dr Lothar Schwinden, who has been kind enough to send me his preliminary reading: loco sacrato / caca supra. After being inscribed, the tablet was folded twice.

230

Galliae Gallia Narbonensis 150. Amélie-les-Bains Provenance: aquatic context. Current Location: unknown. Date: unknown. Material: lead. Measurements: unknown. Reading:

5



10

Col. I KANTAS NISKAT rogamos et dẹ pecamụ KIOSOT SAV̂ATENON LERANOETDE VXNESOAPETEIA +ETELETA NESOA++LA NVKI ------



Col. II NISQIE KILITIVSI METAT VLATEN ++RVET +P+++++

Bibliography: Henry 1847; Bonnefoy 1866: no. 1; CIL XII, 5367, 1; Wünsch 1900: 270; DT 114; Preisendanz 1933: 156; Ponsich 1952: 228–29; Guiter 1957: 24–26; Coromines 1975: 5–15; Coromines 1976: 148–63; de Hoz 1995: 285–89; Abélanet 2000: 74–75; RIG II.2, *L-97; Martin 2005: 389–90; Mees 2009: 47 and 104.

The curse currently under consideration was written on an irregularly shaped lead sheet, which is broken along its central area. The inscription contains 16 lines of capitals, which run from left to right and are arranged into two columns. Given that this tablet has been lost and all editions are based on Puiggari’s drawings, the present reading follows CIL. De Hoz has noted the difficulty of interpreting this text and has claimed that even if we can distinguish some Latin terms in the text, the majority belongs to a pre-Latin language that is difficult to interpret (1995: 285). Coromines (1975 and 1976) proposed calling this ‘pre-Latin’ language ‘Sorothaptic’ and he has connected it to the so-called ‘Urnfield culture’. Coromines has claimed to have identified a range of lexemes in the curses from Amélie-les-Bains, including: 20 in Latin, 9 in proto-Indo European, 48 in Sorothaptic, 13 in Latin-Sorothaptic and 11 in CelticSorothaptic. That said, Coromines’ theories have not won wide acceptance due to both the opaque definition of the language spoken by the peoples of the ‘Urnfield culture’ and also the many uncertainties regarding the reading of these curses, which, after all, are only based on drawings (for a discussion, see de Hoz 1995: 286 and RIG II.2, p. 250).

Image source: Puiggari, apud Henry 1847: Pl. 71, fig. 1. Commentary: this defixio was discovered in 1845 together with another five tablets in the Lo Gros Escaldador spring in Amélie-les-Bains (Pyrénées-Orientales). In addition to the lead sheets (‘roulés couverts d’écriture,’ according to Puiggari apud Henry 1847: 410), excavators uncovered a series of first-century CE coins that were minted in Emporion and Nîmes. Sometime around the year 1849, the tablets were lost and today can only be studied through Puiggari’s sketches, which were later published by Henry. Based on a palaeographic analysis of the letter s, Abélanet (2000: 71–72) has argued that the tablets are not all contemporaneous, but rather fall into two groups. According to this analysis, the earlier group of tablets (154–55), in which we find capital s, is contemporaneous with the coins (i.e., first century CE), while in the second group (150–52) the letter s consists of two strokes (one vertical, the other diagonal and ascending), which points to a date of the third or fourth century CE. This second type of s recalls the shape of the new Roman cursive s. Nevertheless, both the fact that the rest of the letters do not follow the same trend (see n and m in 153–54, made with three and four strokes respectively instead of a zigzagging one, as is customary in new Roman cursive) and the presence of similar invocations in 150 and 154 do not support Abélanet’s theory. On the whole, it strikes me as more prudent to conclude that all the curses were most likely deposited at the same time.

In the present tablet, there has been some consensus among experts about the reading rogamos (I, l. 2; for rogamus) and et depecamụ (I, ll. 3–4; for et deprecamur). Furthermore, scholars have tended to agree that the deities invoked are the Niskas (where the final -t should be taken as a cursive s). This theonym has been connected to Basque ‘neska’ (Lizop 1931, apud de Hoz 1995: 286). However, de Hoz has cautioned us against overreading this connection, since a single word in a religious context could very well be nothing more than a loanword (1995: 288). As this same

231

Sylloge of  Defixiones from the Roman West scholar has pointed out, Niskas can also be compared to the toponym Nescania (located in Hispania Ulterior Baetica), a place where there is a Latin dedication to fonti divino (see CIL II 5, 837). To add yet another comparandum, we must mention the invocation of Neptunus and Niskus found on 451. In this British curse, Tomlin has argued that Niskus can either be a masculine singular form from Niskas or the Celtic form of Neptune (see Tomlin 1997: 457 and 1999b: 562, n. 52 for the two possibilities).

As Lambert (apud RIG II.2, p. 250) and Martin (2005: 389) have argued, kantas (I, l. 1) is a divine epithet related to canta (‘stone’, ‘rock’ or ‘mountain’, the last of which is a common element in toponyms; see Villar 1990: 381-386). Finally, Martin (2005: 390) has interpreted the second column of the text as a list of personal names in the genitive. While Celtic curses do indeed contain lists of names, such lists tend to have bipartite ononomastic formulae consisting of a cognomen and patronymic. In this curse, the second column seems to conform to this regular practice.

151. Amélie-les-Bains Provenance: aquatic context. Current Location: unknown. Date: unknown. Material: lead. Measurements: unknown. Reading: A NICAS QVITE rogamus++++S SSYATISNVMEMA S++NIVELDELA 5 RES++NVQVAI AVTETE CVMA++ B KENVMENE MAXIMIE FLAVCRE ILLIVSSIROES++++QVAE ANTQVID++RVID ASETIV̂AT LAAOKRIOS VCAPOSIMA ATXEXIAIA 5 OS NIAM CAT++ON+++ SNONEVOSTRIM M ++ TATINVM VLLXKI++ KI OHIR

Bibliography: Henry 1847; Bonnefoy 1866: nos. 2 and 3; CIL XII, 5367, 2 and 3; Wünsch 1900: 270; DT 115–16; Preisendanz 1933: 156; Ponsich 1952: 228; Guiter 1957: 24–26; Coromines 1975: 15–27; Coromines 1976: 163– 81; Abélanet 2000: 73–74; RIG II.2, *L-97; Martin 2005: 390; Mees 2009: 47 and 104.

along its edges. Given that the tablet has been lost and all editions are based on the drawings of Puiggari, the present reading follows CIL. An opisthograph, the inscription contains 17 lines, which run from left to right and were written in capitals. The tablet contains some words found in the previous tablet, such as the theonym Nicas (A, l. 1; for Niskas) and the verb rogamus (A, l. 2). Both Abélanet (2000: 73) and Martin (2005: 390) have reconstructed quite (A, l. 1) as aquige(nas), which would be a divine epithet that refers to the deity’s aquatic context. For numema (A, l. 3), Abélanet (2000: 74) has further proposed the reading nuscuema, based on the similarity between the first m in numema and the syllable cu in A, l. 7. Like the previous text, there is no scholarly consensus about the reading or interpretation of this curse.

Image source: Puiggari, apud Henry 1847: Pl. 71, figs. 2 and 3. Commentary: this defixio was discovered in 1845 together with another five tablets (150, 152–55) in the Lo Gros Escaldador spring in Amélie-les-Bains (for the archaeological context, see 150). The curse was written on an irregularly shaped lead tablet that is now broken 232

Galliae 152. Amélie-les-Bains Provenance: aquatic context. Current Location: unknown. Date: unknown. Material: lead. Measurements: unknown. Reading: A DEMETI ------ B ------ AMIKIOV ------

Bibliography: Henry 1847; Bonnefoy 1866: nos. 4 and 5; CIL XII, 5367, 4 and 5; Wünsch 1900: 270; DT 117; Preisendanz 1933: 156; Ponsich 1952: 228–29; Guiter 1957: 28; Coromines 1975: 27–34; Coromines 1976: 182– 93; Abélanet 2000: 73; RIG II.2, *L-97; Martin 2005: 390; Mees 2009: 47 and 104.

archaeological context see 150). The curse was written on an irregularly shaped lead tablet that was broken along its bottom third. An opisthograph, the inscription contains 19 lines of capitals, which run from left to right and have proven extremely difficult to interpret. Given that the tablet has been lost and all editions are based on the drawings of Puiggari, the present reading follows CIL. Despite the difficulties inherent in reading this text, Abélanet (2000: 73), who is supported by Martin (2005: 390), has read Demeti or Demet(r) (A, l. 1), which would be part of the personal name Demetrius. They have also read item (A, l. 2) with a ligature connecting the t and e. The rest of the text has proven indecipherable.

Image source: Puiggari, apud Henry 1847: Pl. 71, figs. 4 and 5. Commentary: this defixio was discovered in 1845 together with another five tablets (150–51 and 153–55) in the Lo Gros Escaldador spring in Amélie-les-Bains (for the

153. Amélie-les-Bains Provenance: aquatic context. Current Location: unknown. Date: unknown. Material: lead. Measurements: unknown. Reading:

AXX+++ ++EAVB++ AXSONIS

Bibliography: Henry 1847; Bonnefoy 1866: no. 6; CIL XII, 5367, 6; Wünsch 1900: 270; DT 118; Preisendanz 1933: 156; Ponsich 1952: 228–29; Guiter 1957: 29; Coromines 1975: 35; Coromines 1976: 193–94; Abélanet 2000: 73; RIG II.2, *L-97; Martin 2005: 390; Mees 2009: 47 and 104.

Lo Gros Escaldador spring in Amélie-les-Bains (for the archaeological context, see 150). The tablet, a small irregularly shaped lead fragment, contains a text of three lines of capitals, which run from left to right. Given that the tablet has been lost and all editions are based on the drawings of Puiggari, the present reading follows CIL. Lambert (apud RIG II.2, p. 250) has read the onomastic element Axona in l. 3.

Image source: Puiggari, apud Henry 1847: Pl. 71, fig. 6. Commentary: this defixio was discovered in 1845 together with another five tablets (150–52 and 154–55) in the 233

Sylloge of  Defixiones from the Roman West 154. Amélie-les-Bains Provenance: aquatic context. Current Location: unknown. Date: unknown. Material: lead. Measurements: unknown. Reading:

5



DOMXSAA NISKASROG MOSETDE TAMVS DINNO NN

archaeological context, see 150). The text was written on an approximately triangular sheet of lead (for a similar shape, see 106). The inscription contains six lines of capitals, which run from left to right. Given that the tablet has been lost and all editions are based on the drawings of Puiggari, the present reading follows CIL. In the text, it is possible to make out the theonym Niskas accompanied by the verb rog[a]mos (ll. 2–3; for rogamus; cf. 150, I, l. 2). Abélanet (2000) and Martin (2005) have suggested the reconstruction depecamus from DETAMVS (ll. 3–4), which would make good sense given the invocation found in the first curse from this collection.

Bibliography: Henry 1847; Bonnefoy 1866: no. 7; CIL XII, 5367, 7; Wünsch 1900: 270; DT 119; Preisendanz 1933: 156; Ponsich 1952: 228–29; Guiter 1957: 29; Coromines 1975: 35; Coromines 1976: 194–95; Abélanet 2000: 73; RIG II.2, *L-97; Martin 2005: 390; Mees 2009: 47 and 104. Image source: Puiggari, apud Henry 1847: Pl. 71, fig. 7. Commentary: this defixio was discovered in 1845 together with another five tablets (150–53 and 155) in the Lo Gros Escaldador spring in Amélie-les-Bains (for the

155. Amélie-les-Bains Provenance: aquatic context. Current Location: unknown. Date: unknown. Material: lead. Measurements: unknown. Reading:

COLOS

Commentary: this defixio was discovered in 1845 together with another five tablets (150–54) in the Lo Gros Escaldador spring in Amélie-les-Bains (for the archaeological context, see 150). It was written on a strip of lead and contains a short, but difficult to decipher, inscription. Scholars cannot even agree which way to orient the text (cf. Bonnefoy 1866 and Abélanet 2000). Given that the tablet has been lost and all editions are based on the drawings of Puiggari, the present reading follows CIL.

Bibliography: Henry 1847; Bonnefoy 1866: no. 8; CIL XII, 5367, 8; Wünsch 1900: 270; DT 120; Preisendanz 1933: 156; Ponsich 1952: 228–29; Guiter 1957: 29; Coromines 1975: 36; Coromines 1976: 195; Abélanet 2000: 73; RIG II.2, *L-97; Martin 2005: 390; Mees 2009: 47 and 104. Image source: Puiggari, apud Henry 1847: Pl. 71, fig. 8.

234

Galliae 156. Mazan Provenance: necropolis. Current Location: Lapidaire de Carpentas. Date: unknown. Material: lead. Measurements: unknown. Reading: (traces)EKE ++NMS ++REV+++ ++++++++CCCAHIVL 5 + E +++M [-c.2-3-]XIATISENOSC (traces)VZOREIOS

Bibliography: Barruol and Barruol 1963: 108–09; Rolland 1964: 562.

the inscription contains 30 names and the jar is filled with the bones of a small chicken. The jar was then pierced with nails and buried in a building with an industrial function (I thank J. Lamont for this information; on this jar, see now Lamont 2021).

Image source: Barruol and Barruol 1963: 108, fig. 18. Commentary: this defixio was discovered in 1963 during the excavations of the necropolis of Saint-Andéol (Mazan) near burial no. 1. Outside of the grave but at the same level as the deceased’s head, archaeologists found an urn that contained the bones of a small chicken, probably an offering to the infernal deities, below which there was the present defixio (see Barruol and Barruol 1963: fig. 4). This method of depositing the tablet can perhaps be explained as an attempt on the defigens’ part to place the curse next to a tomb without being detected. We cannot rule out the possibility that it was deposited as an offering at the time of the burial. Even if this method of depositing the tablet is unparalleled in the Roman West, there it is an Athenian curse (dated to c. 320 BCE) inscribed on a jar:

The curse was written on a lead sheet that is broken into three parts and whose maximum dimensions are (5) × (10) cm. The surface of the tablet is rather corroded, which has made reading the text more difficult. The inscription contains eight lines, which run from left to right and were written mostly in capitals (although in line 2, n and m are written in new Roman cursive). The layout fills the space left by the tablet’s iconography, which according to Barruol and Barruol represents a snake or a boat (1963: 109). It is worth noting that the iconography bears a resemblance to a semicircular drawing on a Greek curse tablet from Corinth (see Jordan 1994b: 120–21). The reading provided here is based on the published drawing of the tablet.

157. Montfo Provenance: well. Current Location: unknown. Date: 50–60 CE. Material: lead. Measurements: 6.5 × 9 cm. Reading: A quomodo hoc plumbu(m) non paret et decadet sic deca dat aetas membra vita bos grano(m) mer(x) eorum qui 5 mihi dolum malu fecerunt item Asuetemeos Secundina qu(a)e illum tulit et Verres Tearus

B et Amarantis et hoc omnia vobis Dii interdico in omni bus sortebus tam celebrare 5 MASITLATIDA concinere necra cantum COL[-c.2-]SCANTVM et omnes deos [---] ta datus [---]

Bibliography: Bacou and Bacou 1975: 18–20; Marichal 1981; Lejeune 1981; AE 1981, 621; Kropp 2008: no. 4.4.1/1; Mees 2009: 70–72; Urbanová 2018: no. 226; Gordon 2019a: 122–23.

Commentary: this defixio was discovered in 1975 during the excavation of a well in the oppidum Montfo (located northeast of Magalas [ancient Baeterrae], Béziers). The tablet, found at a depth of 11 m (see Bacou and Bacou 235

Sylloge of  Defixiones from the Roman West 1975 for the archaeological context), was inscribed on an irregularly shaped lead sheet. An opisthograph, the inscription contains 16 lines, which run from left to right and are dated between the years 50 and 60 CE on palaeographic grounds.

In the second part of the text, the practitioners’ attempts to prevent the gods from participating in the MASITLATIDA (B, l. 5), a word that is possibly Gaulish and refers to an indigenous ritual of unknown characteristics (Lejeune 1981: 52). Furthermore, the defigens urges that the same divinities do not sing the necracantum (B, ll. 5–6) or the songs of the dead (for this hybrid word, see Lejeune 1981: 52). As Marichal has put it, ‘il s’agit d’une espèce d’excommunication, l’équivalent de cette interdiction de sacrifier qui était la peine la plus grave que les druides infligeaient, nous dit César, à ceux qui avaient commis quelque crime ou qui refusaient leur décision lorsqu’il s’était élevée “une contestation relative à un héritage ou a des limites” c’est-à-dire lorsque quelqu’un a subi un dolum malum!’ (1981: 50).

The tablet was written in response to a dolum malu(m) (A, l. 5). The text presents the following orthographic and phonetic features: plumbu for plumbum (A, l. 1) decadet for decidit (A, 2), decadat for decidat (A, ll. 2–3), grano for granum (A, l. 4), mer for merx (A, l. 4), malu for malum (A, l. 5), que for quae (A, l. 7), sortebus for sortibus (B, l. 4) and necrocantum for necracantum (B, ll. 5–6). The curse begins with a similia similibus formula that compares the way in which the lead tablet will fall to the bottom of the well where it was deposited (decadet, A, l.2) with the (immediate) downfall of the curse’s victims (for a full discussion, see Sánchez Natalías 2018: 12). It goes without saying that the defigens had already planned how the tablet would be deposited before inscribing the text. Accordingly, this defixio underscores the close connection between inscribing and depositing a curse. As the text makes clear, the victims’ decline is to be physical (affecting their aetas, membra, vita, A, l. 3) and also economic (including their bos, grano(m), mer(x), A, l. 4). Though we do not learn of all the details that led to the writing of the curse, the phrase qui mihi dolum malu fecerunt (A, ll. 4–5) certain gives us a good idea: dolus malus is often used in a legal context (see 361, l. 5: dolum ṃạḷụm intụḷerunt; 493, ll. 1 and 5: dolum malum adm[isit] and dolum malum adhibet, etc.). Therefore, the defixio appears to be juridical. Next, the list of victims is introduced by item (cf. 2).

Only the victim Asuetemeos (A, l. 6) is qualified with a matronymic phrase: Secundina qu(a)e illum tulit (A, l. 7; the more usual formulation would be quem peperit). Furthermore, this victim has a Celtic name, while his mother has a Latin one. Amarantis is a Greek cognomen (A, l. 9; see Solin 2003: 1154), while Verres (A, l. 8; see OPEL IV, 159) is Latin. There are no known parallels for Tearus (A, l. 8), which Marichal has taken as an allusion to the victim’s home region (the Teari lived in Hispania Tarraconensis, according to Pliny HN III, 23). Gordon (2019a: 122 [side A], and my own): ‘As this lead plaque disappears and drops down, so may the person, the body, the vitality, the cattle, the grain, the goods of those who have defrauded me waste away: Asuetemeos (son of) Secundina who stole it, and Verres Tearus, and Amarantis. I enjoin you, gods, through this (lead tablet) that in all instances (...) not celebrate and not sing... ’

158. L’Hospitalet-du-Larzac Provenance: necropolis. Current Location: MA Millau. Date: unknown. Material: lead. Measurements: 14 × 26 × 0.9–1.3 cm. Reading: Frag. I A insinde · se · bnanom bricto[m i-] n eíanom · anuana sanander[---] na · brictom · uidluias uidlụ[---] tigontias · so · adsags·ona · seue[rim] 5 tertionicnim · lidssatim liciatịṃ eianom · uoduiuoderce lunget utonid ponc · nitixsintor sị[es] duscelinatia inanon anuaṇ[a] esi · andernados brictom · bano[na] 10 flatucias · paulla dona potiti[us] iaia · duxtir · adiegias poti[ta m-] atir paullias · seuera du[xtir] ualentos dona paulli[-c.1-]uṣ adiega · matir · aiías 15 potita dona primụṣ ị[---] abesias 236

Galliae B etic epotiniosco· et[ic] ruficna casta dona B[---] nonus c ̣o etic diligenti ṣoc[---] ulatio·nicn om aucitiọnim[---] 5 aterem potiti ulatucia ṃat[ir] banonias ne · incitas · biontutu in das mnas ueronadas brictas lissinaụ[e] seuerim licinaue · tertioni[cnim] eíabi tiopritom biietutu semiṭ[---] 10 ratet seuera tertionicna[-c.3-]du[---] ne incitas biontutuṣ++++[---] anatia nepi ạndạ+++ [---]ạd incorsonda b+++ [---]+pi·lu dorecon (or donicon?) +s+++ 15 incarata Frag. II A [---]a · senit conectoṣ[---] [---]onda bocca nene+[---] [---]+rionti onda boca ne[---] +on barnaunom ponc nit5 ixsintor sies eianepian digs ne lisatim ne licia[---] tim · ne rodatim · biont[---] utu semnanom sagitiont[---] ias seuerim lissatim licia[---] 10 tim anandognam ạcolụṭ[---] utanit aṇdognamạ[---] da bocca[---] dịomịnẹ[---] B (hand M) aia[-c.3-] cicena[---] nitianncobueđliđat[---] iasuolsonponne antumnos · nepoṇ 5 nesliciatia neosuode neiauodercos · nepon· (hand N) su[-c.3-] biiontutu se mn[---] anom adsaxs.nadoc ̣[---] suet petidsiont sies peti sagitionṭias seụ[---] 5 [er]im tertio lissatiṃ[---] [-c.2-]s anandogna[-c.3-] [-c.3-]ictontias+[---] Bibliography: Fleuriot 1985a and b; Lambert 1985a and b; Lejeune 1985a and b; Marichal 1985a and b; Vernhet 1985a: 2–9; Vernhet 1985b: 96–103; Olmsted 1989; Meid 1990: 47–48; Olmsted 1991: 280–83; Meid 1992: 40 and 43–47; Lambert 1994: 160–72; Koch 1996: 37–40; Lambert 1996a: 65–85; Meid 1996a; Schmidt 1996; Guyonvarc’h 1997: 187–95; RIG II.2, L-98; Mees 2004; Ciurli 2008; Mees 2009: 53–69.

Image source: RIG II.2, L-98 (figs. 141, 143, 145 and 147). Courtesy of P.Y. Lambert. Commentary: this defixio was discovered in 1983 during the excavations of a necropolis outside of a Gallo-Roman village at La Vayssière (L’Hospitalet-du-Larzac). The curse was found inside grave 71, an incineration burial with an array of grave goods consisting of an iron ring

237

Sylloge of  Defixiones from the Roman West and 40 ceramic vessels (dated between 90 and 110 CE). After being broken into two, the tablet was placed on top of the cinerary urn as if it were a lid with one piece placed on top of the other (A I, A and II, A were placed together; see Lejeune 1984: 710, fig. 3. and Vernhet 1985a for the archaeological context).

Latin curse tablets. Perhaps both lexemes are just marking the professions of these women, as happens in other texts (see 52, 101–02, etc.). The second type of onomastic structure is a list of names (IA, ll. 9–16; IB, ll. 1–6) in which individuals are defined by family relations, such as matir (‘mother’), duxtir (‘daughter’) and dona (of unknown meaning, but perhaps related to Latin domina or alternatively a Celtic word for ‘nanny’ or ‘daughter-in-law’).

The curse was written on a lead sheet that must have originally been roughly rectangular. The way in which to defixio was deposited has led to the corrosion and calcareous accretion of its surface. Before being inscribed, the tablet was hammered out and likely broken in two, since the curse is read as if it were really two separate tablets: the text begins with IA–IB and the moves onto either IIA–IIB or IIB–IIA (the exact order is not clear; see RIG II.2, pp. 251 and 261 for a discussion). Marichal (1985a and b) has studied the text’s palaeography and concluded that there were two different authors, whom he refers to as N and M based on the use of these letters at the end of words (the latter uses –m ‘comme une sorte de latinisme’, according to RIG II.2, p. 261). N can be detected first at IIB, ll. 1–6 and appears to have had a less practised hand. M writes with fluidity, uses more archaic letter forms (especially with l and a) as well as đ long i and two different types of m.

The names found in the curse are both Latin (Paulla: IA, l. 10, Potita: IA, ll. 11 and 15, Valens: IA, l. 13, Severa: IA, l. 13, Prima: IIA, l. 15 and Casta: IB, l. 2) and Celtic (Banona: IA, l. 9, Flatucia/Vlatucia: IA, l. 10 and IB, l. 5, Aia: IA, l. 14, Adiega: IA, l. 14, Abesia: IA, l. 16, Vlationos: IB, l. 4, Aucitiona: IB, l. 4 and Epotin(a): IB, l. 1); furthermore, there are some mixed names, which incorporate Latin and Celtic elements (Tertionos: IA, l. 6 and Ruficna: IB, l. 2). Scholars have connected the word bricto[m] (IA, l. 1, cf. brixtia in 163, l. 3) to Irish bricht and Gaulish -brith (‘spell’ or ‘enchantment’). The editors have also singled the lexemes nitixsintor (IA, l. 7, equivalent to Latin defigo) and duscelinatia (IA, l. 8, a compound of dus- ‘bad’ and perhaps nata ‘poem’) as belonging to the realm of magic. Another word to receive scrutiny is adsagsona (IA, l. 4), which could be a theonym derived from the root sag‘to search’. Antumnos (IIB, l. 4), for its part, could be equivalent to Gaulish annwfn ‘the underworld’, while donicon (IB, l. 14) would be connected to Gaulish χdonion ‘terrestrial, human’. Anatia (IB, l. 12) is derived from *ana-tion ‘to love’. In addition to these lexemes, there are also elements related to Latin (e.g., licia-, IIA, ll. 6 and 9 or incitas, IB, l. 11), adapted from Latin (e.g., anderna-, IA, l. 9; perhaps for infernus [cf. 163, l. 2]) or taken directly from Latin (e.g., bocca for bucca, IIA, ll. 2 and 12).

Lambert (2002: 262ff.) has studied the curse’s onomastics and argued that there are two different types of structures. The first consists of a personal name followed by a matronymic. Accordingly, we find seuera tertionicna (e.g., ll. 4–5; ‘Severa, daughter of Tertiona’), who is mentioned no less than five times; she is also associated with the words liciata and lidsata, which, according to Lambert, ‘désignent deux types de sorcières, selon la magie utilisée’ (apud RIG II.2, L-98, p. 262). Although his theory is doubtlessly suggestive, it is also totally unparalleled in the corpus of

Gallia Aquitania 159. Rauranum, Rom Provenance: well. Current Location: M. Saint-Germain-en-Laye. Date: end 3rd century–beginning 4th century CE. [Inv. No.: 82939] Material: lead. Measurements: 10 × 7 cm. Reading: A apecialligarti estiheiontcaticato (or caticno?) atademtissiebotu cnasedemtiticato (or ticino?) 5 bicartaontdibo ṇasociodecipia sosiopurasosio eoẹ+++eiotet sosiopurah++++ 10 suade++ix+o+cn auntaontiodiseịạ

238

Galliae B teuoraiimo (or teuoraiímo?) ihzantaotehi zontantatecom prixtososioberti (or priato?) 5 noipommioateho tissepoteatepri (or poecato?) auimontantate (or auimoatantate or auímonantate?) ontezatimezo 10 ẓiateuoraiimo apesosiobertị ịṃ+++demtiss++ uieie [-c.9-]

Bibliography: Jullian 1897a; Jullian 1898; Wünsch 1900: 268; Nicholson 1904: 129–153; DT 110; Rhys 1905: 107–18; Rhys 1906: 366–70; Dottin 1918: 170–72; Jullian 1921; Chapeau 1924; Preisendanz 1933: 156; Haas 1943: 286–87 and 290–95; Whatmough 1950: 391–94; Pokorny 1953: 306; Haas 1959; Haas 1961; Egger 1962; Egger 1962b; García Ruiz 1967: no. 62; Olmsted 1988: 355–56 and 367–74; Foucher 1991: 32–34; Olmsted 1991: 282– 88; Gager 1992: no. 16; Lambert 1994: 174–76; Meid 1996b; CAG 79: 230-I 10; Guyonvarc’h 1997: 180–186; RIG II.2, *L-103; Meid 2007; Kropp 2008: no. 4.3.4/1; Mees 2009: 94–97 and 101–04; Urbanová 2018: no. 69.

from left to right and were written in capitals, except for certain letters that use forms characteristic of new Roman cursive. Note that the writer employed both types for m (see B, ll. 1 and 3) as well as the use of new Roman cursive s (A, l. 2 and ff.), but capital r (B, l. and ff. 1). The letters measure between 0.4 and 0.8 cm in height. Judging from the text’s palaeographic characteristics, Marichal (apud RIG II.2, pp. 288–91) has dated the tablet between the end of the third century and the beginning of the fourth century CE. The interpretation of the curse has proven controversial: As Lambert has put it, ‘l’analyse de l’inscription est tributaire de conditions difficiles: incertitude d’identification pour tel ou tel tracé, incertitude générale en matière de segmentation du texte, incertitude enfin sur la langue’ (apud RIG II.2, 286). There have been many theories about the curse’s language: Whatmough proposed that the text is written in a magical language, while Haas has proposed that it is in a form of Gaulish containing Latinisms. Alternatively, Egger has argued that the curse is in a Latin heavily influenced by Gaulish. Following Egger, Kropp has also read the inscription as a ‘corrupt’ Latin text. For Urbanová, this unintelligible text presents Gaulish, Greek and Latin features.

Image source: P. M. Duval and R. Marichal apud RIG II.2, *L-103 (figs. 153–54). Courtesy of P.Y. Lambert. Commentary: this defixio was discovered in 1887 during the excavations carried out by J. Blumereau at ancient Rauranum, where a villa and some surrounding structures were found. During the same dig, archaeologists excavated a well that had a diameter of 1.8 m and a depth of 20 m. Between 10 and 12 m down, archaeologists found ‘une quarantaine de ces morceaux de plomb… Leur poids, leur forme et leurs dimensions varient à l’infini… Les uns (au nombre de douze environ) sont de longues lanières de plomb, épaisses, mais étroites et irrégulières, enroulées sur elles-mêmes: ces morceaux ne présentent pas, d’ordinaire, trace de clous. D’autres sont des plaques quadrangulaires également roulées, mais ayant été également clouées: on y aperçoit de deux à quatre clous. D’autres enfin sont des plaques régulières… toutes percées de gros trous carrés, en nombre variant d’un à quatre. Enfin je trouve bon nombre de fragments complètement déformés, véritable amalgame de scories.’ (Jullian 1897a: 132, n. 1). Given the physical characteristics of the items and the context in which they were found, Jullian went on to conclude that they were a collection of defixiones. That said, they were uninscribed (and never fully published). Accordingly, this intriguing cache will not be further analysed here, and instead we will turn our attention to a defixio that was found at a depth on 16 or 17 m in the same well.

Lambert (2002: 295) has recognized several Celtic lexemes in the text (e.g., apecialli (A, l. 1) and sosio (A, ll. 7 and 9, B, l. 4)), as well as Latin ones (pura (A, l. 7 and 9), poteat (B, ll. 6–7), decipia (A, l. 6), demti (A, ll. 3–4, B, ll. 11) and perhaps the adjective bona (in Di-bona, A ll. 5–6; this is generally identified as the theonym Divona). The text is thought to contain a series of Gaulish verbs: uoraiimo (B, ll. 1 and 10) and atepriauimo (B, ll. 7–8), both of which are first personal plural forms and could possibly be taken from the Latin ending –mus, as well as bicartaont (A, l. 5), atecompriato (B, ll. 3–4) and atepriato (B, l. 6; the last two may be derived from atepri(a)-, which can be compared to Irish ath-cren-). As Jullian has noted, the text employs various types of repetition, such as at A, ll. 7 and 9 with the lexeme sosio (sosio pura sosio and sosio purah; also note sosio berti at B, ll. 4 and 11). The meaning of these repetitive sequences is currently beyond our grasp (for a parallel, see 160).

The curse was written on a rectangular lead sheet. An opisthograph, the inscription contains 23 lines, which run 239

Sylloge of  Defixiones from the Roman West 160. Mediolanum Santonum, Villepouge-Chagnon Provenance: necropolis. Current Location: MA Saintes. Date: 2nd century CE. [Inv. No.: 49.2112-1 and 49.2112-2] Material: lead. Measurements: 10 × 8.5 cm. Reading:

5



10



5



10



Tab. I denuntio personis infrascribtis Lentino et Tasgillo uti adsin ad Plutonem quomodo hic catellus nemin[i] nocuit sic IMQVEOLOSICODMA nec illi hanc litem vincere possint quomodi nec mater huius catelli defendere potuit sic nec advocati eorom e[os d]efendere non possint sic il[lo]{s} [in]imicos ATRACATETRACATI GALLARA precata EGDATARA HEHES celata mentis ablata [et] at Proserpinam hinc Tab. II aversos ab hac l[i]te esse quamodi hic catellus aversus est nec surgere potest{i} sic nec illi sic tra(n)specti sin quomodi ille quomodi in hoc m[o]nimont animalia ommutuerun(t) nec surgere possun nec illi NVT[---] ATRACATETRACTI GALLARA precata EGDARATA HEHES celata mentis ablata

Bibliography: Musset 1897; Jullian 1897b and c; Wünsch 1897: 241–44; DT 111–12; Diehl 1910: no. 856; CIL XIII, 11069–70; ILS, 8752; Jeanneret 1917: nos. 111–12; Preisendanz 1933: 156; Foucher 1991: 30–32; Gager 1992: no. 53; Vienne 1992: 213–14; ILA Santons 104; Kropp 2008: nos. 4.3.1/1 and 4.3.1/2; Urbanová 2018: nos. 67 and 68.

The curse was written on two roughly rectangular lead sheets. Each has two trapezoidal protrusions jutting out from opposite sides that make the tablets resemble tabulae ansatae. These ansae were all pierced. If the tablets are placed together with the inscribed faces on the inside, the holes line up, which suggests that they were pierced as a means of ‘closing’ the tablet and keeping the inscription from view. This mechanism, which makes the tablet a sort of diptych, only has one parallel in the Roman West that comes from Pompeii (71), which was also written on two lead sheets that resemble tabulae ceratae.

Image source: Jullian 1897b: 178. Commentary: this defixio was discovered in 1896 during the excavations of a Gallo-Roman funerary monument located along the road that connected Burdigala and Augustodunum (the tablet from Autun [cf. 166], came from the latter city and was deposited in a similar context). During the dig, archaeologists encountered numerous ceramic sherds, a sculpted head and some coins among which there was an as depicting Marcus Aurelius that was minted in 172 CE (for a discussion, see Musset 1897).

The present defixio contains 16 lines, which run from left to right and were written in old Roman cursive dating to the second century CE. Jullian singled out two distinct hands: one wrote I, ll. 3 and 14 and II, ll. 9–12, while the other wrote the rest. This seems unusual but cannot be disproven without a new autopsy of the text. Based on its content, this curse should be grouped among the juridical 240

Galliae defixiones. The text presents the following orthographic and phonetic features: infrascribti for infrascripti (I, ll. 1–2), adsin for adsint (I, l. 3), quomodi for quomodo (I, l. 7; II, ll. 5 and 6), quamodi for quomodo (II, ll. 1–2), potesti for potest (II, l. 3), sin for sint (II, l. 4), traspecto for transpecto (II, l. 4), mnimont for monomento (II, l. 6), ommutuerun for obmutuerunt (II, l. 7) and possun for possunt (II, l. 8). The participle tra(n)specti (‘impaled’) is rare and is not paralleled in any other curse from the Roman West.

and cannot leave the funerary monument (quomodi in hoc m[o]nimont animalia ommutuerun(t) nec surgere possun, II. ll. 6–8; for another text that refers to the deceased with such an analogy, see 2). The repeated ‘formulae’ which include voces magicae (I, ll. 11–13 and II, ll. 9–12) also contain some Latin words. These sections of the curse have proven difficult to interpret. Jullian (1897 b: 180) identified Gallara (I, ll. 11–12; II, l. 9) as a sorceress and the author of the curse, though this is somewhat farfetched and would be without any known parallels. Nevertheless, Foucher has followed Jullian’s interpretation and also distinguished ‘ater: noir et tetra: quatre’ (I, l. 11; II, l. 9), which he translates as ‘Quatre malédictions aussi pour Gallara qui a été sollicitée par eaux dans leurs prières et dans leurs grimoires (ou: qui leur sert d’inspiratrice). Qu’elle reste donc cachée et privée de raison!’ (1991: 31). That said, until the curse tablet is re-examined and a new reading is established, the interpretation and translation of these lines should be taken with a grain of salt.

The text is structured around similia similibus formulae (which are rare in Gaul; for the only known parallel from the region, see 157) and voces magicae that are found at the end of each fragment. Relying on the principles of sympathetic magic, the formulae compare the victims to a tortured and helpless puppy (hic catellus nemin[i] nocuit, I, ll. 4–5). The repeated use of the deictic hic has led scholars to suppose a real puppy was killed during a ritual. By comparison with the puppy, the victims cannot harm the defigens (I, ll. 5–6) or defend themselves in court (I, ll. 8–10). In the second part of the curse, there is a comparison between the puppy’s inability to get up and the victims’ ability to safely (?) leave court (…aversos ab hac l[i]te esse, II, l. 1). The above-mentioned tra(n)specti sin raises questions about the use of the nail in the ritual (perhaps it was used to kill the puppy). What is certain is that the tablet was pierced at the end of II, l. 5 where the victims are referred to as ‘impaled’: (tra(n) specti sin) quomodi ille. Strategically piercing a tablet to increase its efficacy or place emphasis is normal (e.g., 72, 338). Finally, the victims are also compared to the animalia (perhaps for animae?) that are unable to speak

Translation (Gager 1992: no. 53, modified): ‘I denounce the persons written below, Lentinus and Tasgillus, in order that they may depart from here for Pluto and Persephone. Just as this puppy harmed no one, so (...) may they not be able to win this suit; just as the mother of this puppy cannot defend it, so may their lawyers be unable to defend them, (and) so (may) those (legal) opponents be turned back from this suit; just as this puppy is (turned) on its back and is unable to rise, so neither (may) they; they are impaled through, just as this is; just as in this tomb animals/souls have been silenced and cannot rise up, and they (can)not...’

161. Les Martres-de-Veyre Provenance: necropolis. Current Location: M. Bargoin. Date: unknown. Material: lead. Measurements: 10 × 12.9 cm. Reading: A [------] SCOI divos et I nolis M [---]LONA[---] 5 SIIAM[---] TOTLI[---] et ISCESSILII advoc[atus?] IIITAMOL[---] IINT AN B [------] TISCO[---] +CS+++OS LITVTION

241

Sylloge of  Defixiones from the Roman West Bibliography: Audollent 1924: 16; RIG II.2, *L-102; Mees 2009: 99–100.

a reading of the texts. The (extremely well-) preserved grave goods found in these burials date to the end of the first century CE.

Image source: RIG II.2, *L-102 (fig. 152). Courtesy of P.Y. Lambert.

The present curse was inscribed on an irregularly shaped lead sheet, which is broken along the right edge and whose entire surface has been corroded. An opisthograph, the inscription contains 12 lines, which run from left to right and were written in capitals. In the inscription, several words, such as divos (A, l. 2), nolis (A, l. 3) and advoc[atus?] (A, l. 7) can be made out. If the reconstruction of this last lexeme is correct, it appears that this curse should be grouped with the juridical defixiones.

Commentary: this defixio was discovered alongside another curse tablet (today at M. Bargoin, and still unpublished) during the second season of the excavations carried out in the Gallo-Roman necropolis in Les Martres-de-Veyre, which Audollent directed in 1923. Even though Audollent identified the tablets as curses (Audollent, apud RIG II.2, p. 285), he did not provide

162. Murol Provenance: sanctuary. Current Location: unknown. Date: unknown. Material: lead. Measurements: 5.4 × (3.7) cm. Reading: A deus te rog[o qui] dịṣtrale [---] ++ NTRA +[---] [---] +C[---] 5 E+GA[?---] [c]olas se N[---] NV+QV (traces)[---] B contra[---] Martis A[---] vel terti[o---] ut confet[eat] 5 [quod] tulit torq[uem] BES r suis IICR[---] E+ et festul[a-] domine vind[ica?---]

Bibliography: Verdier 1963: 243–44; Tisserand 1985: 74– 76; Verdier 1987: 137; CAG 63–2: 247; Kropp 2008: no. 4.3.3/1; Urbanová 2018: no. 225.

series of coins dated between the first century BCE and the fourth century CE as well as an array of other votive offerings.

Image source: Tisserand 1985: 76.

The curse was written on a rectangular lead sheet, which is missing its right edge. An opisthograph, the inscription contains 15 lines, which run from left to right and were written by a poorly trained hand. The text is written in capitals. Despite the fragmentary nature of the text, the inscription’s content suggests that it should be grouped among the defixiones in fures. This is suggested by the presence of the verb tollo (tulit, B, l. 5) as well as the mention of both a dịṣtrale (A, l. 2; for dextrale; cf. 220) and torques (B, l. 5), which are presumably the purloined objects. Accordingly, the mention of a third part (terti[o], B, l. 3) could be taken as a donation of part of the stolen

Commentary: this defixio was discovered during the excavations of a sanctuary located in the plateau of Rajat (Murol, Puy-de-Dôme) between 1954 and 1957. During the various seasons of excavation, a rectangular area was singled out that was walled off on the eastern and western sides. Archaeologists concluded that the remains were part of a sanctuary, which contained two small structures (the cellae). Near the wall and still within the limits of the sacred space, archaeologists found not only the present defixio, but also many Gallo-Roman ceramic sherds, a 242

Galliae Britannia; cf. e.g., 205 and 356). In such curses, the deity is typically asked to punish the wrongdoer and avenge the theft (vind[ica?] B, l. 8; for parallels, see 120, 240, etc.).

goods’ value to the deity invoked, who is mentioned at both the beginning and end of the curse (e.g., 128 and 451); this formula, however, has not been attested yet outside of

163. Arverni, Chamalières (Puy-de-Dôme) Provenance: sanctuary. Current Location: M. Bargoin. Date: first half 1st century CE. Material: lead. Measurements: 4 × 7.1 × 0.1 cm. Reading:

5



10



andedíon uediíumi diíiuion risunartiumapon arueriíatin lopites sníeđđic sos brixtía anderon c lucion floron nigrínon adgarion aemilí on paterin claudíon legitumon caelion pelign claudío pelign marcion uictorin asiatí con ađđedillí etic se couitoncnaman tonc siíontío meíon poncse sit bue tid ollon reguccambion exsops pissíiumí tsoc cantírtssu ison son bissíet luge dessummiíis luge dessumíis luge dessumíís luxe

Bibliography: Poursat 1973: 444; Lejeune and Marichal 1977: 156–78; Fleuriot 1977; Lambert 1979; Fleuriot 1980; Fleuriot 1981: 106–07; Schmidt 1981; Henry 1984; Lejeune 1984: 703–08; Fleuriot 1987; Kowal 1987; Lambert 1987; Bourgeois 1991: 134–36; Lambert 1994: 150–59; CAG 63/2: 075-002; Evans 1996; Lambert 1996a: 51–65; Lambert 1996b: 98–103; Eska 1998; Romeuf 2000: 40; RIG II.2, L-100; AE 2002, 983; Romeuf 2007; AE 2007, 941; Mees 2007; Mees 2009: 10–21.

like many defixiones: it begins with an invocation of a deity and then continues with a list of victims, which in this case is rounded off with an all-inclusive formula and the unleashing of certain illnesses. The curse ends with a final reiterative formula (cf. 26). The text provided here follows the reading found in RIG II.2. The first two lines are arranged like a heading (they are centred and written in bigger letters) and consist of the invocation of the sanctuary’s titular deity and the infernal powers (‘Au nom de la bonne force des divinités chtoniennes, j’invoque Maponos d’Avernion’, translation by Lambert apud RIG II.2, p. 276). The theonym -mapon(on) (l. 2), which is built from the root *ma(k)kwo- (‘young man’) with the addition of the suffix -ono-, doubtlessly refers to Maponos, a god who was assimilated into Apollo (see Fleuriot 1977: 178–79). The word arueriíatin (l. 2), which accompanies the theonym, is an epithet meaning ‘infernal’ (RIG II.2, p. 275). In the first line, andedíon is built on the root *ṇdhi- (‘under’) and therefore ‘ici «inferieur» est à entendre comme «chtonien»’ (RIG II.2, p. 275). According to Lambert, vediíumi (l. 1) is a verb derived from the root *wedh- (‘to know’ or ‘to make known’), while diíuion is an adjective meaning ‘divine’. Finally, Lambert has translated risun-/artiu (ll. 1–2), a compound of the preposition ris, su and nartio, as ‘au nom de, ou par, la bonne force’ (RIG II.2, p. 275).

Image source: R. Marichal apud RIG II.2, L-100 (fig. 150). Courtesy of P.Y. Lambert. Commentary: this defixio was discovered in 1971 during the excavations of ‘Source des Roches’, a Gallo-Roman sanctuary where there was an aquatic cult that appears to have been dedicated to the deity Maponos. Besides the present curse, a large series of votive offerings have been found in the same stream: three thousand wooden offerings, coins from the Julio-Claudian period, ceramics and several fibulae (for the archaeological context, see Poursat 1973 as well as Romeuf 2000 and 2007). The curse was written on a tabella cum capitulo. This shape is typical in common writing practices and here has been adapted to a magical purpose; so much as is strongly suggested by the use of lead (for parallels, see 127–28). The inscription, written in Gaulish, contains 12 lines which run from left to right. Written in old Roman cursive, they have been dated to the middle of the first century CE on palaeographic grounds. Though it has proven difficult to interpret the inscription, it is clear that the text is structured

After this initial invocation, the curse next delivers the victims to the infernal powers with the phrase lopites sníeđđic sos brixtía anderon (l. 3), which Lambert translates as ‘Poursuis… ceux-là avec la magie des Infernaux’ (apud RIG II.2, p. 276). Here, brixtía anderon 243

Sylloge of  Defixiones from the Roman West The list of names ends with an all-inclusive formula: etic se couitoncnaman tonc siíontío (ll. 7–8), which Lambert translates as ‘et tous ceux qui seraient des ces couipo-, de ces enemies’ (apud RIG II.2, p. 279). In this formula, etic means ‘and’, while naman-/to is ‘enemy’. Next, the text provides a series of curses meant to harm the victim’s well-being. The first of these reads meíon poncse sit buetid ollon reguccambion (ll. 8–9), which Lambert has rendered as ‘si c’est [le parti vainqueur], alors, que ce soit pour eux la complète déformation des os droits’ (RIG II.2, p. 279). Here, ollois an adjective meaning ‘big’, while reguccambion is a compound made of *regut- ‘long bone’ and cambo‘curved’. The second curse in this series, which reads exsops pissíiumi (ll. 9–10; translated ‘aveugle je verrai’) refers to the blindness that the defigens wants his/her victims to suffer (for a parallel, see 250). The next phrase, tsoc cantírtssu ison son bissíet (ll. 10–11), has proven more difficult to interpret. The word cantírtssu is made up of the preposition cante (‘with’) and the root ret- (‘to run’) and has been interpreted as ‘to catch’ or ‘to trap’. Finally, the curse concludes with the phrase luge dessummiíis luge dessumíis luge dessumíís luxe (ll. 11–12), in which the elements luge (from *lug- ‘to place’ or ‘to put’) and dessum(m)íis (‘right’) are repeated three times in quick succession.

is especially noteworthy: the first lexeme contains the root *brig-, which is related to Gaulish –brith ‘magic’, while the second is derived from anderos, equivalent to Latin inferus. This sequence of words is important given that there is a parallel found in 158: brictom andernados (I. A, l. 1). In both cases the phrase is found in the same position (i.e., right before the names of the victims). In this case, the list contains seven names that are all attested elsewhere in Gaul (six are Latin and one Celtic). Each name has the Gaulish accusative ending –on. Within the list of names, there are two main types of onomastic formulae that are used. The first individual in the list is the most thoroughly named with a praenomen, nomen and two cognomina: C Lucion Florion Nigrínon (l. 4). Another three victims (Aemilíon Paternin(on), Claudíon Legitumon and Marcíon Victorin(on), ll. 4–6) are named with a bipartite formula consisting of nomen and cognomen. The two remaining victims (Caelion peign(on) and Claudío pelign(on), ll. 5–6) are named with a bipartite formula consisting of a nomen and the the word pelign(on), which is related to Celtic pell (‘far’) and hence has been interpreted as meaning ‘foreigner’. Lambert has further drawn attention to the lexeme adgarion (l. 4), which follows the name of the first victim. According to his analysis, the word is of vital important: being a compound with *gar (‘appeal’ or ‘demand’), this word suggests that this curse could be a juridical defixio.

164. Aquae Tarbelliacae, Dax Provenance: aquatic context. Current Location: M. de Borda. Date: 4th– 5th centuries CE. [Inv. No.: 2003.14.1] Material: lead. Measurements: 6.5 × 4.7 × 0.2 cm. Reading: Col I. Col. II Leontio bo + Leontio lave + D{e}idio runt Iovino manus 5 pedis quicumqu˹e˺ le[vavit?] an(n)ue[---] [o]culi qui + I++[---] imm[e]rgo I[---] RV++[---] 10 E+[---]

244

Galliae Bibliography: Watier and Gauthier 1977; Gibut 1998; Marco Simón and Velázquez 2000; AE 2000, 925; Kropp 2008: no. 4.3.2/1; ILA Landes, 3; Urbanová 2018: no.224.

Based on its content and the verb involo (II, ll. 1–3, see below), this curse should probably be grouped among the defixiones in fures. The text presents the following orthographic and phonetic features: bolaverunt for volaverunt (II, ll. 1–3), Deidio for Didio (I, l. 3), pedis for pedes (l. 5), quicumqui for quicumque (l. 5) and anue for annue (l. 6). Forms of the verb involo (here in II, ll. 1–3) are quite common in the defixiones in fures (e.g., 220, 271, 304, etc.). The verb imm[e]rgo (l. 8) is apt given the aquatic context in which the curse was discovered (Marco Simón and Velázquez 2000: 269). Even if the text is too fragmentary to be reconstructed in full, the verb immergo (‘to sink’) is probably part of a similia similibus formula with which the act of deposition was compared to the final fate of the victims (for a similar case, see 157). The text also contains an anatomical sequence (II, l. 4 and ll. 5–6), which is used to attack various parts of the victim’s body (here the hands and the feet). After listing potential suspects (I, ll. 1–4), the actual thief appears to be referred to with the phrase quicumqu˹e˺ le[vavit] (ll. 5–6; for this restoration see Marco Simón and Velázquez 2000: 269 and cf. 249, 350, among others.). All names in the curse are common: Leontio (ll. 1 and 2) is of Greek origin (see OPEL III, 23) and Deidio (l. 3, for Didio, see OPEL II, 99).

Image source: Marco Simón and Velázquez 2000: plate 2. Courtesy of F. Marco Simón. Commentary: this defixio was discovered in 1976 when the Fontaine Chaude of Aquae Tarbelliacae/Dax was being surveyed. Inside, workers excavated a portion of a monumental Roman spring, which was built in the second century CE and filled in sometime in Late Antiquity. It was inside of this spring that the folded defixio was originally deposited alongside ceramic sherds and coins dating to the fourth and fifth centuries CE (for the archaeological context, see Watier and Gauthier 1977). The curse was written on a lanceolate lead sheet, now broken into two pieces that fit together. The surface, which shows signs of being hammered, has been cracked in the areas around the fold lines. The inscription contains 10 lines, which run from right to left (for a similar layout, see e.g., 125–26, 464, etc.) and were written in new Roman cursive (measuring between 0.2 and 1 cm in height). The text has been dated to either the fourth or fifth century on palaeographic grounds. The first portion of the text is arranged in two columns of four lines. Starting from l. 5, however, there is only one column. It is quite probable that I, ll. 1–4 were written first based on the size of the letters and the way in which the rest of the text seems to conform to these lines.

Translation: ‘Leontius, Leontius, D(e)idius Iovinus, have stolen... The hands, the feet of whoever has lifted (?)... his eyes... I immerse...’

Gallia Lugdunensis 165. Lutetia, Paris Provenance: necropolis. Current Location: unknown. Date: 4th century CE. Material: lead. Measurements: (8.5) × (15) cm. Reading:

XIRINNALLISOLICIOM SOCSINCONOIOSVOILSOCI ASVNNA (vacat) MOIV

Bibliography: Vacquer 1879; CIL XIII, 3051; DT 108; CIL XIII, 10029, 328; Duval 1960: 150–51; Tomlin 1987b: 25, n. 3; RIG II.2, *L-105; Mees 2009: 98–99.

Commentary: this defixio was discovered during the excavations that Vacquer carried out in Boulevard Arago in Paris. During the 1847 season, archaeologists uncovered two graves which were ‘selon le rite chrétien’ (Vacquer 1879: 111) and dated to the reign of Constantine. In one

Image source: Vacquer 1879: 113.

245

Sylloge of  Defixiones from the Roman West of these graves, excavators found the present curse, folded in half and lodged in the ribcage of one of the deceased. The curse entered into the editor’s personal collection but was subsequently lost. Today, only a plaster cast of the text exists, which is housed on the Musée d’Archéologie Nationale (Saint-Germain-en-Laye, Inv. 31658).

to Duval 1960: 150, who worked with the plaster cast of the curse tablet, the inscription contains three lines, which run from left to right and whose letters measure between 0.4 and 1.2 cm in height. The text has proven exceedingly difficult to interpret: while the editor had picked out several Gaulish words, Lambert has maintained that the presence of the syllables so or soc are not enough to identify the text as Gaulish (2002: 301). Though he has not offered a full reading of the text, Tomlin has suggested that the curse consist of a list of personal names written from right to left. Accordingly, the xir– found at the beginning of the inscription would be the common Gaulish ending –rix.

The curse was written on a roughly rectangular lead sheet, which was in rather poor condition (Vacquer 1879: 111– 12): not only had it lost part of the bottom and right edges, but also had been broken along the centre of the tablet, along what appeared to have been a fold line. According

166. Autun Provenance: necropolis. Current Location: M. Rolin. Date: 2nd century CE. Material: lead. Measurements: 15 × 4.8 × 0.1 cm. Reading: A (crux) B Col. I Onesiforus Musc(u)losus Carpus Attianus 5 nepos Veracis Titus



Col. III (charaktêr) αβρασαξ (charaktêr ) δαμναμευς κομπωθ θιφεριθ γωματου σαβαλθωυθ βισωτορθ δεθερθ

Col. II αβρασα

Bibliography: Marcillet-Jaubert 1979a; BÉp 1979: 662; AE 1979, 407a; Marcillet-Jaubert 1979b; SEG 29, 1028; BÉp 1980: 577; Rebourg 1987: no. 545; Frezouls 1997: 107; Decourt 2004: 159; Kropp 2008: no. 4.2.1/1; Urbanová 2018: no. 65; Gordon 2020b: 196.

to the second century CE based on palaeographic grounds. Nevertheless, it seems more likely that the text was written by a single individual, probably a professional practitioner with specific knowledge (as the use of charaktêres and the mixing of Greek and Latin suggest; contra Gordon 2020b: 196). We do not know the circumstances that led to the writing of this curse. The Latin portion of the text, written in capitals, presents the following list of cognomina in the nominative (presumably the curse’s victims): Onesiforus (l. 1, a Latinized form of Onesiphorus; the name suggests a servile origin (see Solin 2003: 131), Musclosus (l. 2; for Musc(u)losus (see OPEL III, 91)), Carpus (l. 3; see OPEL II, 38), Attianus nepos Veracis (who is the only individual defined by a kin relationship, ll. 4–5; see OPEL I, 208 and OPEL IV, 156, respectively) and Titus (l. 6).

Commentary: this defixio was discovered in c. 1954 during the restoration of the so-called pyramid of Couard, a Roman funerary monument that dates to the first century CE and is located along the road connecting Lugdunum and Augustodunum. The tablet was found deposited at the monument’s base and was still folded. The curse was written on a strip of lead with a wrinkled surface and has lost its top-right and bottom-left corners. An opisthograph, side A contains a cross-like symbol measuring 4.5 cm in height (cf. 259 and 448), while side B contains the defixio proper. The text is made up of 15 lines and has an unusual layout: the first section of the text is written in Latin (ll. 1–6, in capitals); in contrast l. 7 is written in Greek and found in the centre of the sheet, right under the fold line. Finally, to read part 3 (ll. 8–15), which is also written in Greek, the tablet must be rotated 180 degrees and then the text runs down towards the fold line. The inscription, which according to the editor was written by two different hands (1979b: 11), can be dated

The Greek portion of the text (ll. 7–15) consists of a series of voces magicae, most notable of which is the invocation of Abraxas (ll. 7 and 8), whose name appears between two charaktêres made of a x with circles at the ends of each line. In the curses from Gaul, Abraxas appears only here. Marcillet-Jaubert has connected the word δαμναμευς with δαμναμενευς (already documented amongst the Ephesia grammata; cf. section I.4.2.2.3). After being inscribed, the tablet was folded. 246

Galliae 167. Mediolanum, Evreux Provenance: necropolis. Current Location: M. Evreux. Date: 1st–2nd centuries CE. Material: lead. Measurements: vid. reading. Reading:

Frag. I ((4) × (5.4) cm) [---]+++OQR+LIVI[---] [---]++ ut im(m)in[es? or -utio? ---] [---]+RE filiuṃ[---] [---]+END+[---]



Frag. II ((4.2) × (4.9) cm) [ali]ud(?) ver[---] [---]+tibi CVR[---] [---]TIFIPRO+++[---]



Frag. III ((1.6) × (1.8) cm) [---]++[---] [---]ERE[---]



Bibliography: Boüard 1962: 423; Vipard 1987: 380–83, no. 173; Carré and Cliquet 1993: 17; Kropp 2008: no. 4.2.2/1.

Frag. IV ((2.2) × (2.1) cm) [---]O++++[---] [---]+ana+[---] [---]vana[---] Frag. V ((2.4) × (4.2) cm) [---]quisq[ue---] [---]++invitum+[---]

The curse, now in five fragments, contains an inscription written in capitals that runs from right to left (Frag. I–IV) and in boustrophedon (Frag. V). For the first fragment, which contains four lines, Vipard has suggested the following possible reconstructions: im(m)inere, im(m)inens and im(m)inutio. It is worth noting that none of these would be paralleled in the larger corpus of defixiones from the Roman West. The rest of the fragments contain two (Frag. II and IV) or three lines (Frag. III and V). Unfortunately, only a few words can be isolated.

Commentary: in 1958, M. Le Pesant discovered this defixio alongside the fragments of three other tablets (168–70) in the Gallo-Roman necropolis located on rue de Bellevue in the Clos-au-Duc neighbourhood of Evreux. The grave goods uncovered in the necropolis consist of ceramic sherds and pieces of bronze and iron that have been dated to the Neronian period (for the archaeological context, see Boüard [1962: 423] and CAG 27: 288-VI, p. 140).

168. Mediolanum, Evreux Provenance: necropolis. Current Location: M. Evreux. Date: 1st–2nd centuries CE. Material: lead. Measurements: vid. reading. Reading:

Frag. II ((2.5) × ((4.3) cm) ----- Frag. III ((1.1) × (0.5) cm) [---]++[---] [---]si[---]

Frag. I (4 × 5.4 cm) -----[---]Optimi Iovi (?)[---] [------] [------]

Bibliography: Boüard 1962: 423; Vipard 1987: 380–83, no. 174; Carré and Cliquet 1993: 17; CAG 27: 288-VI, p. 140; Kropp 2008: no. 4.2.2/2.

The grave goods have helped date the necropolis to the Neronian period. The curse, now broken into at least three fragments, contains an inscription of at least eight lines, which run from right to left and were written in capitals. If the editor’s reading is correct, the mention of Jupiter is quite unusual within the corpus of defixiones from the Roman West (for a parallel, see 349).

Commentary: in 1958, M. Le Pesant discovered this defixio alongside the fragments of the tablets 167, 169– 70 in the Gallo-Roman necropolis located on rue de Bellevue in the Clos-au-Duc neighbourhood of Evreux.

247

Sylloge of  Defixiones from the Roman West 169. Mediolanum, Evreux Provenance: necropolis. Current Location: M. Evreux. Date: 2nd–3rd centuries CE. Material: lead. Measurements: vid. reading. Reading: Frag. I (3.7 × 8.2 cm) ----- Frag. II ((1.3) × (1.7) cm) [---]ACA++[---] [---]+++++[---]



5

Frag. IV ((1.7) × (2.9) cm) [---]+[---] [------] [---]AM[---] [---]++++++[---] [---]+++[---]

Frag. III ((1.1) × (0.7) cm) [---]+[---] [---]+A+[---] [---]+[---]

Bibliography: Boüard 1962: 423; Vipard 1987: 380–83, no. 175; Carré and Cliquet 1993: 17; CAG 27: 288-VI, p. 140.

grouped together based on shared palaeographic characteristics (1987: 380). Judging from the available drawings and photographs, the text was written in old Roman cursive. Frag. I, which preserves the bottom and right edges of the tablet, contains at least 10 lines, though the published images do not clearly confirm this. The rest of the fragments have two (Frag. II), three (Frag. III) and five (Frag. IV) lines. No words have been satisfactorily reconstructed.

Commentary: in 1958, M. Le Pesant discovered this defixio alongside the fragments of three other tablets (167–68 and 170) in the Gallo-Roman necropolis located on rue de Bellevue in the Clos-au-Duc neighbourhood of Evreux (for the archaeological context, see 167). The tablet is now in five fragments that Vipard

170. Mediolanum, Evreux Provenance: necropolis. Current Location: M. of Evreux. Date: 2nd–3rd centuries CE. Material: lead. Measurements: vid. reading. Reading:



Frag. I ((1.2) × (1.2) cm) [---]RI[---] [---]+++C+[---] Frag. II ((1.5) × (1) cm) [---]+[---] [---]M++[---] [---]+[---]

Bibliography: Boüard 1962: 423; Vipard 1987: 380–83, no. 176; Carré and Cliquet 1993: 17; CAG 27: 288-VI, p. 140.

Frag. III ((1.9) × (1.3) cm) [---]+++[---] [---]+C+[---] [---]+++OLVI[---] [---]+[---] Frag. IV ((0.7) × (1.5) cm) [---]N+[---] [---]++[---]

three other tablets, 167–69, were discovered (for the archaeological context, see the commentary for 167). Since these pieces are quite deteriorated and hard to identify, Vipard has admitted that they could in reality belong to 169. The fragments contain two, three, four and two lines respectively. It has not been possible to reconstruct a single word in these fragments.

Commentary: according to Vipard (1987: 380), these four fragments belong to a fourth defixio from the Gallo-Roman necropolis located on rue de Bellevue in the Clos-au-Duc neighbourhood of Evreux, where

248

Galliae 171. Vindinum, Le Mans (Sarthe) Provenance: pit. Current Location: DAD. Date: unknown. [Inv. No.: 86.72.181.24.2F32.0.3] Material: lead. Measurements: unknown. Reading:

5

+IALOS DM++++(?) IITS NAIOS V+++(?) `SDO+++´ TS Carantio BT(?)++ V COMA+AIO(?)+++

Bibliography: RIG II.2, L-104; Mees 2009: 100–01.

old Roman cursive. The first one (side B) is undeniably a commercial record as is clear from the use of numerals and the symbol for denarii (see RIG II.2, p. 298). The second inscription (on side A), which is arranged perpendicularly to the defixio, is practically illegible, though the presence of the ending -entes suggests that the text is in Latin. The final inscription (reproduced here) was written in capitals and contains five lines which run from right to left. Lambert (apud RIG II.2) has read the name Carantio (l. 4, for Carantius (?), cf. OPEL II, 35) and perhaps Ialo (l. 1, for Iallus (?) cf. OPEL II, 188) as well as the Latin formula D(is) M(anibus) (in l. 1). The use of such an abbreviation is odd (and unattested in the corpus of Latin defixiones), and only further complicates the interpretation of this text.

Image source: RIG II.2, L-104 (fig. 155). Courtesy of P.Y. Lambert. Commentary: this defixio was discovered during the emergency excavation of a ritual pit dated to the first century CE. Despite the difficulties inherent in reading and interpreting this text, it has been included in the present catalogue due to the item’s material characteristics and its archaeological context (a pit similar to a Roman favissa; see 341 and 353 for a parallel). This opisthographic tablet, which is roughly rectangular, contains three distinct inscriptions, each of which was written by a different hand. Two of these inscriptions were written in

172. Autricum, Chartres Provenance: domestic context. Current Location: unknown. Date: 1st century CE. Material: lead. Measurements: see reading. Reading:

5



10

Tab. I (5.2 × 7.9 × 0.2 cm) Vatumaros Senouiri · Cantognatos Virato · Conbarilos · Aberxtobogii · Raros · Esuatexti · Toutisa · Alignati {o} +ẹṭọ · cantipisontas · Sondios · adgario · duti · so · adogarie olusami Locuardicni · tascouidus adgariontâs · eti · so · adgarie Minio Mario

249

Sylloge of  Defixiones from the Roman West

5



10

Tab. II (5.4 × 7.5 × 0.2 cm) Tasguni Abrestubogiu Paulo Tasgunias · Cornuto · Cornilio Lami · Eponicnos · eto · cantipisontas · eti · Cornuti · Toutisa · Vi(m)pịḷi · Felicẹ · uo · Contestạ · duti · so · cantigarie · Sondiobi

Bibliography: Lambert 2014; Stüber 2014; Viret, Maqueda and Willerwal 2014.

eto cantipisontas (I, l. 5 and II, l. 5), the interpretation of which we will return to momentarily. The curse contains bipartite onomastic formulae, which follow different models (for a similar phenomenon, see 163).

Image source: Lambert 2014: 137 and 139. Courtesy of P.Y. Lambert.

The Celtic formula, which consists of the victim’s name in the nominative and his father’s in the genitive, is found in the first list. In the second list, there are two different ways of identifying victims: the first type of onomastic formula consists of two personal names in the nominative case (e.g., Minio Mario, I, l. 10); the second, a variation of the Gaulish structure found in list one, contains the father’s name in the genitive followed by the victim’s name in the nominative: Tasguni Abrestubogiu(s) and Lami Eponicnos (II, ll. 1 and 4, respectively). While this type of inversion is found in other Gaulish texts (as the one from Briona; for further references see Lambert 2014: 144), it must be stressed that it remains quite rare. Given the rarity of this onomastic formula, Lambert (2014: 144) has suggested that it be taken as a sort of magical inversion. However, this seems unlikely, given both other parallels in texts that are not magical and the presence of the more typical structure in an earlier part of this inscription. And indeed, between these two victims another Gaulish victim is identified with a formula that puts the genitive second: Paulo(s) Tasgunias (II, l. 2), though it must be said that in this example we find a matronymic instead of the more frequent patronymic. The third list also uses bipartite formulae to name the victims except for the first, Cornuti (II, l. 6), which the editors claim is an isolated name in nominative (for an indepth discussion of this curse’s onomastic formulae, see Stüber 2014). The names Vatumaros (I, l. 1) and Raros (I, l. 4) are found here for the first time.

Commentary: this Gaulish defixio was discovered in a ditch running along a path in Fille-Dieu (Chartres, ancient Autricum). The curse was found in a stratum dating to the end of the first century BCE, though archaeologists have questioned whether this was where the curse was originally deposited (perhaps it was accidently moved at some point). Whatever its precise context or strata may have been, the defixio was deposited somewhere near a ceramic workshop, a fact that has led the excavators to believe that it was intentionally placed near where the victims worked (for this question, see Viret et al. 2014: 130–31). This curse is actually made up of two separate lead tablets, which are roughly rectangular, similar in size and were folded together (for curses written on more than one sheet, see 158 and 160). Both parts of the defixio are in good condition, although the inscribed area is completely covered with a white layer that has made reading the tablet more difficult. The inscription contains 20 lines, which run from left to right and were written in capitals. The palaeographic characteristics of the text suggests a date of the first century CE and hence slightly later than the stratum in which it was found (see above). The words are separated with interpunctuation found halfway up the letters. The text presents the following orthographic and phonetic features: omission of implosive nasal and final –s (especially in the -os termination in masculine names) and the use of the cluster -st- for -xt- (see II, ll. 1 and perhaps 8).

Lambert has argued that the text should be grouped among the juridical defixiones based on the use of cantipisontas (I, l. 5 and II, l. 5): this lexeme is found at the end of the first two lists and contains the root*kwis, which means ‘to see’. Thus, he has suggested that this accusative plural should be translated as ‘qui sont présents, qui sont témoins’ (2014: 150). If his translation is correct, this would be an allinclusive formula which refers to all those who witnessed

The text is structured around three lists of personal names (the first at I, ll. 1–5; the second running from I, l. 10 to II, l.4; the third at II, ll. 6–8). In these lists, there is generally one name per line, which makes it much easier to identify each victim. The first two lists both contain the sequence 250

Galliae the denounced action (cf. with 302 B, ll. 3–4: qui medius fuerit). Lambert has analysed the phrase sondios adgario (I, l. 6) in a similar vein: apparently, it alludes to the first member of the second list and should be translated as ‘c’est celui-là qui est l’accusateur.’ Accordingly, adgario would be derived from *adgariio- (just like the words adgariontâs, adgarie and cantigarie, I, l. 9 and II, l. 9, respectively).

elusive: Lambert has connected it to Old Irish ollam but does not settle on a clear meaning: it could refer to either to judge presiding over the case or to the defigens (see Lambert 2014: 155 and 159). Finally, Lambert has taken tascouidus (I, l. 8) as a compound of tasco, which in this context would refer to a letter and hence in this content the defixio itself (Lambert 2014: 156–57). This hypothesis finds some parallels in the Latin curse tablets that call themselves carta (cf. 96, 484).

Lambert’s analysis of the verb adogarie (I, l. 7; for aduogariet) is of note: connecting it to *ar-uo-(g)ar, he has translated it as ‘to put a spell on’, which would undoubtedly be appropriate in this context (see Lambert 2014: 153–54). The word olusami (I, l. 7) has proven more

After being inscribed, the curses were placed together with the inscribed faces of the tablets on the inside and then folded in half (Viret et al. 2014: 132–33).

Gallia Belgica 173. Maar Provenance: necropolis. Current Location: unknown. Date: 2nd century CE. Material: clay jar. Measurements: 17 (high) × 11 (diameter) cm. Reading:

(along the body of the jar) Artus fututor aprilis KLSIO(?) Art(um) ligo Dercomogni fututor(em) (just above the foot of the jar, upside down) ABCDEFGHIKLMNOPRRSTVXYZ

Bibliography: Lehner 1893; Wünsch 1897: 29; CIL XIII, 3, 1, 10008, 7; DT 103; Gering 1916; Jeanneret 1917: no. 103; Kropp 2008: no. 4.1.2/1; Urbanová 2018: no. 57.

CE) as well as a yellow clay jar on which the present curse was inscribed (see Lehner 1893, fig. 1). The use of a small clay jar as the surface on which to inscribe a curse is exceptional. The fact that the defixio was written ante cocturam recalls a recipe from the third to the fourth centuries CE found in the PGM (XXXVI, 187) that explicitly calls for an erotic spell of attraction to be incised on an unbaked piece of pottery (for a possible parallel, see the agoge from El Jem which is inscribed on a tegula discussed in Foucher 2000).

Image source: Lehner 1893, figs. 1–3. Commentary: in 1893 this defixio was found in the necropolis of Maar (near Maastricht) inside a cremation burial. In addition to the cinerary urn, archaeologists found a lamp, three coins dating to the reigns of Domitian, Hadrian and the Antonines (minted between 80 and 140 251

Sylloge of  Defixiones from the Roman West The inscription needs to be separated into two, based on the moment at which they were inscribed. The first part is ante cocturam and consists of an alphabetical sequence and the curse proper. The alphabet portion, which is found along the bottom section of the jar, was written in inverted letters and contains an error (P R R instead of P Q R). Scholars have generally interpreted this sequence as a means of reinforcing the spell’s power, but the truth is that in the Roman West, there are almost no parallels for this practice (for the only possible parallel, see 207). For the curse, and despite the differences in interpretation, there is scholarly consensus that the text is an erotic one based on the presence of the word fututor (‘copulator’, ll. 1–2). Here, the victim is first mentioned in the nominative. L. 2 has proven to be more difficult to restore. Even though we would expect an accusative together with the filiation of the victim, the fact is that the text does not conform to grammatical conventions and needs to be reconstructed to be properly understand: Art(um) ligo Dercomogni (filium) fututor(em) (i.e.: ‘I bind Artus the copulator, son of Dercomognus’). Alternatively, Dercomognus could also be taken as the name of Artus’ master. Dercomognus

is a Celtic name, while Artus is a Latin cognomen attested in Gallia Belgica (see OPEL I, 179). The second part of the text, which according to the editors of CIL ‘post argillam coctam manus alia adscriptit’, reads aprilis KLSIO (l. 1; or KESIO according to Lehner and CIL). That said, palaeographic analysis suggests that this part was written by the same hand that wrote the rest of the text (compare, for instance, the shape of r, s and k in both texts). Lehner has argued that aprilis is undoubtedly a personal name (see Kajanto 1982²: 219) and that the signs that follow are just an attempt at reproducing some letters, although they are meaningless (Lehner 1893: col. 205; cf. with 317–21). Even though subsequent editors have omitted this part of the curse, I think it still belongs to the defixio against Artus. Here (and instead of a name) we could take aprilis as a reference to a magical deadline (i.e., the month of April), which could have been followed by a more specific date. Nevertheless, the final KLSIO defies interpretation: even though KL looks like the abbreviation k(a)l(endae), we would still need a number for making full sense of the phrase.

174. Durocortorum, Reims Provenance: necropolis. Current Location: M. Saint Remi (?). Date: 3rd century CE. Material: lead. Measurements: 1–1.8 × 10.5 cm. Reading: Frag. I Frag. II Frag. III IILI VL [---]OTIILLCA[---] Bibliography: Wünsch 1897: 244; Jadart 1901: no. 2154; DT 107; Preisendanz 1933: 155.

magical ritual (on this, see section I.6.2.2 and Alfayé and Sánchez Natalías 2020). The grave goods that were found above the deceased’s left shoulder consisted of 14 coins dating to the Antonine period, among which archaeologists came upon ‘une feuille de plomb amincie au marteau et roulée sur elle-même’ (Jadart 1901: 66). The curse was written on a strip of lead that broke into three pieces as it was being unfolded. The text has proven quite difficult to read and interpret and has been considered practically illegible.

Commentary: this defixio was discovered in Les troit Piliers in 1894 during the excavation that Habert carried out at a Gallo-Roman grave. The burial contains the inhumed remains of an individual whose head, according to Jadart, ‘porte une exostose et d’autres difformités’ (1901: 66). Presumably, the damage to the skull explains its role in a

175. Augusta Treverorum, Trier Provenance: amphitheatre. Current Location: RLT. Date: 4th–5th centuries CE. Material: lead. Measurements: (8.2) × (7.6) cm. Reading:

Frag. I Primus [------] [------] [------]

252

Galliae Bibliography: Wünsch 1910: no. 11; Besnier 1920: no. 17; Preisendanz 1933: 153; García Ruiz 1967: no. 37; Kropp 2008: no. 4.1.3/1; Urbanová 2018: no. 58.

have been nailed up). That said, the editor was right to stress that we do not have the needed information to defend either theory (on this, see section I.6.5.3).

Commentary: this defixio was discovered alongside another 31 curse tablets (27 of which are included in the following pages, cf. 176–202) during the excavations carried out in the amphitheatre of Augusta Treverorum in 1908. Although information about the tablet’s archaeological context is scarce, it is known that they were found in the basement of the amphitheatre in the ‘eastern chamber’. The fact that the tablets were found at different levels within this chamber led Wünsch to hypothesize that curses were originally deposited either in the amphitheatre’s sand (where they would have been buried) or alternatively on the wooden basement’s posts (in which case they would

This curse was still folded up upon its discovery and was covered with burnt organic material. According to Wünsch (1910: 2–3) this layer of organic material can perhaps be explained as the result of the fire that followed the destruction of the building, instead of the result of a magical ritual. When the tablet was unfolded, it broke into seven fragments, of which only one was inscribed. It contains four lines, which run from left to right and were written in new Roman cursive. It appears that the first line includes the name of the victim: Primus. Unfortunately, the remaining lines are illegible according to the editor.

176. Augusta Treverorum, Trier Provenance: amphitheatre. Current Location: RLT. Date: 4th–5th centuries CE. [Inv. No.: 09, 927] Material: lead. Measurements: 8.7 × 18.5 cm. Reading:

quidquid ad`h´ịbẹṭ Rac ̣aṇus et adṿọ[ca]tụṣ (h)a[b]es

Bibliography: Wünsch 1910: no. 12; AE 1911, 148; CIL XIII, 11340, 6; Besnier 1920: no. 18; Preisendanz 1933: 153; Whatmough 1950: 711, VI; García Ruiz 1967: no. 38; Schwinden 1984b: 187–88; Kropp 2008: no. 4.1.3/2.

The presence of the noun adṿọ[ca]tụṣ (ll. 2–3) suggests that this curse should be grouped among the juridical defixiones. Here, both Racanus and his lawyer are targeted. It is worth noting the verb adhibeo (l. 1), which is not found elsewhere in the corpus of defixiones from the Roman West and which here may refer to the communication between Racanus and his lawyer before the court. In this simple text, the defigens hands over ((h)a[b]es, l. 3) to an unnamed deity the victim and his words. There are no known parallels for the name Racanus (perhaps for Ra(e)c(i)anus?; see Kajanto 19822: 153), unless is an error or mis-reading for Arcanus (?).

Commentary: this defixio was discovered in 1908 in the basement of the amphitheatre in Trier. The tablet was specifically found where the northern passageway meets the easternmost room (for the archaeological context, see the commentary to 175). The curse was written on a rectangular lead sheet, which is intact. The inscription contains three lines, which run from left to right and were written in new Roman cursive.

Translation: ‘Whatever Racanus communicates, and his lawyer, (you) hold (them).’

177. Augusta Treverorum, Trier Provenance: amphitheatre. Current Location: RLT (?). Date: 4th–5th centuries CE. Material: lead. Measurements: 8.2 × 18.5 cm. Reading:

quidqu[id---] ------

Bibliography: Wünsch 1910: no. 13; Besnier 1920: no. 19; Preisendanz 1933: 153; García Ruiz 1967: no. 39; Kropp 2008: no. 4.1.3/3.

Commentary: this defixio was discovered in 1908 in the basement of the amphitheatre in Trier. The curse was specifically found in a stratum of sand (for the

253

Sylloge of  Defixiones from the Roman West archaeological context, see the commentary to 175). Like the previous inscription, this curse was written on a rectangular lead sheet and begins with the word quidquid. Next, the text includes a series of random letters placed on the tablet, which Wünsch did not transcribe. Due

to the lack of any published image of the tablet, whose whereabouts are now unknown, it is not possible to provide a new reading. After being inscribed, the tablet was folded.

178. Augusta Treverorum, Trier Provenance: amphitheatre. Current Location: RLT (?). Date: 4th–5th centuries CE. Material: lead. Measurements: 11 × 7.4 cm. Reading:

Matrona

Bibliography: Wünsch 1910: no. 14; Besnier 1920: no. 20; Preisendanz 1933: 153; Whatmough 1950: 711, XII; García Ruiz 1967: no. 40; Kropp 2008: no. 4.1.3/4; Urbanová 2018: no. 59.

amphitheatre in Trier (for the archaeological context, see the commentary to 175). The inscription is written in capitals and contains the name of the victim: Matrona, which is a well attested Latin cognomen (for an example of another curse that only contains the victim’s name, see 368).

Commentary: this defixio was written on a lead sheet that was discovered in 1908 in the basement of the

179. Augusta Treverorum, Trier Provenance: amphitheatre. Current Location: RLT. Date: 4th–5th centuries CE. [Inv. No.: 09, 942] Material: lead. Measurements: 6.7 × 14 cm. Reading:

Ursus +++LICVS nec ̣ anue

Bibliography: Wünsch 1910: no. 18; CIL XIII, 11340, 13; Besnier 1920: no. 21; Preisendanz 1933: 153; García Ruiz 1967: no. 41; Schwinden 1984b: 187; Kropp 2008: no. 4.1.3/5.

right corner. Unlike most tablets from Trier, this curse contains both image and text. The curse’s iconography depicts the victim’s head with a hat or hood. Wünsch has suggested that this headgear could be the attribute of a gladiator.

Image source: Wünsch 1910: Tab. II, fig. 1.

The text is found below the picture of the head and consists of two lines of capitals that run from left to right. Following Wünsch’s interpretation, Ursus is to be identified as the victim (for the name, see Kajanto 1982²: 329). The second line has defied interpretation.

Commentary: this defixio was discovered in 1908 in the basement of the amphitheatre in Trier among the rubble found in the entrance to the western chamber (for the archaeological context, see the commentary to 175). The curse was inscribed on a small roughly triangular lead sheet that is damaged along the top-

After being inscribed, the tablet was folded.

254

Galliae 180. Augusta Treverorum, Trier Provenance: amphitheatre. Current Location: RLT. Date: 4th century CE. [Inv. No.: 09, 943] Material: lead. Measurements: 8 × 5.2 cm. Reading:

5

Ursus Ursula Martinianus Ursacia

Bibliography: Wünsch 1910: no. 19; CIL XIII, 11340, 11; Besnier 1920: no. 22; Preisendanz 1933: 153; Whatmough 1950: 711, VI; García Ruiz 1967: no. 42; Schwinden 1984b: 186–87; Schwinden 1996: 236; Kropp 2008: no. 4.1.3/6; AE 2009, 907; Urbanová 2018: no. 60.

curse was written on a rectangular lead sheet, which is intact and in excellent condition. The inscription, written in new Roman cursive by a trained hand, contains five lines, which run from left to right. We do not know the circumstances that led to the writing of this curse, which lists a series of personal names in the nominative: Ursus (also found in 179), Ursula (l. 2, see Kajanto 1982²: 330), Martinianus (ll. 3–4, see Kajanto 1982²: 212) and Ursacia (l. 5; see Kajanto 1982²: 329).

Image source: Wünsch 1910: Tab. II, fig. 2. Commentary: this defixio was discovered in 1908 in the basement of the amphitheatre in Trier (for the archaeological context, see the commentary to 175). The

After being inscribed, the tablet was folded.

181. Augusta Treverorum, Trier Provenance: amphitheatre. Current Location: RLT (?). Date: 4th–5th centuries CE. Material: lead. Measurements: 5.5 × (9) cm. Reading:

Matrimoni[a A]b[ae et] amicorum [A]ba reddat [pre-] tia damno[---]

Bibliography: Wünsch 1910: no. 22; AE 1911, 149; CIL XIII, 11340, 7; Besnier 1920: no. 23; Preisendanz 1933: 153; Whatmough 1950: 711, VII; García Ruiz 1967: no. 43; Kropp 2008: no. 4.1.3/7; Urbanová 2018: no. 221.

know the circumstances that led to the writing of this curse, which targets Aba (see OPEL I, 1) as well as her spouse and friends (for other extended curses, see 237, 246, 258 and 370). Though there are no known parallels in the corpus of defixiones from the Roman West, the expression reddat [pre]tia damno[rum] (ll. 3–4) clearly alludes to the wish that Aba pay for the wrong she had done.

Commentary: this defixio was discovered alongside an uninscribed tablet in 1908 in the basement of the amphitheatre in Augusta Treverorum (for the archaeological context, see the commentary to 175). The curse was written on a lead sheet broken along its left edge. The inscription contains four lines, which run from left to right (the absence of published images precludes us from knowing if the text was written in new Roman cursive or capitals). We do not

After being inscribed, the tablet was folded. Translation (Urbanová 2018: 269, modified): ‘[I commend/ accurse?] the marriage of Aba and friends (?), may Aba pay the price(?) of the wrongs (she has done?).’ 255

Sylloge of  Defixiones from the Roman West 182. Augusta Treverorum, Trier Provenance: amphitheatre. Current Location: RLT (?). Date: 4th–5th centuries CE. Material: lead. Measurements: 5.5 × 7.7 cm. Reading:

5



inimicum[---] [---]qui [------] [------] i[u]vate(?) [------] Marti et Diane

Bibliography: Wünsch 1910: no. 23; CIL XIII, 11340, 4; Besnier 1920: no. 23; Preisendanz 1933: 153; García Ruiz 1967: no. 44; Kropp 2008: no. 4.1.3/8; Urbanová 2018: no. 61.

The presence of the lexeme inimicus (l. 1) suggests that this curse is a juridical defixio. The editor has put forth the restoration i[u]vate (l. 5, from the verb iuvo, ‘to help, assist’). Although suggestive, so far there are no parallels for such a verb within the corpus of Latin defixiones. In the curse, Mars and Diana are both invoked. These two gods also appear together in 183 (Mars is found alone in 238, 302, 356–57, 400, 430 and 452, while Diana also appears in 343).

Commentary: this defixio was in 1908 in the amphitheatre in Trier in 1908 (for the archaeological context, see the commentary to 175). Inscribed on a lead sheet, the inscription contains seven lines, which run from left to right (the absence of published images precludes us from knowing if the text was written in new Roman cursive or capitals). Ll. 3, 5 and 6 are today illegible.

183. Augusta Treverorum, Trier Provenance: amphitheatre. Current Location: RLT. Date: 4th–5th centuries CE. [Inv. No.: 09, 930] Material: lead. Measurements: (14.3) × 9.3 cm. Reading: A [-c.1-2-]YIBALFOQOIRM[-c.1-2-] (charaktêres) YSP (magical sign) YDMXFVS INABIHTIARO vestro 5 [Di]anam et Martem vinculares ut me vi ndicetis de Ququma Eusebium in ungulas obligetis et me 10 vindicetis B pepostum Eusebium

256

Galliae Bibliography: Wünsch 1910: no. 24; AE 1911, 150; CIL XIII, 11340, 3; Besnier 1920: no. 25; Preisendanz 1933: 153; Whatmough 1950: 711, III; García Ruiz 1967: no. 45; Schwinden 1984a: 68–69; Schwinden 1984b: 185–86; Schwinden 1996: 233–34; Kropp 2008: no. 4.1.3/9; AE 2009, 907; Urbanová 2018: no. 222.

The second box begins with the inabihtiaro vestro (l. 4). Interpreting inabihtiaro has proven controversial. The editor read inabintiaro, which he took as an invocation of the deity Intarabus, who was worshiped in Gallia Belgica. This god was associated with Mars, who appears later in this curse paired with Diana (A, l. 5; for the same pairing, see 182). Inabihtiaro could also be taken as a corrupted form of inhabitatio, in which case it would refer to the deities’ dwelling. That said, the latter suggestion involves numerous corrections and also complicates the use of vestro. The practitioner uses the term vinculares (l. 6), which has been taken as an epiphet describing the gods’ ability to bind and constrain the curse’s victims (Urbanová); the word is otherwise unattested. The defigens goes on to ask that the gods take vengeance on his/her behalf (me vindicetis, A, ll. 6–7 and 9–10). Wünsch has rightly corrected Ququma as Cucuma (A, l. 7, see Kajanto 1982²: 344). The Greek cognomen Eusebius (A, l. 8 and B, l. 2) is well attested (see Solin 2003: 1317 and 1476). On analogy with 184, the phrase pepostum E(u)sebium (B, ll. 1–2), in which the letters in the second line are written upside down, can be corrected to depositum Eusebium. This phrase is used as a means to kill the victim (for depono in this sense, see Lewis and Short, s.v. IB).

Image source: Wünsch 1910: Tab. II, fig. 4 and Tab. III, fig. 1. Commentary: this defixio was discovered in 1908 together with 184. The tablets were found in the northern passageway of the basement of the amphitheatre in Trier 1.5 m below the surface (for the archaeological context, see the commentary to 175). Both texts were inscribed on the same type of lead tablet and, according to Wünsch, were composed by the same hand. This curse was written on a roughly rectangular lead sheet, which has lost its top and bottom edges. An opisthograph, the inscription contains 10 lines, which run from left to right and were written in capitals (but note how b, d and m can be also traced in new Roman cursive). The text is distributed in three rectangular boxes that were drawn on the tablet (for parallels, see 18 and 184). We do not know the circumstances that led to the writing of this curse. In the first box, Wünsch isolated the word Bal (l. 1), which he suggested to be the Semitic Baal (Wünsch 1910: 8), though this identification is rather dubious. Next, there is a series of charaktêres and magical symbols, which are found on another Greek defixio from the same amphitheatre (see 1910: no. 20).

After being inscribed, the tablet was folded. Translation (Urbanová 2018: 277, modified): ‘[I ask] the constraining Diana and Mars to take vengeance on Cucuma on my behalf (?), put Eusebius on the instruments of torture and avenge me. Eusebius has been put down.’

184. Augusta Treverorum, Trier Provenance: amphitheatre. Current Location: RLT. Date: 4th–5th centuries CE. [Inv. No.: 09, 929] Material: lead. Measurements: 9.3 × 7 cm. Reading: A E (magical signs, charaktêres) (magical signs) (magical signs) (magical signs) EE (charaktêres) (magical signs, charktêres) B Prụsiae nom en de posit[um]

257

Sylloge of  Defixiones from the Roman West Bibliography: Wünsch 1910: no. 25; CIL, XIII, 11340, 8; Besnier 1920: no. 26; Preisendanz 1933: 153; Whatmough 1950: 711, VIII; García Ruiz 1967: no. 46; Schwinden 1984b: 188–89; Schwinden 1996: 234–35; Kropp 2008: no. 4.1.3/10; AE 2009, 907; Urbanová 2018: no. 62.

inscription contains a series of charaktêres and magical symbols on side A, which are arranged inside of a drawn rectangle that resembles the boxes found in 183. Among these signs, the editor has noted the presence of six sevenpointed stars, which could be a representation of the seven planets (cf. 6).

Image source: Wünsch 1910: Tab. III, fig. 2.

Side B contains three lines, which run from left to right and were written in capitals. In these lines the name and with it, the victim herself, is killed (for depono in this sense, see Lewis and Short, s.v. IB; cf. 183 side B). Prusia is a Greek personal name (for a discussion, see Solin 2003: 665).

Commentary: this defixio was discovered along with 183 in 1908 in the northern passageway in the basement of Trier’s amphitheatre (for the archaeological context, see the commentary to 175). Both texts were written on the same type of lead tablet and probably by the same hand. The curse was written on an irregularly shaped sheet, which has lost part of its left edge. An opisthograph, the

Translation: ‘The name of Prusia has been put down.’

185. Augusta Treverorum, Trier Provenance: amphitheatre. Current Location: RLT. Date: 4th–5th centuries CE. [Inv. No.: 09,926] Material: lead. Measurements: 8.6 × 12.7 cm. Reading:

-----si tu (H)ostillạm q(ua)e + Racatia {frau} q(u)i · mihi fraude(m) fe[cit] deus nos te q(u)i audis[ti]

Bibliography: Wünsch 1910: no. 26; AE 1911, 151; Besnier 1920: no. 27; Preisendanz 1933: 153; Whatmough 1950: 711, V; García Ruiz 1967: no. 47; CIL, XIII, 11340, 5; Schwinden 1984b: 188; Kropp 2008: no. 4.1.3/11; Urbanová 2018: no. 185.

is qualified with a matronymic phrase q(ua)e + Racatia, l. 1; cf. 187), is accused of having somehow harmed the defigens (q(u)i mihi fraude(m) fe[cit], l. 2; similar phrases are well attested in Britannia [e.g., 237, 365 and 370]). The final phrase, which the editor has reconstructed as deus nos te q(u)i audis[ti sacrificio colemus] (l. 3), alludes to the defigens’ promise to offer the deity a votive offering when the victim is properly punished (for similar promise of reward, see 56, 121, 205, 443, etc.). The name Ostilla (l. 1; for Hostilla, see OPEL II, 186) is a documented Latin cognomen. There are no known parallels for the name Racatia (l. 1).

Commentary: this defixio was discovered in 1908 in the basement of the amphitheatre in Trier (for the archaeological context, see the commentary to 175). The curse was written on a rectangular lead sheet, which is intact. According to Wünsch, the tablet had been reused. So much is suggested by the remains of a cursive text located in the upper half of the tablet.

Translation (Urbanová 2018: 275, modified): ‘...if you, Hostilla, (who was born?) from Racatia (...) who did me wrong, god, we (will honour?) you, who answered...’

In contrast, the lower half of the tablet, which was inscribed after the top half, contains three lines of capitals, which run from left to right. Here, the curse’s victim, Hostilla (who

258

Galliae 186. Augusta Treverorum, Trier Provenance: amphitheatre. Current Location: RLT (?). Date: 4th–5th centuries CE. Material: lead. Measurements: 6 × 3.7 cm. Reading:

[------] [------] viro

Bibliography: Wünsch 1910: no. 27; Besnier 1920: no. 28; Preisendanz 1933: 153; García Ruiz 1967: no. 48; Kropp 2008: no. 4.1.3/12.

whereabouts are unknown. The curse was written on a small fragment of a lead sheet, which has a circular hole in one of its edges (perhaps it had been nailed up?). The inscription, which is hardly legible, contains three lines, which run from left to right (the absence of any published images precludes us from identifying the palaeographic features of the text).

Commentary: this defixio was discovered in 1908 in the amphitheatre in Augusta Treverorum (for the archaeological context, see the commentary to 175). Today, the curse’s

187. Augusta Treverorum, Trier Provenance: amphitheatre. Current Location: RLT (?). Date: 4th–5th centuries CE. Material: lead. Measurements: vid. commentary. Reading:

Nestoria -----quam pep[erit]

Bibliography: Wünsch 1910: no. 28; Besnier 1920: no. 29; Preisendanz 1933: 153; García Ruiz 1967: no. 49; CIL XIII, 11340, 2; Kropp 2008: no. 4.1.3/13.

for the reconstructed tablet are 9 × 6 cm). Although we do not know the number of lines in the inscription or its palaeographic features, there was clearly an onomastic formula that identified the victim with the matronymic phrase quam pep[erit]. This phrase is common in North Africa (e.g., DT 247, 250 and 253), but also in some Late Antique texts from Rome (see 20). Other victims from Trier are also identified by their matronymic (see 185 and 189). The Latin cognomen Nestoria (l. 1) is unparalleled.

Commentary: this defixio was discovered in 1908 in the amphitheatre in Trier (for the archaeological context, see the commentary to 175). Today, its whereabouts are unknown. According to the editio princeps, the tablet was broken into three parts that fit together (the maximum dimensions

188. Augusta Treverorum, Trier Provenance: amphitheatre. Current Location: RLT (?). Date: 4th–5th centuries CE. Material: lead. Measurements: 5.3 × 4.3 cm. Reading: A B ------ -----deus -----Bibliography: Wünsch 1910: no. 29; Besnier 1920: no. 30; Preisendanz 1933: 153; García Ruiz 1967: no. 50; Kropp 2008: no. 4.1.3/14.

archaeological context, see the commentary to 175). Today the curse’s whereabouts are unknown. This opisthographic inscription was written on a lead sheet that was beaten (perhaps with a hammer) to the extent that the text was rendered illegible. The curse was hit before it was deposited (for other parallels, see 432, 433 and 512).

Commentary: this defixio was discovered in 1908 in the basement of the amphitheatre in Trier (for the 259

Sylloge of  Defixiones from the Roman West 189. Augusta Treverorum, Trier Provenance: amphitheatre. Current Location: RLT. Date: 4th–5th centuries CE. [Inv. No.: 1909, 931] Material: lead. Measurements: 13 × 10 cm. Reading:

5



10



bona san(c)ta nomen piạ nomen NOEMNCLIA [-c.2-]OESÊSTVDENITIA tibi santne dia defigo [Ro]danum +VEN +E+ORIT Anula Regula FATTA ạer domina QVEA [-c.4-]+CE tanta famas tu (?) M[-c.2-]R++[-c.2-]RE[-c.2-] carnis Bonarium [-c.3- d]efigo ATT[-c.2-]A[-c.5-]TRATA +[-c.1-]TE+[-c.3-]+TI[-c.2-]++TVCI[-c.4-] TAI[-c.3-]TA[-c.2-]+OTVN

Bibliography: Wünsch 1910: no. 31; AE 1911, 152; CIL XIII, 11340, 1; Besnier 1920: no. 31; Preisendanz 1933: 153; Whatmough 1950: 711, 1; García Ruiz 1967: no. 51; Schwinden 1984c; Schwinden 1996: 235–36; Kropp 2008: no. 4.1.3/15; AE 2009, 907; Urbanová 2018: no. 63.

Although we do not know the circumstances that led to the writing of this curse, the text invokes a female deity in the first line bona san(c)ta nomen piạ (also see ll. 3–4: tibi santne dia (for sancte) and domina, l. 7; for parallels, see 10, l. 1: bona pulcra Proserpina and 215, ll. 3–4: dea sanctissima Sulis, etc.). Following Wünsch, the curse’s victim, [Ro]danus (l. 5, see Kajanto 19822: 203), could be qualified with the matronymic formula quem peperit (l. 5 with quen for quem; cf. 185, 187, etc.). That said, the reading of quem peperit is a problematic conjecture, since it cannot be confirmed by Wünsch’s drawing or the pictures kindly provided by L. Schwinden (RLT). The use of the verb defigo is noteworthy (ll. 4 and 10), since it is rarely attested in the corpus of Latin defixiones from the Roman West (for parallels, see 50, 105, 108, 338, 345 and 530).

Image source: Wünsch 1910: Tab. III, fig. 3. Commentary: this defixio was discovered in 1908 in the basement of the amphitheatre in Trier (for the archaeological context, see the commentary to 175). The curse was written on a rectangular lead sheet, which has lost a good deal of its left edge (especially on the top half) as well as much of its lower third. The inscription has proven extremely difficult to interpret due in large part to its poor condition. It contains a text of 13 lines which run from left to right and were written in capitals (but note the new Roman cursive shape of b, m, and s).

After being inscribed, the tablet was probably folded.

190. Augusta Treverorum, Trier Provenance: amphitheatre. Current Location: RLT (?). Date: 4th–5th centuries CE. Material: lead. Measurements: 8.3 × 7.6 cm. Reading: (traces) R (traces)

260

Galliae Bibliography: Wünsch 1910: no. 8.

along the upper-left and bottom-right corners by the time of the editio princeps. The inscription consists of a series of symbols and jumbled lines, which have not been satisfactorily interpreted.

Commentary: this defixio was discovered in 1908 in the basement of the amphitheatre in Trier (for the archaeological context, see the commentary to 175). Today the curse’s whereabouts are unknown. The curse was written on a rectangular lead sheet, which was broken

After being inscribed, the tablet was folded.

191. Augusta Treverorum, Trier Provenance: amphitheatre. Current Location: RLT (?). Date: 4th–5th centuries CE. Material: lead. Measurements: unknown. Reading: A B ESPE ++CIDSFR

Bibliography: Wünsch 1910: no. 9; CIL XIII, 11340, 14.

(for the archaeological context, see the commentary to 175). Today the curse’s whereabouts are unknown. An opisthograph, the curse was written on a strip of lead, which has four capitals on side A and the traces of six others on side B. The first letter on side A is ‘mirrored’ (i.e., written from right to left).

Image source: Wünsch 1910: Tab. I, figs. 2a and b. Commentary: this defixio was discovered in 1908 among rubble during the excavations of the amphitheatre in Trier

192. Augusta Treverorum, Trier Provenance: amphitheatre. Current Location: RLT (?). Date: 4th–5th centuries CE. Material: lead. Measurements: vid. commentary. Reading: A B +OPAMVN (?) (traces)+++ LR++E+++++ NCLE(?) (crux?)

Bibliography: Wünsch 1910: no. 10; CIL XIII, 11340, 10.

on a triangular lead sheet, which is 7 cm tall (for similar shapes, see 106 and 180). Side A of this opisthographic tablet contains two lines with a list of personal names (according to Wünsch, the end of l. 2 reads –cium; although this reading cannot be confirmed by the published drawing). Side B depicts a type of net, which can be compared to either 196 and 197, where there seems to be a fence-like object, or to be part of a mummy’s body, like the one found on 117.

Image source: Wünsch 1910: Tab. I, figs. 3a and b. Commentary: this defixio was discovered in 1908 during the excavations of the amphitheatre in Trier (for the archaeological context, see the commentary to 175). Today the tablet’s whereabouts are unknown. The curse was written 261

Sylloge of  Defixiones from the Roman West 193–202. Augusta Treverorum, Trier Provenance: amphitheatre. Current Location: RLT (?). Date: 4th–5th centuries CE. Material: lead. Measurements: vid. commentary. Reading:

See commentary.

Bibliography: Wünsch 1910: nos. 3, 4, 5, 6, 7, 8, 15, 16, 17, 21, 30 and 32.

In the centre of the tablet there is a hole, which suggests that the tablet was pierced after being inscribed.

Commentary: here follows a combined file that gives relevant information on some of the curse tablets that were discovered in the amphitheatre in Augusta Treverorum and that Wünsch edited (for the archaeological context, see the commentary to 175). Unlike 175–92, there is little published information about these texts. Given that the tablets’ whereabouts are unknown, it was not possible to conduct an autopsy or even check any images. For those reasons, in the present file, I give the reference to the editio princeps in parentheses, the dimensions (when they are known) and a brief commentary.

197 (= no. 7): 8 × 5.8 cm. Opisthographic lead sheet: on both sides we find a drawing similar to that found in curse tablet no. 192. After being inscribed, the tablet was pierced with a nail. 198 (= no. 16): 6.1 × 4.2 cm. Rectangular copper sheet. It was found folded and is uninscribed. Given the use of copper instead of the more habitual lead, the editor supposed that this item was in reality an amulet and not a defixio. That said, in the sanctuary of Anna Perenna (see 30 and 31) there are defixiones inscribed on copper sheets. Accordingly, it remains possible that this item is also a curse. 199 (= no. 17): 5 × 4 cm. Rectangular copper sheet that was found folded and is uninscribed. For the use of copper, see 198.

193 (= no. 3): 5 cm long. Inscribed lead sheet that is illegible. It was folded after being inscribed.

200 (= no. 21): 3.9 × 4.3 cm. Folded lead tablet that is uninscribed. Discovered together with 181.

194–95 (= nos. 4 and 5): two fragments of a lead sheet found together. They are still folded and contain the remains of an inscription.

201 (= no. 30): three fragments of a lead sheet. The largest measures (6) × (6) cm. It could be a part of 187.

196 (= no. 6): 7.5 × 5.5 cm. Lead sheet on which there is a series of overlapping lines. The editor interpreted this as a fence that symbolically impeded the movements of the defixio’s victim.

202 (= no. 32): 3 × 6.5 cm. Inscribed lead tablet that is illegible.

203. Dalheim Provenance: unknown. Current Location: MNHA of Luxembourg. Date: unknown. [Inv. No.: 1989-100/7] Material: lead. Measurements: (5.65) × (6.3) × 0.1 cm. Reading:

(magical signs) LEBA (magical signs)

262

Galliae Bibliography: Schwinden 1992; Krier 2011: 317.

right edge has been preserved. There is a break along the bottom third of the tablet that coincides with the fold line. The sheet contains an inscription of four lines composed by a series of charaktêres and some capitals that have defied interpretation (LEBA can be made out in l. 2). The charaktêres found in this curse are also attested in other defixiones from Gaul (e.g., 166, 183 and 184), where the names of the victims are inscribed. By analogy, we can posit that LEBA forms part of a personal name.

Image source: Schwinden 1992: 87, fig. 4. Courtesy of L. Schwinden. Commentary: a metal detectorist found this defixio in October 1989 in the Roman vicus in Dalheim. Now the text is housed in Luxembourg’s National Museum of History and Art. While we lack precise information about the tablet’s archaeological context, it is known that it was found in an area of the vicus near the sanctuary and a residential district. The curse could have originally been deposited in either of these contexts.

After being inscribed, the tablet was folded three times: first in half and then each side was folded in towards the centre.

Today, the defixio is fragmentary and corroded in both sides. Of the original edges only the top part of the

204. Wolberg Provenance: on surface. Current Location: M. d’Autelbas. Date: 2nd–3rd centuries CE. Material: lead. Measurements: 3.1 × 6.5 × 0.05–0.1 cm. Reading: A Cam(ulus) or Cam(ulo?)

B G(aius) Curul[---] or Curvẹ[lius?] [---]NIA [---]INV (?)

Bibliography: Schwinden 1988; Schwinden 1989; AE 1989, 537; ILB ² 171; Kropp 2008: no. 4.1.1/1.

The curse was written on an irregularly shaped lead sheet. An opisthograph, it contains four lines, which run from left to right and were written in capitals dating to the second or third century CE. On side A, the letters measure 1.8 cm in height, whereas on side B they are between 0.3 and 1.1 cm. Schwinden (1989) has proposed reading the theonym Camulus on side A, which would be an invocation of the Celtic deity of the same name, who is occasionally equated to Mars. Side B is quite fragmentary but appears to have contained a list of personal names, of which we can only read part of the first: G(aius) Curul[us] (which would be unparalleled) or Curvẹ[lius], in which case the last visible stroke would be part of the cursive e (for this nomen, see OPEL II, 89).

Image source: Schwinden 1989: 86, figs. 1–2. Courtesy of L. Schwinden. Commentary: this defixio was discovered during a land survey in Wolberg (next to Arlon, ancient vicus Orolaunum) in an area that appears to have been a GalloRoman fanum that was built alongside the ancient road that connected Reims and Trier. The tablet, which was found on the ground, was covered with a thick off-white patina and has suffered some damage along its edges.

263

4 Britannia apparently inscribed on both sides in capitals, but since it is now too brittle to be unfolded and the surface is badly corroded, the text is not legible.’

Britannia is the province of the Roman West that has turned up the greatest number of curse tablets. The region is so prolific that the number of defixiones has climbed to 256, and hence it constitutes more than half of the present collection of curses. In general terms, we can say that the tablets come from the southern third of the island and in particular from Aquae Sulis/Bath and Uley, where there are sanctuaries dedicated to Sulis Minerva and Mercury, respectively. The former has yielded 130 tablets, whereas 140 curse tablets have been unearthed at the latter. Of these 140, only 87 are discussed in the following pages since we are waiting for the full edition of the corpus (by R.S.O. Tomlin). In addition to the tablets that hail from Bath and Uley, we must add 40 more defixiones that were discovered in isolation (many of them found thanks to metal detectorists). Small but important collections come from London (eight tablets, 338–44 and below), Ratcliffeon-Soar (five, 349–51 and below) and Pagans Hill (three, 443–45). While the earliest curses from Britannia have been dated between the first and early second centuries CE (see 337, 338 and 369), it was not until the second through to the fourth centuries CE that the practice of writing curses reached its height.

Ratcliffe-on-Soar: Elsdon (1982: 16 and 19, n. 12) mentions the existence of a defixio found nearby a building that should probably be identified as a RomanoCeltic sanctuary. He goes on to note, ‘Outside the building was a surface of gravel and pebbles in which artefacts were buried, including a third folded lead inscription. It is similar in size to the first -Turner- and also in cursive script.’ Ratcliffe-on-Soar: the fifth curse from this village was found by Peter Reeves, an employee at English Heritage, in 1992, alongside a coin dated between 365 and 375 CE. The lead tablet measures 11 × 11 × 0.2–0.3 cm and was pierced by a nail in each of its corners. The curse has not yet been unrolled, but it appears that it does not have an inscription (on this, see Tomlin 1993b: 310, n. 4). Rothwell: in anticipation of a full scholarly edition, Tomlin (2012: 9 and AE 2012, 868) has announced the discovery of a new defixio that was written on a long lead sheet (measuring 4.5 × 8.4 × 0.2 cm) and has been substantially damaged by corrosion. An opisthograph, the tablet contains a text written in new Roman cursive, which has been dated to the third century CE on palaeographic grounds. The text includes words like nomine and mulier and what appears to be a long list of personal names. After being inscribed, the tablet was folded.

Besides the items discussed in this catalogue, we must mention the following curses that are only known through brief mentions in print: Bawtry, Nottinghamshire: Moretti (2015: 108) mentions the discovery of a rolled curse tablet together with ten lead fragments at a Roman shrine. The tablets, which are inscribed, have not yet been published. Lincolnshire: in 2007, the metal detectorist Tom Redmayne found three folded lead tablets. Two of them contained the impression of a coin representing the emperor Valens and were pierced. Although the sheets were not inscribed, they can be dated to the end of the fourth century CE. Currently, they are housed at the British Museum (Inv. no.: 2007.4213.1). More information about these curses can be found at the Portable Scheme Antiquities website (Inv. nos.: LIN-57F021, LIN-57B091 and LIN-5806B6) and in The Guardian (‘Rare find highlights antiquities fears’, Maev Kennedy, Monday, 17 December 2007). For a picture of one of these tablets, see Figure 4.1 in the section I.4.1 of the prolegomenon. London: according to Tomlin (2003: 364), an eighth tablet was discovered in London in the amphitheatre at Guildhall Yard. The curse, a lead sheet measuring 6.2 × 8.8 × 0.1 cm, is now housed in the Museum of London (Inv.: GYE ), but it has not been fully read. The editor explains, ‘Before being folded only once, it was 265

Sylloge of  Defixiones from the Roman West 205. Lydney Provenance: sanctuary. Date: 4th–5th centuries CE. Material: lead. Reading:

5



10



Current Location: Lydney Park Museum. Measurements: 7.5 × 6 cm.

«rediviva» devo Nodenti Silvianus anilum perdedit demediam partem donavit Nodenti inter quibus nomen Seniciani nollis petmittas sanitatem donec perferạ(t) usque templum [No-] dentis

Bibliography: Wünsch 1897: 25; DT 106; ILS 4730; Jeanneret 1917: no. 159; Collingwood 1932: 100–02; Preisendanz 1933: 155; Goodchild 1953; Toynbee 1953: 19–21; Rubio and Bejarano 1955: no. 181; CIL VII, 140; Egger 1964: 15–16; RIB I, 306; Burn 19692: 50; Mann 1971: 220–21; Tab. Sulis, p. 61, no. 28; Versnel 1991: 84; Gager 1992: no. 99; Corby Finney 1994; Tomlin 1999a: 554–55; Adams 1998: 235; Kropp 2008: no. 3.15/1; Tomlin 2017c: 12.46; Urbanová 2018: no. 289; Sánchez Natalías 2022.

The word devo (l. 1) deserves mention, since it may be related to the Celtic word for a deity, deivos (see 373 for a parallel). The defigens offers the deity half the value of the stolen property in an attempt to better his chances of recuperating the lost goods (ll. 4–5, for a parallel, see 443). Furthermore, we find a non permittas formula (ll. 7–9) that is meant to harm the victim’s physical well-being until the stolen ring is deposited in the temple (for a parallel, see 355). The word rediviva (l. 1) was written at a latter time in an attempt to renew the spell’s power (perhaps Silvianus added this on a subsequent visit to the temple of Nodens?). This tactic is unparalleled in the defixiones from the Roman West.

Image source: Collingwood 1932: fig. 28, no. 1. Commentary: this defixio was discovered c. 1817 at Lydney Park in the temple dedicated to the Romano-Celtic deity Nodens, whose cult flourished in the province until the fourth century CE. It was written on a rectangular lead sheet that has been broken into two pieces that fit together. The tablet has also lost its two top- and bottomright corners. Long stored in an oak drawer, this curse has been severely corroded and today is practically illegible (according to Tomlin 1999 a: 555). The inscription contains 13 lines, which run from left to right and were written in capitals.

Various authors, such as Goodchild and Toynbee, have connected this tablet to a specific gold ring (see CIL VII, 1305, RIB II, 2422.14 and Corby Finney 1994) that was discovered in 1786 by a farmer in Silchester, some 50 km south-east of the site of the robbery. The ring’s gemstone depicts the bust of a bearded man in profile around which VENUS can be read. Around the ring’s band a second inscription, which was added at a latter time, is found. This inscription picks up on a Christian legend and reads Seniciane vivas in De[o] (‘Senicianus, may you live in God’; translation from RIB). While the identification of the Senicianus from the ring with the Senicianus from the Lydney curse is undeniably tempting, the fact that this personal name is attested in other curses (213, 303, 456, etc.) problematizes the identification. That said, the identification cannot be ruled out. Finally, the names Silvianus and Senicianus are already attested Latin cognomina (see OPEL IV, 82 and 66, respectively).

Based on its content, this curse should be grouped among the defixiones against thieves. Here, Silvianus denounces the theft of a ring to Nodens. The text presents the following orthographic and phonetic features: anilum for anulum (l. 3), perdedit for perdidit (l. 3), demediam for dimidiam (l. 4), Seniciani for Senecianis (l. 7; see Mann 1971: 220), petmittas for permittas (l. 8) and perfera for perferat (l. 9). Adams, for his part, has rightly proposed the reinterpreted nollis (l. 7) as non illis, a mistake that he attributes to phonetic or scribal confusion (for a parallel, see 250, B, l. 2: non illị permittas...); he also notes the lack of coordination between inter and quibus, since ‘the antecedent which would have been dependent on inter (inter (eos) quibus) has been omitted’ (1998: 235).

Translation (RIB, modified): ‘To the god Nodens: Silvianus has lost his ring and given half (its value) to Nodens. Among those who are called Senicianus, do not allow him health until he brings it to the temple of Nodens. (This curse) comes into force again.’ 266

 Britannia 206. Aquae Sulis, Bath Current Location: Roman Bath Museum. [Inv. No.: batrm 1983.14.b.1] Measurements: 6.8 × 6.8 cm.

Provenance: sanctuary. Date: 2nd century CE. Material: lead alloy. Reading:

5



qụ[i] mihi VILBIAM iṇ[v-]|olavit sic liquat comọ ạqua LLEAT[-c.2-3-]M quị eaṃ [invol-]|avit ++ (vacat) Velvinna Ex[s-]|upereus Ṿerianus Se-|veriṇus Agustalis Com-|itianus Minianus Catus Germanill[a] Iovina (vacat)

related to Vellibia (cf. RIB 1181). Tomlin, however, has rejected such an interpretation on the grounds that this curse would be the only one from Britannia in which the purloined ‘property’ was a woman. Russell (2006), for his part, has offered a new interpretation: vilbiam is a word of British origin (represented by Middle Welsh as gwlf) used to describe a sharp pointed tool, such as a gouge. This could very well be the stolen object. Other stolen tools in Britannia include things like ploughs (236) and axes (452 and 459). Accordingly, Russell’s suggestion is quite welcome since it fits well the context of defixiones against thieves.

Bibliography: Wünsch 1897: 25; Rhys 1880; DT 104; Jeanneret 1917: no. 104; Preisendanz 1933: 155; RIB I, 154; Mann 1971: 221; Smith 1983: 993; Tab. Sulis 4; Adams 1992: 17; Russell 2006; AE 2006, 704; Kropp 2008: no. 3.2/1; Tomlin 2012: 25; Urbanová 2018: no. 242. Image source: Tab. Sulis 4. Courtesy of R.S.O. Tomlin. See also App. IV.3, SD 206. Commentary: this defixio was discovered in 1880 in the sanctuary of Sulis Minerva (Bath) during the excavations carried out by the architect and antiquarian Charles E. Davis, who was in charge of renovating the bath complex. While excavating a small part of the spring’s votive deposit, Davis found four pewter vessels, coins and two curse tablets (on this excavation, see Cunliffe 1988a: 45 with further references).

As far as phrases and formulae are concerned, it is noteworthy that this is the only text from Bath that employs a similia similibus formula: the practitioner asks that the thief sic liquat como aqua (for some parallels, see 118, ll. 17–19: disulvit̂e omnị[a] membra omnis viscida ipsius... and 488, B, ll. 2–5: sic illorum membra liquescan(t) quatmodum hoc plumbum liquescet...).

The curse was found alongside various coins from the reigns of Hadrian, Trajan, Vespasian, Antoninus and Domitian (see the newspaper article ‘The King’s Bath’ published in The Bath Herald (24 April 1880). The curse was written on a rectangular sheet of lead alloy, whose edges have been preserved except for the right side. The inscription contains eight lines of text. While the words run from left to right, the letters of each word have been put in reverse order (i.e., beginning with the last letter and ending with the first).

This curse contains a list of names in the nominative, which belong to those suspected on committing the theft. The list includes Minianus (l. 7, OPEL III, 82), Germanilla (l. 8, diminutive of Germana, whose the ending -illa is more common in Celtic areas, according to Kajanto 1982²: 127; also see OPEL II, 165), the theophoric name Iovina (l. 8, OPEL II, 223) as well as other cognomina that mix Latin and Celtic elements (Velvinna and Catus, ll. 4 and 7, respectively).

Based on its content, we can group this tablet among the curses against thieves. The text presents the following two orthographic features: como for quomodo (l. 2) and Agustalis for Augustalis (l. 6). The lexeme VILBIAM (l. 1) deserves special attention: while the reading is undisputed, the word’s meaning has led to debate. The editors of RIB have claimed that the word is a female personal name,

After being inscribed, the tablet was cut. Translation (Tomlin 1988a: 112, modified): ‘May he who has stolen a gouge from me become as liquid as water … who has stolen it. Velvinna, Exsupereus, Verianus, Severinus, A(u)gustalis, Comitianus, Minianus, Catus, Germanilla, Jovina.’ 267

Sylloge of  Defixiones from the Roman West 207. Aquae Sulis, Bath Current Location: Roman Bath Museum. [Inv. No.: 655] Measurements: 5.2 × 6.5 cm.

Provenance: sanctuary. Date: unknown. Material: lead, tin and copper. Reading:

ABCDEFX

Bibliography: Tomlin 1983: 336–37; AE 1983, 632; Tab. Sulis 1; Velaza 2003: 955; Kropp 2008: no. 3.2/3; Urbanová 2018: no. 207.

brought 128 curse tablets to light. The individual items constituting this large cache will be discussed in the following pages.

Image source: Tab. Sulis 1. Courtesy of R.S.O. Tomlin.

This curse was written on an irregularly shaped sheet of alloyed metal and has been wholly preserved. The inscription contains a single line written in capitals. Tomlin has argued that we have an alphabetic sequence followed by an X, which can be interpreted as ABC def(i)x(io), given that the tablet was found alongside other curses (see commentary to Tab. Sulis 1). Velaza, following the interpretation found in AE, has suggested that we have an alphabetic inscription (instead of a curse tablet) that simply jumps towards the end of the alphabet (see Velaza 2003: 995, n. 11). Even though the latter suggestion appears more probable (in fact, the term defixio has been attested only once in the curse tablets from the Roman West; on this, see Blänsdorf 2019 and I.2, with n. 23), the fragmentary nature of the sequence, together with the use of lead and the context of the deposit argue in the other direction. Hence, it remains an open question.

Commentary: this lead tablet was discovered in the sacred spring of the sanctuary of Sulis Minerva (Aquae Sulis, Bath) during the excavations carried in the years 1979–980. The excavations, which were carried out in the southern half of the spring (c. 29 sq. m, according to Cunliffe 1988: 3), had two main objectives: to address issues related to the spring’s contamination and to fix structural problems that arose from Major Davis’s archaeological efforts in 1879. After draining this section of the spring, archaeologists found a votive deposit made of some 12,000 coins and a wide range of domestic and personal objects (e.g., earrings, bead necklaces, wooden combs and spindle whorls), anatomical votive offerings made of bronze and ivory, jugs, paterae, vessels, a candle holder, gems, etc. (for a full discussion, see Henig et al. 1988: 5–53). In addition to these objects, excavators

208. Aquae Sulis, Bath Provenance: sanctuary. Date: unknown. Material: lead, tin and copper. Reading:

Current Location: Roman Bath Museum. [Inv. No.: 688] Measurements: 3.7 × 6 cm.

Britivenda Venibelia

268

 Britannia Bibliography: Tomlin 1984: 336; AE 1984, 621; Tab. Sulis 2; Kropp 2008: no. 3.2/4; Urbanová 2018: no. 184.

The inscription contains two lines of capitals, which run from left to right. Some of the letters are mirrored (i.e., written from right to left), such as the e or the b. This orientation may explain the difference between the bs, rather than a change of hands (contra Tomlin). The circumstances that led to the writing of this curse are unknown since the text only provides two female names of Celtic origin: Britivenda (found also in 209) and Venibelia (attested here for the first time).

Image source: Tab. Sulis 2. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a roughly rectangular sheet.

209. Aquae Sulis, Bath Provenance: sanctuary. Date: unknown. Material: lead, tin and copper. Reading:

Current Location: Roman Bath Museum. [Inv. No.: 623] Measurements: 7.6 × 7.6 cm.

Br{p}itvenda Marinus Meṃorịṇa

Bibliography: Hassall and Tomlin 1982: 397; AE 1982, 662; Tab. Sulis 3; Kropp 2008: no. 3.2/5; Urbanová 2018: no. 185.

a fracture at l. 3. The inscription contains three lines written in capitals, which run from left to right. The circumstances that led to the writing of this curse are unknown. The text comprises a list of personal names in the nominative. Here we find a Celtic name (l. 1, Britvenda, probably for Britivenda as in 209) and two Latin ones: Marinus (l. 2; see OPEL III, 58) and Memorina (l. 3, see OPEL III, 75).

Image source: Tab. Sulis 3. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a quadrangular sheet, which is in a good state of preservation despite some oxidation and

After being inscribed, the tablet was folded twice.

210. Aquae Sulis, Bath Provenance: sanctuary. Date: 275–400 CE. Material: lead, tin and copper. Reading:

5



Current Location: Roman Bath Museum. [Inv. No.: 477] Measurements: 9.9 × 6.8 cm.

[D]ocimẹdis [p]ẹrdidi(t) manicilia dua qụi illas involavi(t) ut mentes sua(s) perdat et oculoṣ sụ[o]ṣ in fanọ ubị deṣtina(t)

269

Sylloge of  Defixiones from the Roman West Bibliography: Tomlin 1986: 430, 432–33; AE 1986, 465; Tab. Sulis 5; Adams 1992: 5–6, 14; Kropp 2008: no. 3.2/6; Urbanová 2018: no. 244.

person singular verbs: perdidi(t), involavi(t), destina(t) (ll. 2, 4 and 9, respectively). This is a common phenomenon in ‘Vulgar’ Latin (cf. Adams 1992: 6). The lexeme manicilia deserves comment: a diminutive formed from manica (‘sleeve’), Tomlin has translated the word as ‘gloves’ (for other parallels, see 351 and 367). Beginning in the fifth line, the defigens turns attention to the requested punishment, which is introduced by an ut that Adams (1992: 6) has taken as introducing a desire. This veers from the interpretation of Tomlin, who argued for ‘an indirect command dependent upon a verb like rogat … in ellipse’ (1988a: 115). Thus, Docimedis expresses his desire for vengeance in the following terms: ut mentes suas perda(t) et oculoṣ sụ[o]ṣ in fanọ ubị deṣtina(t). Such a punishment is unparalleled. The name Docimedis (also attested in 300) is a masculine personal name that could be Celtic in origin (derived from doci-).

Image source: Tab. Sulis 5. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on an irregularly shaped lead sheet, now broken into two parts that fit together. We can see a fold on the right edge, which was intentionally made before the sheet was inscribed. The inscription contains nine lines of text, which run from left to right and were written in a combination of new Roman cursive (l. 1) and capitals (ll. 2–9). This is the only inscription from Bath that mixes new Roman cursive and capitals (see 214 for a combination of old Roman cursive and capitals).

After being inscribed, the tablet was folded.

Based on its content, this curse should be grouped among the defixiones against thieves. In this case, Docimedis reports the theft of his manicilia (ll. 2–3). Throughout the inscription, we can observe the omission of final t in third

Translation (Tab. Sulis 5, modified): ‘Docimedis has lost two gloves. May (the person) who has stolen them should lose his minds [sic] and his eyes in the temple where (she) appoints.’

211. Aquae Sulis, Bath Provenance: sanctuary. Date: unknown. Material: lead alloy. Reading: 5

Current Location: Roman Bath Museum. [Inv. No.: 673] Measurements: 6 × 6.2 cm.

A B [------] sṭ˹ragulu˺m qm [------] p˹erdidi˺ anima(m) [------] EE+[-c.1-]IATDAEM+ [------] VM (traces) [------] AE ++ [i]nv[o-] (traces) [---]S+ lavit +E++ ˹nisi˺ N[-c.3-]T[---] ṣ[an]g[u]ine ˹ṣua˺

Bibliography: Tab. Sulis 6; Kropp 2008: no. 3.2/7; Urbanová 2018: no. 245.

been written backwards has suggested dyslexia to the editor (Tomlin 1988a: 116).

Image source: Tab. Sulis 6. Courtesy of R.S.O. Tomlin.

Based on its content, this curse should be grouped among the defixiones in fures. This is supported by the lexeme stragulum (the stolen object, B, l. 1) as well as the final phrase used to lay out the victim’s punishment (B, ll. 5–8). The text presents the following peculiarities: qm for qm (B, l. 1, referring to the neuter noun stragulum) and sua for suo (B, l. 7). The defigens’ mention of the thief’s blood (nisi ṣ[an]g[u]ine ṣua, B, ll. 6–7) and life (anima(m), B, l. 2), which are common elements in the requested punishments documented at Bath.

Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a quadrangular sheet. An opisthograph, the tablet is wholly preserved, though it is rather deteriorated on side A, which is practically illegible. The inscription contains 14 lines of text, which run from left to right and were written in capitals. Within each word, the letters are organized in an illogical manner (see starugulm for stragulum, B, l. 1; peridid for perdidi, B, l. 2; isni for nisi, B, l. 6; aus for sua, B, l. 7). The confusion between e and p (e.g., B, l. 2) and the fact that some letters have

After being inscribed, the tablet was folded once. Translation (Tab. Sulis 6): ‘… the rug which I have lost,… (his) life… has stolen… unless with his own blood.’ 270

 Britannia 212. Aquae Sulis, Bath Provenance: sanctuary. Date: unknown. Material: tin. Reading:

Frag. I [---]VO[---]



Frag. II ((3.3) × (2.6) cm) [---]++[s-] angu[ine---] noctis [---] qui mih[i---] VVI[---]

5

Current Location: Roman Bath Museum. [Inv. No.: 399] Measurements: vid. reading.

Bibliography: Tab. Sulis 7; Kropp 2008: no. 3.2/87.

composition and patina. The portion that we have contains seven lines of capitals, which run from left to right. While the circumstances that led to the writing of this curse are unknown, the lexeme sang[uine] (fragment II, ll. 1–2) recalls the classic way of punishing guilty thieves (see 246, 299, etc. for parallels), and it does suggest that this curse should be grouped with the defixiones in fures. It’s possible that the word noctis may have been part of a magical deadline (for parallels see 267, 456–67, etc.; for a discussion see Sánchez Natalías 2019a).

Image source: Tab. Sulis 7. Courtesy of R.S.O. Tomlin. Commentary: these three irregular fragments of a tin sheet were discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207). Two of the pieces fit together (forming fragment II), whereas the third (fragment I, which belongs to the upper edge of the tablet) has been grouped with the other two based on its

213. Aquae Sulis, Bath Provenance: sanctuary. Date: 175–275 CE. Material: lead and tin. Reading:

Current Location: Roman Bath Museum. [Inv. No.: 473] Measurements: 9.4 × 5.2 cm.

A [d]ẹae Suli donavị [arge-] ntiolos `ṣ´ex quos perd[idi] a nomin[i]bus infrascripṭị[s] ̣ deae exactura est (vacat) 5 Seniciaus et Saturniṇụs {ṣẹd} et Anṇ[i]ola carta picta perṣc ̣[ripta] B An[i]ola Senicianus `S´aṭurnins

Bibliography: Hassall and Tomlin 1981: 370–72; AE 1982, 658; Tab. Sulis 8; Adams 1992: 14; Gager 1992: 193–94; Kropp 2008: no. 3.2/8; Urbanová 2018: no. 246.

Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on sheet made of lead and tin that must have originally been rectangular. Currently, the surface is quite uneven, and the edges have been severely

Image source: Tab. Sulis 8. Courtesy of R.S.O. Tomlin. 271

Sylloge of  Defixiones from the Roman West damaged with the exception of the bottom one, which is almost entirely intact. An opisthograph, the inscription contains nine lines of text, which run from left to right and were written in old Roman cursive.

on top of an unfinished e (A, l. 2), f with the shape of e in infrascriptis (A, l. 3), how the first n in Senicianus is underlined while the second is composed of a single vertical stroke (probably because the author of the text read IV instead of NV, A, l. 5), the simplification of geminated Aniola (B, l. 1) and the writing of the first s in Saturninus on top of another letter (B, l.3).

Based on its content, this curse should be grouped among the defixiones in fures, since we learn about the theft of [arge-]ntiolos sex (A, ll. 1–2). Also attested in 259, argentiolos is notable: a diminutive of argenteus, this word is usually associated with coinage from around the year 400 CE. The text, however, has been dated to the third century due to palaeographic considerations. Another noteworthy lexeme is exactura (A, l. 4), which according to Tomlin (Tab. Sulis, p. 119) and Adams (1992: 14) could be a neologism formed in -iura from the verb exactio. Alternatively, and following Adams, it could also be a periphrastic future, with deae misspelt (for dea).

Senicianus (A, l. 5; B, l. 2 see OPEL IV, 66) is a variant of Senicio, which is very common in the Celtic provinces. Saturninus (A, l. 5; B, l. 3 see OPEL IV, 51) is an extremely widespread Latin cognomen, whereas Ann[i]ola (A, l. 6; B, l. 1) is a diminutive of the nomen Annia (see OPEL I, 119). After being inscribed, the tablet was pierced with a nail as can be adduced from the hole found at the last line of side A and the first line of the opposite side, where the hole has damaged the name Anniola. Since the personal names are found in a different order at A, ll. 5–6, Tomlin’s suggestion that the defigens altered the order of the list so that this name in particular would be affected by the nail is certainly within the realm of possibility.

The phrase nomin[i]bus infrascripṭị[s] (A, l. 3) here precedes the list of victims and is found (with slight variations) in other curses (see 160: denuntio persona infrascribtis…). The phrase carta picta perṣc ̣[ripta] (A, l. 6) is attested here for the first time and has sparked a debate among commentators. The editor adduced a recipe against insomnia (Liber Medicinalis, 982) as a comparison to argue that one of the meanings of perscribere can be ‘to copy out’ (Tomlin 1988a: 119; for a different interpretation see Chiarini 2019: 148–49). Following Tomlin, the numerous errors sprinkled throughout the text could be a result of this process, such as the following: s written

Translation (Tab. Sulis 8): ‘I have given to the goddess Sulis the six silver coins which I have lost. It is for the goddess to exact (them) from the names written below: Senicianus and Saturninus and Anniola. The written page has been copied out. Anniola. Senicianus. Saturninus.’

214. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 612] Material: lead and tin. Measurements: 8.5 × 5.9 cm. Reading: A PETIO ROVE · te · Victoria · vind(ices?) · CVN (vacat) Minịc ̣ị 5 Cunomolius Minervina ussor Cunitius ser(v)us Senovara ussor Lavidendus ser(v)us 10 Matṭonius ser(v)us Catinius Ẹx̣sactoris fundo eọ Methịanụ[s -c.3-] [-c.3-] dono

272

 Britannia B [---]+MICVS TPIASV (vacat) GINENINSV[S] GIENVSVS

petitio, we could also read peto (attested in 457 and 22) Accordingly, I propose the reconstruction peto (et) rogo te, Victoria, (ut) vindices. Next, we find a list of the victims’ names in the nominative case. Some of these are qualified with a patronymic (ll. 4–5 and 11), while others are singled out as wives (A, ll. 6 and 8; note ussor for uxor) or slaves (ll. 7 and 9–10; note serus for servus).

Bibliography: Hassall and Tomlin 1982: 400; AE 1982, 664; Tab. Sulis 9; Kropp 2008: no. 3.2/9; Urbanová 2018: no. 186. Image source: Tab. Sulis 9. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on an irregularly shaped sheet. Today the curse tablet is broken in two fragments that fit together. An opisthograph, the inscription contains 18 lines of text, which run from left to right and were written in a mix of capitals (A, ll. 1–3, 5–8 and B, 1–3) and new Roman cursive (A, ll. 4, 9–12 and B, l. 4). Also of note is how at A, ll. 2–3 the words have been separated with interpunctuation comprised of a single dot.

In this curse we find several Celtic names: Cunomolius (A, ll. 4–5; the syllable Cun- is written in l. 4, where due to a crack in the lead, the author of the text decided to rewrite the entire name in the following line), Senovara (A, l. 8; based on the prefix Seno-), Mattonius (A, l. 10; derived from the root matto), Catinius (A, l. 11; from Cattus/ os), and also Lavidendus and Cunitius (A, ll. 9 and  7). Furthermore, we have the theophoric Minervina (A, l. 6, see OPEL III, 82 for the cognomen Minervinus). Side B also contains a series of personal names, which may be preceded by the lexeme [ini]micus (B, l. 1); if not, it is part of another cognomen, such as Limicus, Comicus, etc.). For the final name, which is written in capitals and new Roman cursive (B, ll. 3–4), the following readings could be possible: Ingenuus, Igennus, Ingenuinus or Ingenui.

The circumstances that led to the writing of this curse are unknown. The curse opens with the phrase petio rove te, which Tomlin has reconstructed as pe(ti)tio rogo te or oro te, based on the parallels for rogo te in 340 and 361. The same author has identified Victoria (A, l. 3) and reconstructed vind[-c.3-] as the epithet vindex, which sometimes qualifies Jupiter or Fortune. Since the epithet is rare, a more attractive option in my view would be to take vind as a form of the verb vindico, meaning ‘punish’ or ‘avenge’ (OLD 5–7), which is found in a handful of texts (e.g., 106, 115, 120, 183, 240, 340, 365, 482 or 492). Furthermore, 120 and 240 provide especially tantalizing parallels with the verb rogo (see 120, ll. 4–5: te rogo oro obsecro uti vindices and 240, ll. 2–4: rogo [s]anctissimaṃ maiestatẹṃ tuam u[t] vindices). Instead of the editor’s

After being inscribed, the tablet was folded five times. Translation (Tab. Sulis 9, modified): ‘… I ask and beseech you, Victory, to avenge … Cunomolius, (son?) of Minicius, Minervina (his?) wife, Cunitius (their?) slave, Senovara (his?) wife, Lavidendus (their?) slave, Mattonius (their?) slave, Catinius (son?) of Exsactor … Methianus … I give…’

273

Sylloge of  Defixiones from the Roman West 215. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 2nd century CE. [Inv. No.: 638] Material: lead, tin and copper. Measurements: 9.8 × 7.1 × 0.08 cm. Reading: A Docilianus Bruceri deae sanctissim(a)e Suli 5 devoveo eum [q]ui carạcellam meam involaverit si vir si femina si servus si liber 10 ut [-c.1-2-]VM dea Sulis ṃạximo letum [a]digat nec ei so ṃnum permit-

B tat nec natos nec nascentes do [ne]c ̣ caracallam meam ad tem5 plum sui numi nis per[t]ulerit

Bibliography: Hassall and Tomlin 1981: 372, 374–77; AE 1982, 660; Tab. Sulis 10; Adams 1992: 6–8, 15 and 20; Tomlin 2002: 167; Kropp 2008: no. 3.2/10; Tomlin 2017c: 11.06; Urbanová 2018: no. 247. Image source: Tab. Sulis 10. Courtesy of R.S.O. Tomlin. See also App. IV.3, SD 215.

cm. The curse tablet has been dated to the second century CE on palaeographic grounds. The tablet’s layout is quite noteworthy: as Tomlin has stressed (1988a: 122), the heading, which is centred and written with larger letters, resembles a monumental inscription, which may have inspired the practitioner in this case (for the relationship between public epigraphy and curse tablets, see Sánchez Natalías 2016b).

Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a rectangular sheet, which, back in 1988, was broken into four pieces that fit together. My autopsy of the curse tablet (December 2009), however, showed that it is now broken in six pieces, of which only four fit together. An opisthograph, the inscription contains 19 lines of text, which run from left to right and were written in elegant capitals measuring between 0.3–0.7

Based on its content, this curse should be grouped among the defixiones in fures. Here, Docilinus reports the robbery of various articles of clothing to Sulis. Note the author’s sophisticated use of lexemes like devoveo (A, l. 5, without parallels in the corpus from Bath) and carạcellam (A, l. 6; caracallam, B, l. 3). In addition, note also that there are some problems with agreement and case usage: ṃạximo letum (for maximo leto, A, l. 11) and templum sui (for templum suum, B, ll. 4–5). The mention of the goddess 274

 Britannia in l. 10 is itself noteworthy, since it is the only time in the corpus from Bath where the name appears in the nominative case (dea Sulis). This demonstrates that her name is indeed Sulis and not *Sul, as some scholars had previously suggested.

for himself and his freedmen present and future (natis nascentibus)’ (1988a: 167). Adams, for his part, has homed in on the phrase and has tried to flesh out its meaning. Among other suggestions, nascentes could here refer to children still in the uterus, whereas natus would refer to new-born children. In an attempt to distinguish the two participles, he has suggested that ‘the thief should not have children or even be able to cause them to be conceived’ (1992: 20). For a parallel that deals with children, see 443, l. 9: …[nec nat]ọṣ sanos habẹ[a]ṇṭ.

On the phrasal level, ṃạximo letum [a]digat (A, ll. 11–12), and nec ei soṃnum permittat (A, ll. 12–B, l. 1) are also of note. The first, reconstructed by Adams (1992: 7) as ut eum/illum maximo leto/letum adigat contains the defigens’ wish for Sulis to kill the thief. For parallels, see 530, l. 16: …ṇec[et]i[s] eum pes(s)imo leto and especially, 361, ll. 7–8: …rogo te ut eos maxịmọ ḷẹto adigas; in fact, the editor has argued that 361 was actually written by the same hand based on lexical and palaeographical similarities.

Docilianus (A, l. 1) is a Latin cognomen already attested (see OPEL II, 104), while Brucerus (A, l. 2) is not documented, unless the author of the text wrote this in place of Brucetus (a known local name).

The second noteworthy phrase from this defixio, nec ei soṃnum permittat, is a common non permittas formula that seeks to impede the victim’s vital functions and well-being (for some parallels, see 237, 305, etc.); what is especially noteworthy is how the phrase continues: nec natos nec nascentes (B, ll. 1–2). If the stolen object is not returned, the curse should also deprive the thief of his progeny. According to Tomlin, this reference to descendants parallels the almost legal language found in CIL VI 8063 and XII 3702, where ‘…a patronus provides a burial place

After being inscribed, the tablet was pierced with a nail at A, l. 10. Translation (Tomlin 2002: 167): ‘Docilianus son of Brucetus to the most holy goddess Sulis. I devote him who has stolen my hooded cloak, whether man or woman, whether slave or free. May the goddess Sulis inflict him with the greatest death, and not allow him sleep or children now nor in the future, until he has brought my hooded cloak to the temple of her divinity.’

216. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: 399] Material: lead alloy. Measurements: vid. reading. Reading:

5

Frag. I ((3.2) × (2.5) cm) [---invo-] laver[it---] fuerit[--- dona-] vi si l[iber si servus(?) quicum-] que co[---]



Frag. III ((2.3) × (3.2) cm) [---]+[---] [---qui]cumquẹ[---] [---]quicumque[---]



Frag. V ((2.2) × (0.9) cm) [---]++[---] [---?qu]ịc ̣[cumque---]

Bibliography: Tab. Sulis 11; Kropp 2008: no. 3.2/11.



Frag. II ((2.5) × (1.7) cm) [---]VE[---] [---]DIBAT+[---] [---si l]iber[si servus---] [---]+++[---]



Frag. IV ((1.3) × (1.5) cm) [---]ho[c---] [---]fuerị[t---]

Fragment I has five lines, fragments II and III four lines, and fragment IV and V have two lines. In fragment I part of the tablet’s lower-left corner has been preserved, which originally was probably rectangular in shape, while part of the lower edge of fragment IV has also been preserved. None of the original edges survives on the other fragments.

Commentary: these five fragments come from the sanctuary of Sulis Minerva (for the archaeological context, see 207). They were written in capitals, which run from left to right, and may not have been written by the same hand.

275

Sylloge of  Defixiones from the Roman West 217. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: 399] Material: lead alloy. Measurements: vid. reading. Reading: Frag. I ((2.3) × (4.3) cm) [---]Marin[us?---] [---]quas per[didi?---] [---]+TIOLO meo[---] [---i]nvolaverit si[---] traces



Frag. II ((1.4) × (2.2) cm) Dạṇ+[---] res me[as---] +E++[---]

Bibliography: Tab. Sulis 12; Kropp 2008: no. 3.2/12



Frag. III ((1.7) × (1.7) cm) NIT+[--- res?] meas[---] infr[ascript---?] [---]

actually belong to the same tablet. The first and second contain part of an upper edge, while the third comes from a lower-left corner of a tablet that was probably rectangular in form. Fragment I contains four lines, whereas fragments II and III have three, all of which run from left to right.

Commentary: these three fragments come from the sanctuary of Sulis Minerva (for the archaeological context, see 207). They were written in capitals that are palaeographically similar, though they may not

218. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: 20005] Material: lead and tin. Measurements: (5.2) × (8.5) cm. Reading:

5

Col. I [---] involaverit [---]LLVM invola[verit---]virus ++ muliẹris [---]ILLIDO [-c.2-3-]VS [---]+++LLVM Vitali



Col. II Espeditus Tatirụm

Bibliography: Tab. Sulis 13; Kropp 2008: no. 3.2/13.

Despite the fragmentary nature of the text, we can group it among the defixiones in fures, given the presence of a form of the verb involo (col. 1, ll. 1–2) and the mention of virus ++ muliẹris (col. 1, l. 3, where virus should be taken for viri). The latter could be taken as the remains of an all-inclusive formula with which the defigens attempted to identify the guilty party, although it is never expressed in genitive. The form Vitali (l. 5) could be a dative of Vitalis, or from Vitalianus (for both cognomina, see OPEL IV, 176). The last two lines, written in cursive, contain the names Espeditus (l. 6, for Expeditus; see OPEL II, 131) and Tatirụm (l. 7, perhaps a Celtic name).

Image source: Tab. Sulis 13. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), has been broken into five pieces that fit together. The inscription contains seven lines, which run from left to right and have been arranged in two columns, which were written by different hands: the first column (ll. 1–5) was written in capitals, whereas the second (ll. 6–7) was written in old Roman cursive and written upside down (i.e., you must turn the tablet 180 degrees to read it).

After being inscribed, the tablet was folded twice. 276

 Britannia 219. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: RBS 80 CS 3] Material: lead and tin. Measurements: vid. reading. Reading: 1 2 2a 3 4 5 5a 6 6a 7 8 9



Frag. I ((4.6) × (5.6) cm) luciumio[---] cittimediu+x̣ṣ[---] uibẹc ̣[---]traceọṣ[---] estạịdimaui[-c.2-]+++[---] tittlemṃacatacimluci[---] lendiierandant+[-c.2-]nno (or d) a (or n)[---] [+]ục ̣[-c.2-3-]miotouesularạ+[-c.2]+irando+[---] [-c.4-]+m noṭtanou+m (or a) dịi[---] [-c.6-]cịị++eleubarrau+[-c.1-2-]++[---] [---]staginemse[-c.2-]++[---] [---]++fẹr[---] ̣ [---]+r+[---]



Frag. II ((2) × (1.7) cm) [---]luio [---]ai (or mi or n) qtit [---]rii

Bibliography: Tomlin 1987a: 21–24; Tab. Sulis 14; Lambert 1994: 174; RIG II, 2, *L-108; Mees 2005: 177– 80; Mees 2009: 36–39.

a Celtic text written in the Latin alphabet, since ‘[i]t was technically possible (…) to write a curse tablet in British, using the Latin alphabet.’ (1988a: 129).

Image source: Tab. Sulis 14. Courtesy of R.S.O. Tomlin.

Lambert (apud RIG II, 2) has proposed several different readings: titremimcatacim (l. 4), lundiie (l. 5) and nocta (l. 6). He suggests isolating the lexeme randa (l. 5, derived from *randa) and reads eleu (l. 6a) in reverse to yield vel e-, which could indeed be Latin. He further points out things with a Latinate appearance, like statginem (l. 7), traceos (l. 2a), est[---]imaui (l. 3) and mediu (for medium?, l. 2). In l. 5a ([+]ục ̣[-c.2-3-]miotouesularạ), Lambert has reconstructed a series of personal names in the accusative and vocative cases: [Lu]c[iu]m Ioto Vesula Ra[---]. In his opinion, ‘[l]a répéitition de luciumio (lignes 4 luci[ et 5a, début) nous autorise cependant à considérer ce texte comme un charme magique, donc un texte latin où l’on a pu incorporer des éléments étrangers’ (RIG II, 2: 308).

Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), has been broken into five pieces. Four of the pieces fit together so that we can reconstruct the top-left corner of the tablet (here fragment I), while the last piece (fragment II) comes from the bottom of the right edge. All the fragments have been well preserved, as I was able to confirm during my autopsy (December 2009). The inscription contains 12 lines text, which run from left to right and were written in four different styles: capitals (ll. 1–6: the heading and following lines in smaller letters), old Roman cursive (l. 2a, which is between ll 2–3 and can be dated to the third century CE on palaeographic grounds), smaller capitals (ll 5a and 6a, also written between the main lines) and new Roman cursive (ll. 7–9).

Mees (2005 and 2009), in turn, has proposed different readings and interpretations for certain words, such as luciu- (ll. 1, 4 and 5a), which he has interpreted as ‘to consign’; staginem (l. 7) ‘tin’; lendii (l. 5) from Old Irish lend or Welsh llenn (‘cloak’), which he has taken as the stolen object; erand (l. 5), which he translates as ‘burning’; aidis (l. 3) would be ‘temple’ (cf. Latin aedes); catacim (l. 4), which could be a second stolen item, and finally tittlemma (l. 4), could be the verb ‘steal’ or ‘take’. Mees also has proposed the reconstruction nocta nou[a]m (l. 6), which would be the magical deadline for the curse to take effect (for parallels, see 340 and 444,

The first part of the text (ll. 1–6) is quite strange: according to Tomlin, it ‘is not reversed or enciphered, but is clearly not Latin (…) [and there] seem[s] to be some un-Latin combinations of letters’ (1988a: 128). In these lines, we can find some personal names: Lucium (l. 1), Citti (l. 2), [Mod]esta (l. 3), Idimavi (l. 3; without any known parallels), Cataci (l. 4), Notta (l. 6) and Cunobarrus (l. 6a). Given the presence of these personal names, Tomlin has argued that we do not have a pseudo-inscription, but rather 277

Sylloge of  Defixiones from the Roman West etc.; for magical deadlines, see Sánchez Natalías 2019a). Furthermore, Mees has interpreted vesula (l. 5a) as an epithet for Sulis, ‘the good’, based on an analogy with Vesunna Preticoriorum worshipped in Périgueux.

The difficulties inherent in deciphering Celtic preclude us from settling on a linguistic interpretation for the text. That said, the context of deposition and the media employed make clear that this is indeed a curse tablet.

220. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: 631] Material: lead, tin and copper. Measurements: 6.3 × 6.3 × 0.1 cm. Reading:

nomen rei qui destrale involaverit

Bibliography: Hassall and Tomlin 1982: 401-402; AE 1982, 665; Tab. Sulis 15; Kropp 2008: no. 3.2/14; Urbanová 2018: no. 248.

written in capitals, which measure between 0.4 and 0.8 cm in height. The text runs from left to right, as do the letters; that said, e is consistently ‘mirrored’ (i.e., written from right to left). Based on its content, this curse should be grouped among the defixiones against thieves. Here, the theft of a destrale (ll. 2–3, for dextrale) is reported. The phrase nomen rei (l. 1) merits attention, since we have the legalistic word reus used to identify a victim whose name is unknown. We find another variant of this phrase nomen furis in other curses (see 221, 350, 359, etc.).

Image source: Tab. Sulis 15. Courtesy of R.S.O. Tomlin. See also App. IV.3, SD 220. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a roughly quadrangular sheet made of lead, tin and copper, whose form resembles a tabula ansata (for other curses with this shape, see 337, 463 and 480). The curse tablet is in an excellent state of conservation as I was able to confirm during an autopsy (December 2009). The inscription contains four lines

After being inscribed, the tablet was folded twice. Translation (Tab. Sulis 15): ‘The name of the culprit who has stolen (my) bracelet (is given).’

221. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: 523] Material: lead, tin and copper. Measurements: 5.8 × 7.7 cm. Reading: A nomen furis qui ḷater+

278

 Britannia B R+VET donat[[u]] ur

Bibliography: Hassall and Tomlin 1982: 398–99; AE 1982, 663; Tab. Sulis 16; Kropp 2008: no. 3.2/15.

curse should be grouped among the defixiones against thieves. This formula was used to identify and hand over the victim to the deity (for parallels, see 350 and especially 359). For the final line of side A and the first line of side B, the editor has proposed the reading involaverit, which would undoubtedly be appropriate in this context.

Image source: Tab. Sulis 16. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on an approximately pentagonal sheet, which is wholly preserved, despite corrosion and tearing along the edges. An opisthograph, the inscription contains six lines of text, which run from left to right and were written in capitals. As the formula nomen furis (…) donatur makes clear, this

After being inscribed, the tablet was pierced with two nails. Translation (Tab. Sulis 16): ‘The name of the thief who… is given.’

222. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: 675] Material: lead and tin. Measurements: 2.3 × 6.9 cm. Reading: A Ṣenianus Magnus ̣ Ṃ++++ọ B Lucianu[s] Marc ̣elliaṇus [M]allianus Mu[t]ata Medol+++ GEAĈVS

Bibliography: Tab. Sulis 17; Kropp 2008: no. 3.2/16; ­Urbanová 2018: no. 187.

contains seven lines of text, which run from left to right and were written in capitals. The circumstances that led to the writing of this curse are unknown and the text comprises a list of personal names in the nominative case. Following Tomlin, Mu[t]ata (B, l. 3) can be taken either as a variant of the Celtic Mutacus (Mutaca is attested in OPEL III, 92), while GEAĈVS can be read as part of the preceding name Medol[---], B, l. 3) or the end of a different name.

Image source: Tab. Sulis 17. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on an irregularly shaped tablet made of lead and tin. An opisthograph, the inscription

279

Sylloge of  Defixiones from the Roman West 223. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: 20,001] Material: tin (?). Measurements: 3.5–3.8 cm (diameter). Reading:

5



adixoui deịạna deịeda andagin uindiorix cuamịịn ạị

Bibliography: Tomlin 1987a: 19–21; Tab. Sulis 18; Lambert 1994: 174; RIG II, 2, *L-107; Mees 2005: 177; Mees 2009: 34–36.

patronymic qualifying Vindiorix, which could either be Celtic or perhaps from the Greek κύαμος. Lambert (RIG II.2), for his part, has argued that this text could be taken as an example of ‘Vulgar’ Latin. In that case, deuina would be taken as divina (l. 2), andagin for andangina (l. 4) and cummen(d)ai for commenda[t] (ll. 6-7). For the first line, Lambert has offered the following reading: ad (D?)Ixoui(am), in which Ixui[am] or Dixoui[am] (l. 1) would be a theonym derived either from Isca or*Dexsovī (‘formé sur le nom de la “droite” cf. Dexsiva dans le Vaucluse’, RIG II.2, p. 305). Accordingly, Lambert’s entire reconstruction reads: ad (D?)Ixoui(am) / deuina(m) / ? deieda(m) / andagin(am) / Vindiorix / cummen(d)-/ aṭ. Still, this proposal is highly speculative.

Image source: Tab. Sulis 18. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a tin disc plated with bronze. Given the presence of a ring for hanging the disc, it resembles a pendant. This medium is unparalleled in the present corpus. Perhaps the disc had originally belonged to one of the victims and re-purposed for inscribing the curse? The thought is tempting, but purely speculative. In any case, and as I was able to confirm during an autopsy (December 2009), the disc is whole and in excellent condition. It contains an inscription of seven lines, which run from left to right and was written in capitals (ranging from 0.2 to 0.5 cm in height).

Mees (2005 and 2009), in turn, argues that the text is Celtic. This scholar has taken adixoui (l. 1) as a word derived from the root digs-, which ‘[s]eemingly continuing an earlier tigs- (or stigs-), a form related both to the Greek stigma and Modern English stick, this root is strikingly similar to defixio’ (2009: 35). In his analysis, cuamenai (for cuamịịnạị ll. 6-7) can be connected to the Medieval Irish tuidmen. Between these parts, Mees reads a list of personal names: Devina, Deieda, Andangin and Vindiorix (ll. 2–5).

The circumstances that led to the writing of this curse are unknown and the interpretation of the text has been a matter of debate. According to the editor, the tablet contains a list of Celtic personal names, though the only secure reading is Vindiorix (l. 5). Although this name is without any known parallel, it likely derives from the elements *uindo-, ‘white’ and *rix, ‘king’. The word deịạna (l. 2) can be read as deuina, in which case it would be a name derived from *Deiuos, ‘god’. Finally, Tomlin has suggested that cuaniimai (ll. 6–7) could be a

Despite the difficulties inherent in interpreting this inscription, its layout makes it very likely that it comprises a list of names. In fact, this type of text would have several parallels from Bath, such as 209, 235, 300, 301, etc.

224. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: 399] Material: lead alloy. Measurements: (2.1) × (3.6) cm. Reading:

[---]+++[-c.2-]ḷḷ+dE[---] [---]+em det nisi [---] [---] in vero d(e)AE S[uli (or is?)---] 280

 Britannia Bibliography: Tab. Sulis 19; Kropp 2008: no. 3.2/17.

preserved. The inscription contains three lines, which run from left to right and were written in capitals. The circumstances that led to the writing of this curse are unknown. The lexeme nisi could be part of the conditions for punishment (for parallels, see 211, 270, etc.).

Commentary: this irregularly shaped fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207). Only a small portion of the original lower edge has been

225. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: RBS 80 CS 3] Material: lead alloy. Measurements: vid. reading. Reading:

Frag. I ((1.9) × (3.9) cm) [---d]eae Su[li---] [---]++[---]



Bibliography: Tab. Sulis 20; Kropp 2008: no. 3.2/18.

Frag. II ((2.4) × (2.8) cm) SEQV[---] et M[---] f[---] ̣

which run from left to right and were written in capitals by different hands. Fragment I is broken into two pieces that fit together and preserve part of the original tablet’s upper edge. Fragment II preserves a small part of the original left side.

Commentary: these two fragments were discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207). Fragment I preserved two lines, while fragment II contains the remains of three lines, all of

226. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: RBS 80 CS 3] Material: lead alloy. Measurements: (1.6) × (2.1) cm. Reading:

[--- dea?] Ṣulis t[ibi---] [---]EN[---]

Bibliography: Tab. Sulis 21; Kropp 2008: no. 3.2/19.

see 207). The upper edge of the original tablet has been preserved along with the remains of two lines, which run from left to right and were written in capitals. The first line contains the name of the sanctuary’s patron deity.

Commentary: this fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context,

227. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: 399] Material: lead alloy. Measurements: (2) × (2.4) cm. Reading:

D+[---] ho[c---]

Bibliography: Tab. Sulis 22.

see 207). The lower-left corner of the tablet, which was probably rectangular, has been preserved. The inscription contains two lines, which run from left to right and were written in capitals.

Commentary: this fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context,

281

Sylloge of  Defixiones from the Roman West 228. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: 307] Material: lead alloy. Measurements: (4) × (3.8) cm. Reading: [---]NVS [---]+V [invo]lavit [-c.3-] (traces) Bibliography: Tab. Sulis 23; Kropp 2008: no. 3.2/20.

see 207) and belonged to the upper-right corner of the tablet. The remains of four lines, which run from left to right and were written in capitals, have been preserved. In the first line, NVS is probably the ending of a personal name.

Commentary: this fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context,

229. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: RBS 80 CS 3] Material: lead alloy. Measurements: 3.4 × 2.2 cm. Reading:

[---]+PITA+[---] [---]da fạḷsuṃ[---] ̣

Bibliography: Tab. Sulis 24.

context, see 207) and conserves a portion of the upper edge of the original tablet along with the remains of two lines, which run from left to right and were written in capitals.

Commentary: this small fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological

230. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: 695] Material: lead alloy. Measurements: unknown. Reading:

++ + sẹ[---]+ ------

Bibliography: Tab. Sulis 25.

context, see 207) and conserves a portion of the upper edge of the original tablet along with the remains of three lines, which run from left to right and were written in capitals.

Commentary: this small fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological

231. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: RBS 79 CS 3] Material: lead alloy. Measurements: (1) × (1.1) cm. Reading:

[---]ser[vus?---]

282

 Britannia Bibliography: Tab. Sulis 26; Kropp 2008: no. 3.2/21.

context, see 207) and conserves a portion of the upper edge of the original tablet along with the remains of one line, which runs from left to right and was written in capitals.

Commentary: this small fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological

232. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: 680] Material: lead and tin. Measurements: c. 15 × 13 cm. Reading:

[---]VENDI

Bibliography: Tab. Sulis 27.

face, we can see the remains of an inscription that runs from left to right and was written in capitals. According to Tomlin, VENDI may be the ending of a Celtic name in the genitive.

Commentary: this defixio, which is still folded, was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207) and has been broken into four pieces that fit together. On the outside

After being inscribed, the tablet folded six times.

233. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 399] Material: lead alloy. Measurements: (1.9) × (1.9) cm. Reading:

[---]+ER+[---]

Bibliography: Tab. Sulis 28.

the archaeological context, see 207). The remains of one line, which runs from left to right and was written in capitals, have been preserved.

Commentary: this irregularly shaped fragment was discovered in the British sanctuary of Sulis Minerva (for

234. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 399] Material: lead alloy. Measurements: (0.9) × (1.3) cm. Reading:

[---]+[---] [---]+N[---]

Bibliography: Tab. Sulis 29.

two lines, which run from left to right and were written in capitals, have been preserved.

Commentary: this small irregularly shaped lead fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207). The remains of

283

Sylloge of  Defixiones from the Roman West 235. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 2nd century CE. [Inv. No.: 657] Material: pewter. Measurements: 14.5 cm (diameter). Reading:

5



Severianus fil(ius) Brigomall(a)e Patarnianus filius Matarnus ussor Catonius Potentini Marinianus Belcati Lucillus Lucciani Aeternus Ingenui Bellaus Bellini

Bibliography: Tab. Sulis 30; Adams 1992: 12; Kropp 2008: no. 3.2/22; Urbanová 2018: no. 188.

(l. 1; Brigo is a common element in Celtic names). While the inclusion of a matronymic is frequent in aggressive magic in general, it is rather rare in British curses (for parallels, see 303, 360 and 454). The lexeme ussor (for uxor, l. 3) is found elsewhere in the collection from Bath (214). The text’s Celtic names are Catonius (derived from cattus/os, l. 4), Lucillus (from Lucco, l. 6; so is Luccianus, his father), Bellaus (unparalleled, but apparently related to Bellinus, both of which are in l. 8) and the patronymic Belcatus (no known parallels, but from the name-element Bel, l. 5). The Latin names are Patarnianus (for Paternianus, l. 2; see OPEL III, 127; see Adams for the opening of e to a), Matarnus (for Maternus, l. 3; OPEL III, 65; again, see Adams), Potentinus (derived from potens, l. 4; see OPEL III, 156), Marinianus (l. 5; see OPEL III, 58), Aeternus and Ingenuus (both in l. 6; see OPEL I, 47 and II, 194 respectively).

Image source: Tab. Sulis 30. Courtesy of R.S.O. Tomlin. See also App. IV.3, SD 235. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a pewter plate, a medium that is not found elsewhere in the collection of curses from the Roman West. The inscription contains eight lines, which run from left to right and were written by a trained hand in an elegant old Roman cursive: the letters were carefully formed and the first letter of each word is elaborate. Based on palaeographic criteria, this curse has been dated to the second century CE. As my autopsy confirmed (December 2009), the tablet is in excellent condition and only has a small crack, which may have been caused by the folding of the plate.

After being inscribed, the plate was folded twice. Translation (Tab. Sulis 30): ‘Severianus son of Brigomalla, Patarnianus (his?) son; Matarnus (and his?) wife; Catonius (son of) Potentinus; Marinianus (son of) Belcatus; Lucillus (son of) Luccianus; Aeternus (son of) Ingenuus; Bellaus (son of) Bellinus.’

The circumstances that led to the writing of this curse are unknown. The text contains a list of personal names in the nominative case. The victims are qualified by a patronymic, except for the first, who is qualified by matronymic instead. Severianus fil(ius) Brigomall(a)e

284

 Britannia 236. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 677] Material: lead and tin. Measurements: 8.8 × 3.6 cm. Reading:

5



10



si cus vomerem Civilis involaṿit ut aṇ[imam] su{u}a(m) in teṃplo deponat [?si n]o[n] vom[erem -c.1-] ++VB [-c.2- si se]rvus si liber si lịbertinus [-c.2-] VNAN++O finem faci[a]m

Bibliography: Tab. Sulis 31; Adams 1992: 10; Kropp 2008: no. 3.2/23; Urbanová 2018: no. 249.

back vowel and another vowel in hiatus’ (1992: 10)). We must note the unusual expression aṇ[imam] su{u}a(m) in teṃplo deponat (ll. 4–6), through which the defigens establishes the punishment that is to take place unless the thief returns the stolen property (++VB-/[-c.2-] in ll. 8–9 could form part of the verb in the non si clause). The allinclusive phrase used to identify the victim (ll. 9–11) has the unusual additional of si libertinus. Also note finem faci[a]m (ll. 13–14), in which it is not clear whether finem refers to the life of the victim or the loss of the ploughshare. Finally, Civilis is a Latin cognomen that is well attested in Britannia (see OPEL II, 59).

Image source: Tab. Sulis 31. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), has been broken into five fragments which fit together and form part of what must have been a rectangular tablet. The fragments conserve 14 lines of text, which run from left to right and were written in old Roman cursive.

After being inscribed, the tablet was folded three times.

Based on its content, this curse should be grouped among the defixiones against thieves. In this case, Civilis reports the robbery of a ploughshare. The text presents the following orthographic features: cus for quis (l. 1) and suum for suam (l. 5, although for Adams the second u of suum ‘represents the glide [w], inserted between a

Translation (Tab. Sulis 31): ‘If anyone has stolen Civilis’ ploughshare (I ask) that he lay down his life in the temple [? unless]… the ploughshare, whether slave or free or freedman … I make an end to…’

285

Sylloge of  Defixiones from the Roman West 237. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 616] Material: lead and tin. Measurements: 9.1 × 7.8 cm. Reading:

5



10



15

dẹae Ṣuli Minerv(a)e Sọlinus dono ṇụmini ṭuo maiesṭaṭị pạx̣sa(m) ba(ln)earẹṃ et [pal-] leum [nec p]ermittạ[s so]mnum nec san[ita]tẹm{+}eị qui ṃihi fr(a)udem [f]ecit sị vir si femị[na] si servus s[i] ḷ[ib]er nissi {ṣ}ṣe retegens istas s[p]ecies ad [te]mplum tuụm detulerit [-c.2-3-li]beri sui `ve´l SON[-c.1-2-]SVA e[t?] qui [-c.1-]+[-c.1-2-]++[-c.5-]DEG+++[---] ei quoque [-c.3-]+[-c.2-]x̣E+[---] +[-c.4-]+[-c.2- so]mnum ne[c sanitate-] m+[-c.3-]++[-c.3-]N++[-c.4-p]al{ụ}l[e]ụm et reli{n}q[ua]s nissi ad [te]mplum tuum istas res retulerint

Bibliography: Tab. Sulis 32; Reynolds 1990: 381; Adams 1992: 14–15 and 18–19; Kropp 2008: no. 3.2/24; Tomlin 2017c: 12.34; Urbanová 2018: no. 250.

the use of s[p]ecies (l. 8) as a synonym for res (l. 15; cf. Adams 1992: 18–19). This curse employs several of the most common formulae found in the collection of British defixiones: dono ṇumini ṭuo maiesṭaṭị... (ll. 2–3) is used to transfer ownership of the stolen property to the deity so that the theft becomes sacrilegious; [nec p]ermittạ[s so]mnum nec san[ita-] tẹm (ll. 4–5 and 13) seeks to attack the victim’s physical well-being; qui ṃihi fr(a)udem [f]ecit (ll. 5–6) and the all-inclusive sị vir si femị[na] si servus si ḷiber (ll. 6–7) are both used to identify the guilty party. L. 9 has been damaged and Tomlin has offered the reconstruction s(p)onsa sua or perhaps s(p)on[s(a)e] suae, while Reynolds has proposed soror sua. Either way and in conjunction with [li]beri sui, we have a ‘hereditary’ defixio, which not only seeks to harm the victim, but also his family (for parallels, see 246, 258 and 370). Finally, Solinus is a well attested Latin cognomen (on this, see OPEL IV, 87).

Image source: Tab. Sulis 32. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a rectangular sheet made of lead and tin. Currently, it is split into three pieces that fit together. While three sections are missing from the lateral edges, the top and bottom edges are whole. The inscription contains 15 lines, which run from left to right and were written in old Roman cursive. Based on its content, this curse should be grouped among the defixiones in fures. Here, Solinus reports the theft of various articles of clothing, which he dedicates to the patron deity of the sanctuary. This repetitive and formulaic text is riddled with vulgarisms and orthographic errors, which are probably the result of the defigens’ lack of attention and urgency when writing the curse. We find the following orthographic features in the text: paxsa for pexam (l. 3), Minerve for Minerv(a)e (l. 2), nissi for nisi (ll. 7 and 14), banearem for balnearem (l. 3), frudem for fr(a)udem (ll. 5–6, see Adams 1992: 14ff.), eei for ei (l. 5), ssei for sei (l. 3), paluleum for pallium (l. 13), relinquas for reliquas (l. 14) and retulerint for retulerit (l. 15). As far as diction is concerned, note the phrase pạx̣sam balnearem (l. 3; a reference to the stolen bathing garment) as well as

After being inscribed, the tablet was folded four times. Translation (Tab. Sulis 32): ‘Solinus to the goddess Sulis Minerva. I give to your divinity (and) majesty (my) bathing tunic and cloak. Do not allow sleep or health to him who has done me wrong, whether man or woman, whether slave or free, unless he reveals himself and brings those goods to your temple… his children or his… and(?) who… to him also… sleep or [health]… cloak and the rest, unless they bring those things to your temple.’ 286

 Britannia 238. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: RBS 80 CS 3] Material: lead alloy. Measurements: (3.8) × (3.7) cm. Reading:

deo Marti +[---] do[no?] maiest[ati tuo---] sạcellum +[---] [-c.2-]nisi E+[---] ------

Sulis (see 302 for a parallel). Although fragmentary, the formula do[no?] maiest[ati tuo---] (l. 2) likely alludes to a dedication made by the defigens to the deity (perhaps the stolen property, part of its value or even the victim); such a gift is meant to rouse the god to action. The lexeme sạcellum (l. 3) has been debated: for Tomlin, it refers to a small sanctuary (perhaps referring to the place where the dedication should take place), whereas Reynolds has taken it as a vulgarization of sạc[c]ellum (a small bag), which would be the stolen object. Even though both explanations are compelling, the former better fits with the text itself.

Bibliography: Tab. Sulis 33; Reynolds 1990: 381; Kropp 2008: no. 3.2/25; Urbanová 2018: no. 251. Image source: Tab. Sulis 33. Courtesy of R.S.O. Tomlin. Commentary: this fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207) and belonged to the upper-left corner of a tablet that was probably rectangular. The remains of four lines, which run from right to left and were written in old Roman cursive, have been preserved.

Translation (Tab. Sulis 33): ‘… to the god Mars … [I] give to [your] majesty… shrine… unless…’

The circumstances that led to the writing of this curse are unknown. What jumps out is that the tablet invokes Mars, even though it was deposited in the sanctuary of

239. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 621] Material: lead and tin. Measurements: (6.6) × 9.8 cm. Reading:

5



deae Suli Minervạẹ Docca dono numini tuo pecuniam quam [-c.5- a]misi id est (denarios) (quinque) eṭ ịṣ [q]ui [eam involaveri]ṭ si ser[vu]ṣ ṣ[i liber] [si vir si femina] exṣigat[ur -c.5-] ̣ [---]+[---]

Bibliography: Tomlin 1982: 403; AE 1982, 666; Tab. Sulis 34; Kropp 2008: no. 3.2/26; Urbanová 2018: no. 252.

was probably rectangular. The inscription contains six lines, which run from left to right and were written in old Roman cursive by a trained hand.

Image source: Tab. Sulis 34. Courtesy of R.S.O. Tomlin.

Based on its content, this curse should be grouped among the defixiones in fures. Here, the text reports the theft of five denarii, which are dedicated to the deity, so that she would become personally invested in punishing the thief. The use of the verb amitto (l. 3, [a]misi) instead of the more

Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), is now broken in four fragments that fit together and constitute the upper portion of a tablet that 287

Sylloge of  Defixiones from the Roman West common perdo (see 356 and 460) is noteworthy. Also of note are the symbol for denarii (see 349, 443 and 455 for parallels) as well as exsigatur for exigatur (l. 5; parallels in 243 and 241). Tomlin has proposed the reconstruction of the all-inclusive formula si ser[vu]ṣ ṣi liber [si vir si femina] (ll. 4–5), which the defigens uses to identify the guilty party. Docca is a Celtic personal name, which is first attested here and from which both Doccalus and Doccius are derived.

After being inscribed, the tablet was pierced by a nail in its upper-right corner. Translation (Tab. Sulis 34): ‘Docca to the goddess Sulis Minerva. I give to your divinity the money which I have lost, that is five denarii; and he who [has stolen it], whether slave or [free, whether man or woman], is to be compelled…’

240. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 655] Material: lead and tin. Measurements: (4.6) × 6.1 cm. Reading:

5



deae Sul[i] Minervae rogo [s]anctissimaṃ maiestatẹṃ tuam u[t] vindices ab hiṣ [q]ui [fra-] [ude]m fecerụnt ut ẹi[s per-] mittas nec ̣ ṣ(o)mnum [nec] [---]+[---]++[---]

Bibliography: Tab. Sulis 35; Reynolds 1990: 381; Kropp 2008: no. 3.2/27; Urbanová 2018: no. 253.

We do not know the particular details that led to the writing of this curse, in which the defigens denounces [q]ui [fraude]m fecerụnt (for parallels, see 185 and 237). It is worth stressing that the text begins with a phrase analogous to that found in the inscription from Merida dedicated to Proserpina (120, ll. 2–4): per tuam maiestatem te rogo oro obsecro uti mi vindices. Also, the phrase ut ẹi[s per-] mittas nec ̣ ṣ(o)mnum (ll. 5–6) is a typical non permittas formula used to harm the victim’s physical well-being.

Image source: Tab. Sulis 35. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), is now broken in six fragments that fit together. The tablet, originally rectangular, has lost the bottom and part of the right edges and has suffered serious damage at ll. 3–5. The inscription contains seven lines, which run from left to right and were written in old Roman cursive.

Translation (Tab. Sulis 35): ‘To the goddess Sulis Minerva. I ask your most sacred majesty that you take vengeance on those who have done (me) wrong, that you permit them neither sleep [nor…]’

241. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 679] Material: lead and tin. Measurements: (5.5) × (5.9) cm. Reading:

5



[-c.1-2-]Q++[---] et invola[vit (or -verit)---] duo de[-c.2-]++[---] ad hui(u?)s GAR+[---] DEVENIAT si lib[er] si ser(v)us si puer [si]

uella si vir s[i ] [---]

288

 Britannia Bibliography: Tab. Sulis 36; Kropp 2008: no. 3.2/28; Urbanová 2018: no. 254.

of two objects (l. 3: duo de[---], perhaps de[stralia], as the editor has conjectured). We must highlight the lexeme deveniat (l. 5), which is first found here in the collection from Bath and could constitute an attempt to write deus inveniat. We also find the all-inclusive formula si lib[er] si ser(v)us si puer [si]

uella si vir [si]… (ll. 5–7; perhaps ending with mulier or femina) that was used to identify the guilty party.

Image source: Tab. Sulis 36. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a sheet made of lead and tin that must have originally been rectangular. Today, it has suffered severe damage: only part of the upper-left corner has been preserved. The rest of the edges as well as a considerable portion of the upper third of the tablet have been lost. The inscription contains seven lines, which run from left to right and were written in old Roman cursive.

After being inscribed, the tablet was folded. Translation (Tab. Sulis 36): ‘…and has stolen… two... whether free or slave, whether boy or girl, whether man [or woman]…’

Based on its content, this curse should be grouped among the defixiones against thieves. Here, we learn of the theft 242. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 594] Material: lead and tin. Measurements: 4.8 × 6.8 cm. Reading: A illorum anima ḷạṣ(s)et[ur?] Titumus Ṣedileubi Sedịạc ̣5 u«ṣ» QVEPANVM

B Exsibuus Lothuius Mascntius Aesibuas 5 Petiacus

Bibliography: Tab. Sulis 37; Reynolds 1990: 381; Kropp 2008: no. 3.2/29; Urbanová 2018: no. 189.

the defigens seeks to diminish the life force of his/her victims, whose names follow. Titumus is a Celtic personal name (A, l. 3; perhaps related to Tetumus, found in CIL V, 4883), while Sedileubi (A, l. 4) could be a patronymic derived from Sedebelia (for this question, see the new word division of A, ll. 4–5, which Reynolds has proposed). On side B, we find several names attested for the first time: Exsibuus (B, l. 1), Lothuius (B, l. 2), Aesibuas (B, l. 4; a theophoric derived from Aesus) and Petiacus (B, l. 5). Mascntius (B, l. 3), for its part, could be the Latin cognomen Maxentius (see OPEL III, 68).

Image source: Tab. Sulis 37. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on an irregularly shaped sheet made of lead and tin. An opisthograph, the inscription contains 10 lines of text, which run from left to right and were written in old Roman cursive by a trained hand. The circumstances that led to the writing of this curse are unknown. It opens with the formula illorum anima ḷạṣ(s)etur (A, ll. 1–2). There is an interesting parallel from North Africa: lassetur [-c.14-] {h}animam et {i}spiritum deponat) (cf. DT 250 B, ll. 12–13). In the present curse,

After being inscribed, the tablet was pierced with an iron nail. Translation (Tab. Sulis 37): ‘May their life be weakened [list of names follows].’ 289

Sylloge of  Defixiones from the Roman West 243. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 687] Material: lead and tin. Measurements: 7.5 × (4.6) cm. Reading:

5



10



[---] dono ti[bi---]ream [---]L[-c.1-]SIVIO meọ [---e]x(i)gas pe[r? sa-] [nguinem? e]ius qui has [involave]rit vel qui [medius? fuer]it si femina [---]+O+ ṣị liber [---]++[-c.1-2-] SA+[-c.1-]++ [---]+++ [---]++VM pertuleri(t)

Bibliography: Tab. Sulis 38; Reynolds 1990: 381; Kropp 2008: no. 3.2/30; Urbanová 2018: no. 255.

dedicates the stolen item (or part of its value) to the deity (l. 2, [au]ream(?)). Despite the fragmentary nature of the text, we can recognize several formulae that are frequent in the collection from Bath: [e]x(i)gas pe[r sanguinem e-] ius (ll. 4–5, which is used to ask that the thief pay for his misdeeds with blood, as in 246, 249, etc.); the all-inclusive formula to identify the victim (ll. 7–8) and perhaps the classic non permittas formula to affect the well-being of the suspect, if we accept the reconstruction san[itatem] (l. 9) proposed by Tomlin (see 215, 237 for some parallels).

Image source: Tab. Sulis 38. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a sheet that was originally rectangular. The tablet has been severely damaged: the left edge has been lost and there are numerous instances of tearing. The inscription contains 11 lines, which run from left to right and were written in old Roman cursive by a trained hand.

After being inscribed, the tablet was folded four times. Translation (Tab. Sulis 38): ‘…I give to you… [that] you may exact (them) [through the blood of him] who has [stolen] these, or who has [been privy to it,] whether woman …[whether] free …has bought…’

This curse should be grouped among the defixiones in fures, as is suggested by the phrase dono ti[bi] found at the beginning of the text. With this phrase, the defigens

244. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 676] Material: lead, tin and copper. Measurements: 4.8 × 9.1 cm. Reading:

Col. I qui involaverit si ser(v)us si lib[e]r «L»ENTE++

Col. II Totịa (?) anima(m) suam [q]ụ(i) [i]ṇṿolav[erit?] [---]a(m) meam [---]N++O+ (traces) (traces) qu(i) in290



Col. III vọlav[erit?---] [---] VESEL

 Britannia

Bibliography: Tab. Sulis 39; Kropp 2008: no. 3.2/31; Urbanová 2018: no. 256.

qui involaverit (col. I, l. 1, col. II, ll. 3 and 8 and col. III, l. 1) which is an obvious allusion to the thief; si ser(v)us si lib[e]r (col. I, l. 2), which is a classic all-inclusive formula used to identify the wrongdoer; and anima(m) suam (col. II, l. 3), which forms part of the penalty that the victim ought to pay (for parallels, see 237 and 242). Totia (col. II, l. 1) is a feminine Celtic personal name already attested.

Image source: Tab. Sulis 39. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on an irregularly shaped sheet, which was hammered flat before being inscribed. The text contains eight lines, which have been arranged in three columns whose text runs from left to right and was written in old Roman cursive by a poorly trained hand. This curse should be grouped among the defixiones in fures, as can be deduced from the following three phrases:

After being inscribed, the four edges of the tablet were folded towards the centre. Translation (Tab. Sulis 39): ‘Who has stolen, whether slave or free … his life … [who] has stolen … my… who has stolen…’

245. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 672] Material: lead and tin. Measurements: 5.8 × 9.2 cm. Reading:

qui c ̣alamaea ṇegạt ̣ ṣanguine ̣ +++INEN+++[---] de[s]t[in]at

Bibliography: Tab. Sulis 40; Kropp 2008: no. 3.2/40; Urbanová 2018: no. 257.

with the identification of the lexeme c ̣alamaea (l. 1) as a Vulgar form of calumnia(m), the editor has argued that ‘the text would be evidence of an “ordeal spring” role for the spring of Sulis, a way of testing oaths’ (Tomlin 1988a:  159). The mention of sanguis (l. 2) presumably references the proposed punishment for the target of curse (as in 246, 249, etc.).

Image source: Tab. Sulis 40. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a rectangular sheet, which has been fully preserved despite some corrosion. The inscription contains four lines, which run from left to right and were written in old Roman cursive. Judging from its content, and given the verb nego (l. 2) together

After being inscribed, the tablet was folded six times. Translation (Tab. Sulis 40): ‘[Let him] who denies (making) false accusation (?)… blood… (she)… appoints.’ 291

Sylloge of  Defixiones from the Roman West 246. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 598] Material: lead and tin. Measurements: (4.4) × (10.8) cm. Reading:

5



+++ (traces) [---]++[-c.2-]+rịpuit ut ẹ[o]rum pretium [---? et e]x̣igaṣ học per sanguinem eṭ sa[nitatem sua]ṃ et suorum nec ante illos patịạr[is bibere? nec] ṃạnducare nec adsellare nec [meiere? -c.8-]ius hoc [-c.1-]BISOVERIT (vacat)

Bibliography: Tab. Sulis 41; Adams 1992: 12; Kropp 2008: no. 3.2/33; Urbanová 2018: no. 258.

verb patior) that is meant to harm the victim’s physical well-being. For [-c.1-]BISOVERIT at the end of the text, Tomlin has proposed the reading [u]bi soverit or [a]bisoverit, which would be related to the idea of payment (for more thoughts in this regard, see Adams’ commentary, 1992: 12).

Image source: Tab. Sulis 41. Courtesy of R.S.O. Tomlin. Commentary: these two fragments were discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207) and make up the lower right corner of a defixio that was originally rectangular. The inscription is extremely fragmentary, of which six lines, which run from left to right and were written in old Roman cursive by a trained hand, have been preserved.

As far as diction is concerned, the following three verbs are noteworthy: rapio (l. 1) is used as a synonym for the more common involo; the verb manduco (l. 5) is not common in Classical Latin, though it is attested in other British curses (e.g., 365 and 366); finally, adsello (l. 5) is a technical term from medicine meaning ‘defecate’, where we might expect verbs such as defecare or cacare; according to the editor, this is ‘perhaps the only… instance of deliberately non-“Vulgar” language in the Bath tablets’ (1988a: 161).

Based on its content, this curse should be grouped among the defixiones in fures. The text is so fragmentary that we only can read part of the proposed punishment (ll. 2–4): the practitioner asks Sulis to exact the value of the stolen property in blood (Tomlin 1988a: 160; the lexeme pretium is apt in the context exacting compensation for the theft in blood, per sanguinem suam). The phrase [sua]ṃ et suorum (l. 4) suggests that the punishment was meant to be extended to the victim’s kin (for parallels, see 237, 258, 370, etc.). In ll. 3–4 we have a take on the common non permittas formula (though here with the uncommon

After being inscribed, the tablet was folded. Translation (Tab. Sulis 41): ‘…has stolen, that… the price [of them and] exact this through [his] blood and [health] and (those) of his family, and not to allow them [to drink or] eat or defecate or [urinate] before he has …-ed this.’

247. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 399] Material: lead and tin. Measurements: (2.9) × (3.8) cm. Reading:

[---]+A+E+NA+[---] [---]fecit do[n---] [---i]n fano su[o (or -lis?) ---] [---]+[---]

292

 Britannia Bibliography: Tab. Sulis 42; Kropp 2008: no. 3.2/34; ­Urbanová 2018: no. 259.

written in old Roman cursive by a trained hand. In this highly fragmentary text, we can make out the verb fecit (l. 2), which could have formed part of the well attested phrase qui fraudem fecit (for parallels, see 237, 357, 365 and 370). The mention of a fano (l. 3) suggests two possibilites: first, that the victim was being offered to the deity in the sanctuary (see 220, 221, 249, etc.); or second, that the temple of Sulis was the place in which the stolen property had to be returned (see 215, 237, 250, etc.).

Image source: Tab. Sulis 42. Courtesy of R.S.O. Tomlin. Commentary: this fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207) and originally formed the upper part of a tablet that was probably rectangular. The inscription contains four lines, which run from left to right and were

248. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: RBS 80 CS 3] Material: lead and tin. Measurements: (2.9) × (2.7) cm. Reading:

[---] (vacat) PV[---] [---]dono ti[bi---] [---p]aḷliuṃ[---]

Bibliography: Tab. Sulis 43; Kropp 2008: no. 3.2/35.

left to right and were written in old Roman cursive. The reconstruction of the phrase dono ti[bi---] [---p]aḷliuṃ (ll. 2–3) allows us to suppose that the curse is a defixio against thieves.

Commentary: this small fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207) and conserves the upper edge of a tablet that was originally rectangular. The inscription contains three lines, which run from

Translation (Tab. Sulis 43): ‘…I give to you… cloak…’

249. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 615] Material: lead, tin and copper. Measurements: 5.8 × 7.5 cm. Reading: B dono si ṃuḷ[ie]r si ba(ro) si servus si lib er si puẹr si puel la euṃ latr[on]5 em qui rem ipsa m iṇvolavị[t] d euṣ [i]nvenia[t]

A a[e]ṇ[um me]ụm qui levavit [e]x̣c onic[tu]ṣ [e]sṭ templo Sulis dono si mulier si baro si ser vus si libẹr si pu˹er˺ sị puella 5 et qui hoc fecerit san gu(in)em suum in ips˹um˺ aen ˹um˺ fundat

293

Sylloge of  Defixiones from the Roman West Bibliography: Tomlin 1984: 334–35; AE 1984, 620; Tab. Sulis 44; Adams 1992: 15–17, 20–21; Gager 1992: no. 95; Kropp 2008: no. 3.2/36; Urbanová 2018: no. 260.

Perhaps due to the more complex layout (right to left), the text contains a range of errors, such as pure for puer (A, l. 4), ipsmu for ipsum (A, l. 6) aen-/mu for aen-/um (A, ll. 7), as well as the omission of several letters: sanguem for sanguinem (A, l. 6), exconictus for exconfictus (B, l. 1) and ba for baro (B, l. 2). As far as diction is concerned, note the use of levare to mean ‘rob’ (A, l.1; likely a colloquial use of the verb). Also of note is the lexeme [e]x̣conic[tu]ṣ (A, ll. 1–2), which is here found for the first time; as Tomlin has suggested, the reading should be *excon(v)ictus for exconfixus, (meaning ‘utterly cursed’) from the verb configo with an intensifying ex-. The lexeme baro (A, l. 3; B, ll. 1–2) is also a curiosity: given its presence in the all-inclusive formula where it is contrasted with mulier, it must mean ‘man’ in this context, although in Classical Latin it bears the meaning ‘idiot’. For Adams (1992: 15–17), this lexeme of Germanic origin should be taken to mean ‘warrior’. Finally, latr[on]em is a synonym for the more common furem (B, ll. 4–5; on the word, see Adams 1992: 20–21).

Image source: Tab. Sulis 44. Courtesy of R.S.O. Tomlin. See also App. IV.3, SD 249. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was now broken into six pieces that fit together and form a nearly complete rectangular tablet, which has nevertheless suffered from corrosion and been damaged along its edges. An opisthograph, the inscription contains 14 lines of text, which run from right to left and were written in old Roman cursive by a trained hand. Based on its content, this curse should be grouped among the defixiones against thieves. Here, the defigens reports the theft of a bronze vessel and urges Sulis to punish the wrongdoer. The thief is offered to the deity in her temple (templo Sulis dono, A, ll. 2–3), where, much like in a sacrifice, the culprit ought to spill his or her blood in the stolen vessel (sangu(in)em suum in ips˹um˺ aen˹um˺ fundat, A, ll. 5–7). This request is without known parallels. In an attempt to identify the victim, the defigens employs two well-attested formulae (A, ll. 3–4 and B, ll. 1–4) as well as the phrase qui rem ipsam iṇvolavị[t] deuṣ [i]nvenia[t] (B, ll. 5–7). This latter phrase is also unattested and here alludes to the goddess’s ability to locate the guilty party.

After being inscribed, the tablet was folded twice. Translation (Tab. Sulis 44): ‘(The person) who has lifted my bronze vessel is utterly accursed. I give (him) to the temple of Sulis, whether woman or man, whether slave or free, whether boy or girl, and let him who has done this spill his own blood into the vessel itself. I give, whether woman or man, whether slave or free, whether boy or girl, that thief who has stolen the property itself (that) the god may find (him).’

250. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 691] Material: lead, tin and copper. Measurements: 4.1 × 6.4 cm. Reading: A deae Suli+[---] N (or M) (traces) (traces) [---]+is qu[i]

B ṣi servus si liber {si} qui{s}cumq[ue] erit non illị permittas nec oculos nec sanitateṃ nisi caecitaṭẹṃ ọrbitatemque quoad vixerit 5 nisi haec ad fanum [---]

294

 Britannia Bibliography: Tab. Sulis 45; Adams 1992: 1–2; 4–5; Kropp 2008: no. 3.2/37; Urbanová 2018: no. 261.

as it is used as ‘a rhetorical device to emphasize what it follows by confounding expectation; one expects a concession (‘except for…’), but finds the curse repeated instead’ (Tomlin 1988a: 166). For example, the only type of sight that the victim is granted is being blind (i.e., no sight) before returning the stolen goods. The second use of nisi better conforms to expectations. For another rhetorical use of nisi, see 305, B, ll. 2–5).

Image source: Tab. Sulis 45. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a rectangular sheet made of lead, tin and copper. Although the defixio is almost entirely preserved, it has lost the lower-right corner, is missing sections along its edges and has been greatly damaged by corrosion. An opisthograph, the inscription contains nine lines of text, which run from left to right and were written in old Roman cursive.

For Tomlin, the lexeme ọrbitatemque (B, l. 5) refers to the desire that the victim be childless (for a similar idea, see [nec nat]ọṣ sanos hab ẹ[a]ṇṭ…in 443, l. 9). For Adams, however, the word means ‘blind’: ‘in co-ordination with caecitatem, orbitatem cannot reasonably be given any meaning other than blindness, particularly since the linking particle is -que (…) From the earliest period -que was often used to link synonyms or opposites’ (1992: 2). Finally, the phrase quoad vixerit (B, l. 4) has a legalistic ring to it (it is documented in the Digest (Vocabularium Iurisprudentiae Romanae s. v. vivo, 1422, 26) and here constitutes the magical time frame during which the curse ought to take effect (for parallels, see 350, ll. 3–4: …usque die`m´ qụọ moriatur and 459, ll. 6–7: …(ad) diem mortịs; for temporal expressions in curse tablets, see Sánchez Natalías 2019a).

The circumstances that led to the writing of this curse are unknown. The text contains several formulae that are well documented in the British curse tablets, such as the allinclusive formula used to identify the guilty party (B, l. 1), which is followed by the phrase {si} qui{s}cumq[ue] erit (B, ll. 1–2). Tomlin has taken this last phrase as a combination of si quis erit and quicumque erit, while Adams (1992: 4–5) has seen it as a hapax of great interest: ‘si quiscumque stands to quicumque as si quis stands to qui.’

After being inscribed, the tablet was folded in half.

The text also contains a non permittas formula (B, ll. 2–3), with which the defigens attempts to impair the victim’s physical well-being; the condition under which the victim’s physical impairment should end is laid out in a nisi clause (B, l. 5). As the editor notes, there is another use of nisi that at first sight defies expectations, in so far

Translation (Tab. Sulis 45): ‘To the goddess Sulis… … whether slave or free, ‹if› whoever he shall be, you are not to permit him eyes or health unless blindness and childlessness so long as he shall live, unless [he…] these to the temple.’

251. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 611] Material: lead, tin and copper. Measurements: 11.1 × 8.1 cm. Reading: de(ae) Suli Ṃine[r]ṿ- (vacat) ae ẹos qui Ạmaliam +[---] «T»RASVENDET ṣtilum LA[-c.1-]++[---] CORREGEN eṭ C+GEET++ faṇ[o? ---] 5 «ṭ» SVV dea++tẹ do+++[---] e (or u)t q(u)o HABVNIT[-c.1-]SETRO deam e (or u)t san(g)ụẹne suạ+[-c.1-]BIT qui mẹ VITIS et MALV (or E)+IC++EM (vacat) Doc ̣igeṇius V (or E)TEANE[---] ̣

295

Sylloge of  Defixiones from the Roman West Bibliography: Tab. Sulis 46; Reynolds 1990: 381; Kropp 2008: no. 3.2/38; Urbanová 2018: no. 262.

personal name belonging to the defigens (unless it is an error for Aemiliam). Trasuendet (l. 3), could be a form of the verb tra(n)suendo, though the future tense and singular number (cf. ẹos qui, l. 2) complicate this interpretation. More secure is the reading of a form of dono in do+++ (l. 5). In l. 6, habunit appears to be the verb habeo (perhaps for habuerit?). The phrase sanụẹne suạ (l. 7) can be reconstructed as sanguine suo, which recalls the classic request for payment for the purloined object (perhaps the ṣtilum found in l. 3?) with the thief’s blood. Docigenius (l. 9) appears to be a Celtic personal name first attested here and derived from the root doci-. Reynolds, for her part, has conjectured the reading of a form of malefacio in l.8.

Image source: Tab. Sulis 46. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), has been broken into seven pieces that fit together to form a tablet that is roughly rectangular. The curse has been very corroded, especially its right half. The inscription contains nine lines, which run from left to right and were written in old Roman cursive. Despite problems of textual interpretation, this curse should be grouped among the defixiones in fures. For the first lines, Tomlin has proposed the following reading: de(ae) Suli Ṃine[r]ṿae ẹos qui… (ll. 1–2). Next, we find amaliam (l. 2), which may be an unattested

After being inscribed, the tablet was folded six times. Translation (Tab. Sulis 46): ‘To the goddess Sulis Minerva…’

252. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 684] Material: lead and copper. Measurements: (3.6) × (6.6) cm. Reading:

5

[---?tib]i q[u]er[or] [---]+++[-c.2-]+ ex{x}igi [--- si servu]s si liber hoc tulerit (vacat) [--- non il]li perṃittas ịn sangu(i)ne [---]ṣui ++[---]

Bibliography: Tab. Sulis 47; Adams 1992: 2–3; Kropp 2008: no. 3.2/39; Urbanová 2018: no. 263.

a defixio against thieves. We find the following linguistic and orthographic features in the text: the gemmination of x in exxigi (l. 2) and sangune for sanguine (l. 4). Note the use of the verb tollo as a synonym for the more common involo (for the form tulerit, see Adams [1992: 2–3] on late perfect forms of the tollo as well as the parallels in the curses from 337 and 523).

Image source: Tab. Sulis 47. Courtesy of R.S.O. Tomlin. Commentary: this fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207) and originally formed part of the top-right corner of a tablet. Despite being very corroded, we can still read five lines of text, which run from left to right and were written in old Roman cursive. Though it is quite fragmentary, the presence of certain formulae (i.e., the all-inclusive in l. 1 and non permittas in l. 4), allow us to classify this curse as

After being inscribed, the tablet was folded in half. Translation (Tab. Sulis 47): ‘… I complain [to you]… be exacted… [whether slave] or free, has taken his… you are [not] to permit [him] in blood… his…’

296

 Britannia 253. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 399] Material: lead alloy. Measurements: (4.4) × (4.9) cm. Reading: A [---]SAGILIANO+[---] [--- an]tequam in fa[no---] [---]+++[-c.4.-]LEF+[---] [---]+ER+[---] B [---]DESIMILI[---] [---]DIC+++[---] (vacat)

Bibliography: Tab. Sulis 48; Kropp 2008: no. 3.2/40.

left to right and were written in old Roman cursive by a trained hand. The circumstances that led to the writing of this curse are unknown. The phrase [an]tequam in fa[no] (l. 2) could form part of a magical time frame (on these, see Sánchez Natalías 2019a).

Image source: Tab. Sulis 48. Courtesy of R.S.O. Tomlin. Commentary: these two lead fragments were discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207) and formed part of the upper edge of a defixio. An opisthograph, the inscription contains the remains of four lines of text, which run from

After being inscribed, the tablet was folded. Translation (Tab. Sulis 48): ‘…before… in temple…’

254. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 399] Material: lead alloy. Measurements: vid. commentary. Reading:

[---]qu[i invola-] [vi]t cab[al-]lar[e]ṃ si[---] [--- si vir si f]ẹmin[a] si ser(v)us ++[---] [--- ?si libe]r [---]dea Sul[is---]

Bibliography: Tab. Sulis 49; Kropp 2008: no. 3.2/ 41; Urbanová 2018: no. 264.

has preserved three lines. All text runs from left to right and was written in old Roman cursive. Based on the preserved text, this curse should be grouped among the defixiones against thieves. Note that, the lexeme cab[al]lar[e]ṃ likely refers to a horse blanket similar to the one documented in Edictum de Pretiis Rerum Venalium, XIX, 33 (τάτης καβαλλαρικός).

Image source: Tab. Sulis 49. Courtesy of R.S.O. Tomlin. Commentary: these two fragments were discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207) and formed part of a single defixio. The first fragment, measuring (3.3) × (3.3) cm, has preserved part of the original upper edge and contains four lines of text. The second, measuring (3) × (1.6) cm,

Translation (Tab. Sulis 49): ‘(The person) who [has stolen] (my) horse blanket (?), whether [man or] woman, whether slave [or free]… goddess Sulis…’ 297

Sylloge of  Defixiones from the Roman West 255. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 619] Material: lead and tin. Measurements: (3.8) × (3.9) cm. Reading:

{d}deae [Suli---] Victorini[us ? or -anus---] +[-c.1-2-]MESP[---] [-c.3-]IT[-c.2-]++[---]

Bibliography: Tab. Sulis 50; Kropp 2008: no. 3.2/ 42.

corner of a defixio. The inscription contains the remains of four lines, which run from left to right and were written in old Roman cursive. The circumstances that led to the writing of this curse are unknown. The invocation of the patron goddess of the sanctuary appears to be followed by a Latin cognomen that could be reconstructed as Victorinus or Victorianus (see OPEL IV, 168–69).

Image source: Tab. Sulis 50. Courtesy of R.S.O. Tomlin. Commentary: these two small fragments were discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207) and made up the top left

256. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 3rd century CE. [Inv. No.: 206] Material: lead, tin and copper. Measurements: 5.3 × 11.3 cm. Reading:

Severa Dracontius Spectatus Innocentius 5 Senicio C̣andidianus Applicius Belator Surilla 10 Austus Carinianus

298

 Britannia Bibliography: Hassall and Tomlin 1981: 372–74; AE 1982, 659; Tab. Sulis 51; Kropp 2008: no. 3.2/43; Urbanová 2018: no. 190.

in the nominative case. Two of these names are Celtic: Belator and Austus (ll. 8 and 10, respectively). The rest are Latin names: Senicio (l. 5, see OPEL IV, 66), Candidianus (l. 6, see OPEL II, 30), Carinianus (l. 11; see OPEL II, 37), while Dracontius (l. 2, from draco, see OPEL II, 109) and Innocentius (l. 4; see OPEL II, 194) are common in the third century CE. Applicius (l. 7) is first found here and probably is derived from the Latin verb applico. This curse has been dated to the third century CE on palaeographic and onomastic grounds.

Image source: Tab. Sulis 51. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a roughly rectangular sheet made of lead, tin and copper. The inscription contains 11 lines, which run from left to right and were written in old Roman cursive.

Translation (Tab. Sulis 51): ‘Severa, Dracontius, Spectatus, Innocentius, Senicio, Candidianus, Applicius, Belator, Surilla, Austus, Carinianus.’

The circumstances that led to the writing of this curse are unknown. The text consists of a list of personal names

257. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 689] Material: lead, tin and copper. Measurements: (7.2) × (6.1) cm. Reading:

5



[--- perm]ittas [--- somn]um nec sanita[tem --- n]isi tandiu TA[---]IAT quandiu hoc [--- ill]ud se habuerit [--- s]i vir si femina et [---] si ancilla

Bibliography: Tomlin 1984: 338; Tab. Sulis 52; Kropp 2008: no. 3.2/44; Urbanová 2018: no. 265.

us to group this curse among the defixiones against thieves. Furthermore, the editor has underscored the structure tandiu… quandiu, hoc… [ill]ud (ll. 3–5), a magical time frame, which ‘may have had the sense of “so long as he shall live” (…) or “so long as he shall retrain the stolen property”’ (Tomlin 1988a: 178; for magical time frames, see Sánchez Natalías 2019a).

Image source: Tab. Sulis 52. Courtesy of R.S.O. Tomlin. Commentary: this fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207) and belonged to the lower-right corner of a rectangular tablet. Seven lines of text, which run from left to right and were written in old Roman cursive, have been preserved.

After being inscribed, the tablet was pierced with an iron nail at l. 3. Translation (Tab. Sulis 52): ‘…you are [not] to permit [… sleep] or health… except for as long as… it shall find itself (?)… whether man or woman and… or slave woman.’

Although the text is rather fragmentary, the presence of a non permittas formula (ll. 1–3) and an all-inclusive formula (ll. 5–6, for si ancilla, see 449 for a parallel) lead

299

Sylloge of  Defixiones from the Roman West 258. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 601] Material: lead and tin. Measurements: 6.5 × 4.9 cm. Reading: A dẹ[o] Ṃerc ̣urio D+[-c.2-]ISAMVALV++ DIN[-c.2-3-]ITSÂNV++ [C]ivilis++fuerit de 5 +VNIIA+IAMT+T++[---] Trinni (?) familiam [---] Velvalis (?)[---] am sụam (vacat)

B Markelinum familia[m] Velorigam et famili[am] [s]uam Morivassum et [f]ạṃiliam Riovassum ẹ[t] 5 familiam Minoven+++ ̣ et familiam sua[m] [---]++[-c.1-2-]M[-c.1-]DES++

Bibliography: Tab. Sulis 53; Reynolds 1990: 381; Kropp 2008: no. 3.2/45; Urbanová 2018: no. 191.

while Markelinum (B, l. 1) seems to be a vulgarization of Marcellus (see OPEL III, 54–55). We also find the Celtic names Veloriga (B, l. 2, formed from the name element Vell- and the suffix -riga), Morivassus (B, l. 3, from the elements mori- and *uassum) and Riovassus (B, l. 4, derived from *uassos and ri(g)o), which are all found here for the first time. For Tomlin, Velvalis (A, l. 7) is another previously unknown Celtic name, which can be compared to Velvinna and Velva. Reynolds has conjectured valetudo for A, ll. 2–3.

Image source: Tab. Sulis 53. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a roughly oval sheet. The tablet has irregular edges and its surface is quite uneven and filled with imperfections. An opisthograph, the inscription contains 15 lines of text, which run from left to right and were written in old Roman cursive.

After being inscribed, the tablet was folded once and perhaps pierced with a nail at A, l. 2.

The circumstances that led to the writing of this curse are unknown. After invoking Mercury (this is unique in the corpus from Bath, but see 238 where Mars is summoned), the text targets a series of victims and their families (cf. 246 and 370). The name Trinni (A, l.6) is first attested here and may be a variant of Trenus,

Translation (Tab. Sulis 53): ‘To the god Mercury… Civilis… shall have been… family of (?) Trinnus… his… Mar(c)el(l)inus (and his) family. Veloriga and her family. Morivassus and (his) family. Riovassus and (his) family. Minoven… and his family…’

300

 Britannia 259. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 20,004] Material: lead, tin and copper. Measurements: 14.9 × 9.6 cm. Reading: A B+ B ++AT+AD+ITAMO[---] c ̣onq[u]er[or] tibi Sulis Armịnịa (u)t Verecundinum (?) Ter[en]ṭi c ̣[ons]ụmas qui argeṇtiolọs duos mihi +++ 5 MANDICE[-c.5-]revavit no[n il]l[i p]er mittạs nec sedere nec iacere [ne]c ̣ (traces) a[m]bulare n[ec] ṣomn[um nec] sanitatem [?] c ̣ụṃ quantocius coṇṣumas et itẹr[u]ṃ 10 DET+AESTACT (traces) [-c.2-]NVS++ [-c.2-3-]MENSI+ION (traces) ṇọn perveniat (vacat)

Bibliography: Tab. Sulis 54; Kropp 2008: no. 3.2/46; Urbanová 2018: no. 266.

initial invocation and the denouncement of the theft, the text contains a non permittas formula (B, ll. 5–9), with which the defigens seeks to harm the victim’s well-being. This formula asks for the goddess not only to deprive the thief of sleep and health but also to prevent him from sitting, lying down and walking (sedere, iacere [cf. 365] and ambulare [cf. 352]). The curse ought to take effect quantocius (B, l. 9), a magical time frame similar to that documented in 452 and 367 (for magical time frames, see Sánchez Natalías 2019a). The curse contains the name of the defigens, Arminia (B, l. 2, for the cognomen Arminius see OPEL I, 171). What is more surprising is that the text also names the thief: Verecundinum Ter[en]ṭi (B, l. 3, for Verecundinus, see OPEL IV, 157–58; the patronymic is widely documented in Britannia; see OPEL IV, 113).

Image source: Tab. Sulis 54b. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on an irregularly shaped sheet made of lead, tin and copper. The tablet is quite corroded, especially along the right edge of side B. An opisthograph, the inscription contains 13 lines of text, which run from left to right and were written in old Roman cursive. Based on its content, this curse should be grouped among the defixiones against thieves. Side A of the tablet only contains one line in which we read the letter b followed by a cross. This symbol was made by the defigens and resembles several pseudo-inscriptions (see 321). For Tomlin, it could be a mark to show ownership (for a similar case see 448, side A), although his proposal cannot be proven. On side B, 12 lines of text describe how Arminia reports to Sulis the theft of two silver coins. After the

Translation (Tab. Sulis 54): ‘…I, Arminia, complain to you, Sulis, (that) you consume Verecundinus (son of) Terentius, who has (stolen…) two silver coins from me. You are not to permit (him) to sit or lie (or… or) to walk (or) (to have) sleep (or) health, (since) you are to consume (him) as soon as possible; and again… (not) to reach…’ 301

Sylloge of  Defixiones from the Roman West 260. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 462] Material: lead and tin. Measurements: 6.3 × (3.1) cm. Reading:

Dacuṣ AGA AQVER+[---] N[-c.2-3-]T[---] recentis(s)imi+++[---] 5 capitulare«m» [---] EM SVPPLIC+[---] si ser[vus si liber---] sọmnus[---] +V[---] 10 ++[---] [+]GE+[---] (traces) (traces) (traces)

Bibliography: Tab. Sulis 55; Kropp 2008: no. 3.2/47.

Despite the fragmentary nature of the text, the presence of an all-inclusive formula (l. 7) used to identify the victim as well as the lexeme sọmnus (l. 8), which is common in non permittas formulae, lead us to group this curse among the defixiones against thieves. Dac(c)us (l. 1, a Celtic personal name) denounces the theft of a capitularis (l. 5), an article of clothing similar to a hat or hood (for parallels, see 343 and 441).

Image source: Tab. Sulis 55. Courtesy of R.S.O. Tomlin. Commentary: these three fragments were discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207) and formed the left third of a tablet that was probably rectangular in origin. Though it is an opisthograph, one of the sides has been so corroded that it is practically illegible. The preserved portion of the text contains 14 lines, which run from left to right and were written in old Roman cursive by a trained hand.

After being inscribed, the tablet was folded twice.

261. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: RBS 79 CS 3] Material: lead alloy. Measurements: 2.4 × 1.6 cm. Reading:

[---]+INETIONIS SORVI+[---] [---]STA fraudem[---] [---]++ab[---]

the archaeological context, see 207) and formed the upper edge of a defixio. An opisthograph (side B, however, is illegible), the inscription preserves the remains of three lines of text, which run from left to right and were written in old Roman cursive. The words have been separated from one another by the inclusion of small spaces.

Bibliography: Tab. Sulis 56; Kropp 2008: no. 3.2/48. Image source: Tab. Sulis 56. Courtesy of R.S.O. Tomlin. Commentary: these three small fragments were discovered in the British sanctuary of Sulis Minerva (for 302

 Britannia 262. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 419] Material: lead, tin and copper. Measurements: 6.6 × 5.1 cm. Reading:

5

deae ++L++ ṃịṇẹ+ Exib[uus? ---] donaṿ[it] i[l]los qui ++BAN[---] DE[-c.2-]VON+[-c.1-]INELO[-c.2-] sunt [-c.3- si servus] ṣị l[iber si bar]o si m[u]l[i]e[r SA[---]

Bibliography: Tab. Sulis 57; Adams 1992: 15–17; Kropp 2008: no. 3.2/49; Urbanová 2018: no. 267.

Despite the fragmentary state of the text, the use of an all-inclusive formula (ll. 3–4, for baro, see Adams 1992: 15–17) suggests that this curse should be grouped among the defixiones against thieves. After the initial invocation (deae, l. 1), the practitioner hands the guilty party over to the deity. If we accept the reconstruction of the word sa[nguine(m)] (l. 5), we can assume that the victim was to pay for his or her crime in blood (this is common at Bath: for some parallels, see 271, 299, etc.).

Image source: Tab. Sulis 57. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was broken into seven fragments that fit together and belonged to the top-left corner of a tablet that was probably rectangular. Although it has been seriously corroded, the inscription preserves the remains of five lines, which run from left to right and were written in old Roman cursive.

After being inscribed, the tablet was folded several times. Translation (Tab. Sulis 57): ‘To the goddess… Exibuus has given those who… are… [whether slave] or [free, whether man] or woman…’

263. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 590] Material: lead, tin and copper. Measurements: (10) × (15.9) cm. Reading: A A BC ER NI B (traces) (traces) (traces) [---]ESVEB (traces) 5 [---]NO (traces) [---]N+CI++ (traces) [---] ++ CVRILLAS+ [---] [---]S++++[-c.2-]R si+[---] [------] 10 [---]+invo[lavit or -erit ? ---] [---]LO[---]

303

Sylloge of  Defixiones from the Roman West Bibliography: Tab. Sulis 58; Kropp 2008: no. 3.2/50.

Despite the difficulties provided by such a tablet, the editor has proposed the reading of an all-inclusive formula (B, l. 8: [si servus] s[i libe]r si…) and a form of the verb involo (B, l. 10). This leads us to believe that this curse should be grouped among the defixiones in fures. As far as ABCERNI (A, ll. 1–4) goes, there are several possibilities of interpretation: an alphabetic sequence (maybe similar to 207), a personal name (although Abcernus is not attested), or a code known only to the writer. Given the fragmentary nature of the text, none of these possibilities can be proven.

Image source: Tab. Sulis 58b. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was broken into three fragments that fit together and form an irregularly shaped tablet. An opisthograph, the inscription contains 15 lines of text, which run from left to right and were written in capitals (side A) and old Roman cursive (side B). The poor state of conservation precludes a reconstruction of the text.

After being inscribed, the tablet was folded in thirds (first the left side and then the right).

264. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 596] Material: lead and tin. Measurements: (3.6) × (8.1) cm. Reading:

Enica c ̣ọṇqueror ti[bi?---] [---]+T anulis[-c.1-2-] DEHI+[---] [---]+++[-c.2-?]STAT[-c.2-]AM[---] [---]S[---]

Bibliography: Tab. Sulis 59; Kropp 2008: no. 3.2/51.

The circumstances that led to the writing of this curse are unclear, though the mention of some anulis (l. 2) suggests the context of a defixio against thieves. Enica is a feminine personal name of Celtic origin, which is attested as a women’s name in the regio IX (cf. CIL V, 7641). For the phrase c ̣ọṇqueror ti[bi] (l. 1), see 259 and 365.

Image source: Tab. Sulis 59. Courtesy of R.S.O. Tomlin. Commentary: these two fragments, which were discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), made up the top-left corner of the tablet that was originally rectangular. The inscription contains four lines, which run from left to right and were written in old Roman cursive.

Translation (Tab. Sulis 59): ‘I, Enica, complain to you...’

265. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 600] Material: lead, tin and copper. Measurements: 2.9 × (5.9) cm. Reading: Ọc`o´nea deae Suli M(inervae) dọno [ti]bi pạ nnum+ si quis eum (traces)

304

 Britannia Bibliography: Tab. Sulis 60; Kropp 2008: no. 3.2/52; ­Urbanová 2018: no. 268.

Oconea denounces the theft of a pannus (ll. 2–3). This tablet is the first from Bath to abbreviate the goddess’ name as Sulis M(inervae) (ll. 1–2). The lexeme pannus is odd, since it ‘usually means “rag” or “patch”, an unlikely object for theft’ (Tomlin 1988a: 191). Accordingly, we may want to conjecture panna (a cooking vessel such as the one in 271) instead. Oconea is a Celtic name without attested parallels.

Image source: Tab. Sulis 60. Courtesy of R.S.O. Tomlin. Commentary: these two fragments, which were discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), made up the upper edge of a tablet that was originally rectangular. The inscription contains four lines, which run from left to right and were written in old Roman cursive.

After being inscribed, the tablet was folded. Translation (Tab. Sulis 60): ‘Oconea to the goddess Sulis M(inerva) I give you a pan (?). If anyone [has stolen] it…’

Based on its content, this curse should probably be grouped among the defixiones against thieves. Here,

266. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 664] Material: lead and tin. Measurements: 4.6 × 5.5 cm. Reading:

5



Lovernisca d[onat] eum qui sive v[ir] {i}sive femina s[i]ve puer sive puella qui ṃafortium ̣ «ị[n]ṿola»verit

Bibliography: Tomlin 1984: 337; AE 1984, 622; Tab. Sulis 61; Kropp 2008: no. 3.2/53; Urbanová 2018: no. 269.

Based on its content, this curse should be grouped among the defixiones in fures. Here, Lovernisca denounces the theft of an article of clothing (a ṃafortium, l. 5) to Sulis. ̣ Tomlin has connected this lexeme to Greek μαφόρτιον, which is documented in Greek papyri from the third century CE, and to mafurtium, which in Nonius Marcelus XIV is used in place of the more common ricinium (‘veil’). The text contains an all-inclusive formula (ll. 2–4; where we find sive in place of the expected si) used to identify the victim. The name Lovernisca (l. 1) is a Celtic feminine name derived from *louerno (‘fox’).

Image source: Tab. Sulis 61. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a rectangular sheet, which today is broken into two parts that fit together. There are several breaks in the sheet, which is now missing its top-left and bottom-right corners. The inscription contains six lines, which run from right to left and were written in old Roman cursive by a trained hand. The letters are ‘mirrored’ (i.e., written from right to left), which has led to some copying errors.

Translation (Tab. Sulis 61): ‘Lovernisca [gives] him who, whether [man] or woman, whether boy or girl, ‹who› has stolen (her) cape.’

305

Sylloge of  Defixiones from the Roman West 267. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: RBS 80 CS 3] Material: lead alloy. Measurements: (5.6) × (6.5) cm. Reading:

5



[-c.4-]ẹọcorotis perdedi lạ[enam] [pa]lleum sagum ̣ paxsam do[navi] [-c.2-]+[S]ụlis ut hoc ante dies novem [si li]ber si ser(v)us si [li]bera si serva [si] pu˹er˺ si puelḷ[a i]n rostr[o] s[uo] defera[t ---] caballarem ṣị [ser(v)us si liber si] serva si libera si pụer [si puella] in suo rostro defer[at] (vacat?)

Bibliography: Tomlin 1987 b: 364–67; Tab. Sulis 62; Kropp 2008: no. 3.2/54; Urbanová 2018: no. 270.

the Greek τάτης καβαλλαρικός, a word lost in Latin but found at Edictum de Pretiis Rerum Venalium, XIX, 33). Having ‘lost’ three different cloaks, the practitioner is careful to describe each one in the most precise detail possible so that Sulis can find them within the delimited magical time frame: ante dies novem (l. 3, see Reynolds 1990: 381–82, on the relationship with the nundinae; also, see Marco Simón 2010b; for magical time frames in general, see Sánchez Natalías 2019a). For parallels, see 450, 456, etc.

Image source: Tab. Sulis 62. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), has been broken into two parts that fit together and formed a good deal of the upper edge and lateral sides of the tablet. The inscription contains nine lines, which run from right to left and were written in old Roman cursive.

The curse’s all-inclusive formula, which is used to identify the victim, is noteworthy: after the normal [si li]ber si ser(v)us… (ll. 4–5), it is found again with the gender changed (si [li]bera si serva, ll. 7–8). The phrase in rostro suo deferat (ll. 5–6 and 9) has been a matter of debate: for the editor, it refers to the thief’s ‘snout’ (OLD 1), whereas Reynolds has translated it as ‘on its peg’ or ‘in its bench’. The editor’s proposal is more suitable in the context of return of the stolen property to the temple. Finally, -ẹọcorotis (l. 1) is probably part of a Celtic personal name.

Based on its content, this curse should be grouped among the defixiones against thieves. Here, the theft of several articles of clothing (a tunic, three cloaks and a horse blanket) is reported. It appears that the layout of the text (from right to left) led to several of orthographic errors, like pure for puer (l. 5) as well as the author’s subsequent correction of several letters (ll. 4, 5, 7 and 8). Furthermore, the writer increased the size of letters at the end of the lines so as to fill the extra space. Also, note serus for servus (l. 4), and the use of the verb perdo (l. 1) instead of the more common involo.

After being inscribed, the tablet was folded once and perhaps pierced with a nail.

One of the stolen objects was a sagum, which Wild notes ‘is the Gallic native/military cloak par excellence while the pallium is the Greek national equivalent, carrying the aura of intellectual life (...) into the Roman world. The laena is an old-fashioned term for an Italian cloak’ (apud Tab. Sulis 62). The fourth purloined object is a paxsa (cf. 237 for a parallel), while the fifth is a caballarem (cf. 254), which Tomlin has connected to

Translation (Tab. Sulis 62): ‘I, […]eocorotis, have lost (my) Italian/Greek/Gallic cloak and (tunic), (which) I have given… Sulis, that (he) may bring it down in his snout before nine days, [whether] free or slave, whether free woman or slave woman, [whether] boy or girl,… horse blanket, [whether slave or free, whether] slave woman or free woman, whether boy [or girl], bring down in his snout.’

306

 Britannia 268. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 697] Material: lead, tin and copper. Measurements: 5 × 8.3 cm. Reading:

5



deae Suli (vacat) si quis balniarem Cantissen(a)e inv[o]la[v]erit si ser(v)us ṣi liber (traces) (traces) MENA (traces)

Bibliography: Tab. Sulis 63; Kropp 2008: no. 3.2/55; Urbanová 2018: no. 271.

(ll. 2–3, a bathing garment) is reported. The text presents the following orthographic features: balniarem for balnearem (l. 2), Cantissene for Cantissenae (l. 3) and serus for servus (l. 4). Balnearem is an adjective, which should be understood as modifying an omitted paxsam (see 237 for a parallel). The text also has the usual all-inclusive formula (l. 4) meant to help identify the victim. Cantissena (l. 3) is a Celtic personal name (probably for a man) that is unparalleled.

Image source: Tab. Sulis 63. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on an irregularly shaped sheet made of lead, tin and copper. The tablet’s surface is wrinkled and has been corroded, especially the bottom section. Accordingly, the final two lines cannot be read. The inscription contains six lines, which run from left to right and were written in old Roman cursive.

After being inscribed, the tablet was folded once. Translation (Tab. Sulis 63): ‘To the goddess Sulis. If anyone has stolen the bathing tunic of Cantissena, whether slave or free,…’

Based on its content, this curse should be grouped among the defixiones in fures. Here, the theft of a balniarem

269. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 3rd century CE. [Inv. No.: 417] Material: lead, tin and copper. Measurements: (5.3) × 9.8 cm. Reading: (traces) quịeṣcit +++LIT sanitatem invictus ṇisi eịdem loco ipsum pallium [re]duc ̣at (vacat)

307

Sylloge of  Defixiones from the Roman West Bibliography: Tab. Sulis 64; Kropp 2008: no. 3.2/56; ­Urbanová 2018: no. 272.

Based on its content, this curse should be grouped among the defixiones in fures. We must highlight the use of the lexemes quịeṣcit (l. 2), and invictus (l. 2). The latter logically cannot be applied to the thief unless it is an error for victus. The presence of sanitatem (l. 2) likely alludes to the common non permittas formula used to harm the victim’s physical well-being until the stolen object is returned (ipsum pallium [re]duc ̣at, ll. 3–4).

Image source: Tab. Sulis 64. Courtesy of R.S.O. Tomlin. Commentary: these five fragments were discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207) and formed the bottomright corner of a tablet, which originally must have been rectangular. Four lines of text have been preserved that run from left to right and were written in new Roman cursive that can be dated to the third century CE on palaeographic grounds.

After being inscribed, the tablet was folded five times and cut with a knife. Translation (Tab. Sulis 64): ‘… rests… health unconquered unless he brings the cloak itself back to the same place.’

270. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 275–400 CE. [Inv. No.: 671] Material: lead, tin and copper. Measurements: 8.8 × 5 cm. Reading:

5



10



Minerv(a)e de(ae) Suli donavi furem qui caracallam meam involavit si ser(v)uṣ si liber si baro si mulier hoc donum non redemat nessi sangu(i)nẹ suo

Bibliography: Tomlin 1983: 336, 340; AE 1983, 636; Tab. Sulis 65; Adams 1992: 15–17; Kropp 2008: no. 3.2/79; Urbanová 2018: no. 278.

last four lines are squashed together and l. 11 runs up the margin. For what it is worth, according to a pediatrician consulted by the editor, the author of the text may have suffered from dyslexia (see 211 for a parallel).

Image source: Tab. Sulis 65. Courtesy of R.S.O. Tomlin.

Based on its content, this curse should be grouped among the defixiones in fures. After the denunciation of the theft of an article of clothing, the guilty party is handed over to the goddess Sulis. The text presents the following orthographic features: Minerve for Minervae (l. 1), de for deae (l. 2), serus for servus (l. 6), nessi for nisi (l. 10), sangune for sanguine (l. 11) and redemat for redimat (l. 10). The tablet contains a typical all-inclusive formula meant to identify the victim (ll. 6–8) with the attested variant si baro si mulier

Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a rectangular sheet, which is almost complete. The small breakage found along the upper edge is surely ancient, since the practitioner avoided it when writing the text. The inscription contains 11 lines, which run from left to right and were written in a combination of old and new Roman cursive. Note that the 308

 Britannia Translation (Tab. Sulis 65): ‘To Minerva the goddess Sulis I have given the thief who has stolen my hooded cloak, whether slave or free, whether man or woman. He is not to buy back this gift unless with his own blood.’

(for baro, see Adams’ commentary 1992: 15–17). Also note the victim’s impossibility of escaping punishment expressed with the typical formula non redemat nessi sangu(i)nẹ suo (ll. 10–11).

271. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 275–400 CE. [Inv. No.: 614] Material: lead and tin. Measurements: 13.4 × 9.9 cm. Reading:

5



10



Exsuperịus donat pannum ferri qui ilḷi innọc[entiam?]NFAM TVSC+++[---] Sulis si vir [si femin]a s[i] ser(v)us si liber họ[c ---] iḷḷ[---] et[-c.1-]+[-c.1-]ER++[---] suạs invọla[veru]n[t] s[i] vir si femina ṣ[ati]sfecerit ̣ «san»guịṇ[e] ill[o]rum học devindic ̣ẹṣ ṣ[i?] (vacat) q[u]is ạenụṃ mịhi inṿọlav[i]t (vacat)

Bibliography: Tab. Sulis 66; Reynolds and Volk 1990: 381; Kropp 2008: no. 3.2/57; Urbanová 2018: no. 273.

inclusive formula with si vir [si femin]a… (ll. 5–6, 9–10), which is used to identify the victim. Also note the formula asking for the thief to pay for his crime in blood: ṣ[ati]sfecerit «san»guịṇ[e]. The more usual formula uses a form ̣ of the verb redimo but see 299 for a parallel. Reynolds, for her part, has conjectured the reading in noc[te] instead of innoc[entiam?] (l. 3). The reading innoc[entiam?], however, is to be preferred, since this lexeme is attested elsewhere in the Bath corpus (305 A, ll. 4–5 where the reading is secure). Exsuperius (l. 1) is a well attested name between the third and fourth centuries CE (see OPEL II, 131).

Image source: Tab. Sulis 66. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was broken into five parts that fit together, was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207). The tablet, originally rectangular, has lost a good amount of its central portion. The inscription contains 13 lines, which run from left to right and whose lines transition from new Roman cursive (l. 1) to old Roman cursive (ll. 2–13). In Tomlin’s appraisal, the tablet ‘is much more of a palaeographical embarrassment, since it blurs the two categories into which Latin cursive writing is divided’ (Tomlin 1988a: p. 200).

After being inscribed, the tablet was folded nine times. Translation (Tab. Sulis 66): ‘Exuperius gives an iron pan (?). (The person) who… innocence for him… of (?) Sulis, whether man [or woman], whether slave or free, … this… and… have stolen his…, whether man or woman, is to have given satisfaction with their blood. You are to reclaim (?) this [if] anyone has stolen the vessel from me.’

Based on its content, this curse should be grouped among the defixiones against thieves. Here, Exsuperius denounces the theft of a pannum ferri (ll. 1–2; note pannum for pannam and see 265 for a parallel). The tablet contains a classic all309

Sylloge of  Defixiones from the Roman West 272. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: RBS 80 CS 3] Material: lead alloy. Measurements: (3.2) × (2.1) cm. Reading:

nec ṣ[omnum?---] N+[---] T+[---]

Bibliography: Tab. Sulis 67; Kropp 2008: no. 3.2/58.

the remains of three lines of text, which run from left to right and were written in old Roman cursive. Despite the fragmentary nature of the text, Tomlin has conjectured the formula nec somnum nec sanitatem.

Commentary: this small fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207) and preserves part of the upper edge and

273. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: RBS 80 CS 3] Material: lead. Measurements: (2.4) × (4) cm. Reading:

[---]in SACL+[---] [---]hoc invọḷạ[vit or -verit?---] [---]++++[---]

Bibliography: Tab. Sulis 68; Kropp 2008: no. 3.2/59.

together and has lost all its original edges. The inscription contains the remains of three lines of text, which run from left to right and were written in old Roman cursive. Despite the fragmentary nature of the text, the verb involo (l.2) suggests that this is a defixio against thieves.

Commentary: this small fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207). It has been broken into two pieces that fit

274. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: RBS 80 CS 3] Material: lead. Measurements: (2.7) × (2.4) cm. Reading:

[---]+IA (vacat) [---] dea Suliṣ [---] [---]NEM ++[---]

Bibliography: Tab. Sulis 69; Kropp 2008: no. 3.2/60.

fit together, which form part of the upper edge of a larger sheet. The inscription contains the remains of four lines of text, which run from left to right and were written in old Roman cursive.

Commentary: this small fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207). It has been broken into two pieces that

275. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: RBS 80 CS 3] Material: lead alloy. Measurements: (2.2) × (2.9) cm. Reading:

Minerv[a or -ae---] AM C̣ocuṣ[---] LVM PELL[---]

310

 Britannia Bibliography: Tab. Sulis 70; Kropp 2008: no. 3.2/61.

a tablet. The fragment preserves three lines, which run from left to right and were written in old Roman cursive. The editor has picked out the Celtic name Coccus (l. 2) as well as the theonym of the sanctuary’s patron deity (l. 1).

Commentary: this fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207) and belonged to the top-left corner of

276. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 602] Material: lead alloy. Measurements: 6.4 × 8 cm. Reading: (traces) (traces) [---]A [-c.3-4-]qui SVIB+[---] (traces) [--- s]ị se[rv]us s[i ---] 5 ++L+[---] si ṿir si [f]em[ina ---] [---]+discẹbịt+[---] (traces)

Bibliography: Tab. Sulis 71; Kropp 2008: no. 3.2/62.

defixiones in fures. The editor has also signalled out the form discẹbịt (l. 6). As he points out, ‘[i]f this is intended as the future of disco, it is a “Vulgar” confusion between conjugations (2nd for 3rd). The verb disco is unattested in curse tablets, but just possibly equivalent to inveniat’ (Tomlin 1988a: 226).

Image source: Tab. Sulis 71. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), has been broken into two pieces that fit together. The tablet is quite corroded and has lost its original edges and much of its bottom third. It conserves seven lines, which run from left to right and were written in old Roman cursive.

After being inscribed, the tablet was folded once. Translation (Tab. Sulis 71): ‘…whether slave or [free]… whether man or woman… will learn (?) …’

Though quite fragmentary, the presence of an all-inclusive formula (ll. 4–5) allows us to group this text among the

277. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 619] Material: lead alloy. Measurements: (2) × (2.7) cm. Reading: (traces) [-c.2-]+ORI+A[---] tacituri[---] + 311

Sylloge of  Defixiones from the Roman West Bibliography: Tab. Sulis 72.

into two pieces that fit together. They formed part of the tablet’s left edge. We find the remains of four lines, which run from left to right and were written in old Roman cursive. The lexeme tacituri (l. 3) likely refers to the desire that the victim remain silent.

Commentary: this small irregularly shaped fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207) and has been broken

278. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 619] Material: lead alloy. Measurements: (2.2) × (2.2) cm. Reading:

[---]+veṇịạ[t?---] noṇ AN++[---] [---]+[---]

Bibliography: Tab. Sulis 73; Kropp 2008: no. 3.2/63.

(for the archaeological context, see 207). It preserves the remains of three lines of text, which run from left to right and were written in old Roman cursive.

Commentary: this small, irregularly shaped fragment was discovered in the British sanctuary of Sulis Minerva

279. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 207] Material: lead, tin and copper. Measurements: 2.9 × 4.4 cm. Reading:

O[-c.1-2-]O+ DOMX

Bibliography: Tab. Sulis 74; Kropp 2008: no. 3.2/64.

contains two lines, which run from left to right and were written in a mix of cursive (l. 1) and capitals (l. 2). DOMX finds a parallel in 154 (l. 1), where it has been interpreted as an abbreviation for domina.

Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on an irregularly shaped sheet, which appears to be entirely preserved. The inscription

After being inscribed, the tablet was folded once.

280. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: RBS 79 CS 3] Material: lead. Measurements: (3.4) × (8) cm. Reading: (traces)O++[-c.2-3-]O+D+ (?) (traces, vacat)++++++M (?) (traces, vacat)++RIAM (?) (traces, vacat)+++ Bibliography: Tab. Sulis 75.

belonged to the bottom portion of a curse. Very corroded, the fragment preserves four lines written in old Roman cursive, but the letters are so poorly formed that the text is almost completely illegible.

Commentary: this small, irregularly shaped lead fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207) and originally

312

 Britannia 281. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: RBS 80 CS 3] Material: lead. Measurements: (1.9) × (1.7) cm. Reading:

5

[---]IQAS [---]+TVS+IVS [---]IIRVLV [---]VI [---]+E+RNFI

the archaeological context, see 207). It preserves part of the upper edge of the original tablet and the remains of five lines, which run from left to right and were written in old Roman cursive. The text lacks any sense and it is not possible to offer a reconstruction.

Bibliography: Tab. Sulis 76. Image source: Tab. Sulis 76. Courtesy of R.S.O. Tomlin. Commentary: this small rectangular lead fragment was discovered in the British sanctuary of Sulis Minerva (for

282. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 683] Material: lead. Measurements: 8.1 × 6.5 cm. Reading: (traces) Cupit[i]aṇ[us? ---] ++AVISIVS(vacat)++[---] (vacat)

Bibliography: Tab. Sulis 77; Kropp 2008: no. 3.2/65.

which today is quite corroded and has been broken into two pieces that fit together. It conserves the remains of three lines, which run from left to right and were written in old Roman cursive. Tomlin has proposed the reconstruction of the personal name Cupitianus (l. 2, see OPEL II, 87–88).

Image source: Tab. Sulis 77. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on an irregular lead tablet,

After being inscribed, the tablet was folded twice.

313

Sylloge of  Defixiones from the Roman West 283. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 20.019] Material: lead alloy. Measurements: (6.7) × (5.3) cm. Reading: [--- ? B]itilus Linu[s ---] (?)Bit[i]l{u}us Lin[us ---] (vacat)

Bibliography: Tab. Sulis 78; Kropp 2008: no. 3.2/66; Urbanová 2018: no. 192.

and were written in old Roman cursive by different hands (note, for instance, the different trace of the sequence us in each line). The lines seem to repeat the same personal name, which has been reconstructed as Bitillus (a diminutive from the Celtic Bitus) and the Latin cognomen Linus (see OPEL III, 28).

Image source: Tab. Sulis 78. Courtesy of R.S.O. Tomlin. Commentary: this small irregularly shaped fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207). It preserves the remains of two lines of text, which run from left to right

After being inscribed, the tablet was folded.

284. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 416] Material: lead, tin and copper. Measurements: (3.9) × 7.9 cm. Reading: (traces) rẹs MCA+[---]ESIL`A´ Q[-c.2-3-]MALI EGNENV+ A[v]iṭị[us (?) -c.1-2-]+ Bibliography: Tab. Sulis 79.

left to right and were written in old Roman cursive by a trained hand. Despite difficulties in reading the text, it is possible to reconstruct the nomen Avitius (l. 2; vid. OPEL I, 231).

Commentary: this fragment, which comes from the British sanctuary of Sulis Minerva (for the archaeological context, see 207), has been broken into three pieces that fit together and once formed the bottom of a rectangular tablet. It preserves the remains of three lines of text, which run from

After being inscribed, the tablet was folded four times.

285. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 399] Material: lead alloy. Measurements: (3.1) × (4.1) cm. Reading:

[---]L[---] [---]DOIL[---] [---]+++MODA+[---] 314

 Britannia Bibliography: Tab. Sulis 80.

fit together. It preserves the remains of three lines of text, which run from left to right and were written in old Roman cursive.

Commentary: this fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207) and was broken into two pieces that

After being inscribed, the tablet was folded once.

286. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: RBS 80 CS 3] Material: lead alloy. Measurements: (1.4) × (2.3) cm. Reading:

[---] [---]ụt quand[iu ---] [---]+[---]

Bibliography: Tab. Sulis 81; Kropp 2008: no. 3.2/67.

(for the archaeological context, see 207). It preserves the remains of three lines of text, which run from left to right and were written in old Roman cursive.

Commentary: this small irregularly shaped fragment was discovered in the British sanctuary of Sulis Minerva

287. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 20,003] Material: lead and tin. Measurements: (2.9) × (6.6) cm. Reading:

[---]+[---] [---]ṣol[vat? ---]

Bibliography: Tab. Sulis 82.

Roman cursive. If the editor’s reconstruction is correct, we may have a case of the formula sanguine suo solvat, which is well attested in Bath and which was used to demand payment for a stolen object with the victim’s own blood (for parallels, see 271, 337, etc.).

Commentary: this fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207). It originally belonged to the bottom edge of a rectangular tablet. It preserves the remains of two lines of text, which run from left to right and were written in old

After being inscribed, the tablet was folded.

288. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 175–275 CE. [Inv. No.: 399] Material: lead alloy. Measurements: vid. reading. Reading:

Frag. I ((1.9) × (2.1) cm) [---]+suuṃ[---] [---]+ụṭẹ[---]



Frag. II ((1.3) × (1.2) cm) [---]+EL [---]++[---]

Bibliography: Tab. Sulis 83.



Frag. III ((1) × (1.6) cm) [---]c ̣um[---]

The editor has provided a drawing and transcription of only the three best preserved pieces, whose texts are so fragmentary that we cannot offer a reconstruction. They were written in old Roman cursive.

Commentary: these three small fragments were discovered alongside another twelve fragments in the British sanctuary of Sulis Minerva (for the archaeological context, see 207). 315

Sylloge of  Defixiones from the Roman West 289. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: 670] Material: lead alloy. Measurements: (5.4) × (4.3) cm. Reading: [---]++ṇọc ̣+[---] [---]P[-c.1-]+[---] (vacat) +++ (vacat) + (traces) (traces) Bibliography: Tab. Sulis 84.

two pieces that fit together. Due to the level of corrosion, the remains of five lines, which run from left to right and were written in old Roman cursive, cannot be interpreted.

Commentary: this irregularly shaped fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207). It has been broken into

After being inscribed, the tablet was folded.

290–98. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: vid. commentary.] Material: lead alloy. Measurements: vid. commentary. Reading:

vid. commentary.

Bibliography: vid. commentary.

293 (=Tab. Sulis 88) [Inv. No.: RBS 80 CS 3]: this tablet consists of the remains of 40 small fragments, half of which are without any text, while others preserve one or two letters (written in old Roman cursive); others are illegible. One fragment, measuring (1.7) × (2.3) cm, was pierced by a nail, which has left a hole measuring some 3 mm in diameter.

Commentary: Tomlin describes this collection of fragments from the sanctuary of Sulis Minerva in a cursory fashion (for the archaeological context, see 207). Given the brief nature of the published account, I provide here a joint entry on these several curses in order to give each item a number in the present catalogue as well as provide bibliographic references (in parentheses), inventory numbers (in square brackets), the tablets’ dimensions together with a brief commentary.

294 (=Tab. Sulis 89) [Inv. No.: RBS 79]: this tablet consists of the remains of 50 lead fragments (without text or illegible), scraps of lead and a Gloria Exercitus coin that has been quite corroded.

290 (=Tab. Sulis 85; Kropp 2008: no. 3.2/68) [Inv. No.: RBS 79]: this fragment, measuring (2.2) × (2) cm, preserves the remains of four lines, which run from left to right and were written by a trained hand in old Roman cursive. The text reads: (traces)/ [---]+sum+[---]/ [---] ACVSAVE [---]/ [---]qu(o)d fra[udem fecit?]. For this final phrase, see 237.

295 (=Tab. Sulis 90; Kropp 2008: no. 3.2/71) [Inv. No.: 681]: this defixio, measuring approximately 3.2 × 10 cm, is folded seven times. The visible text, written in old Roman cursive, reads: d[eae Suli (?) ---]/ qu[i involavit (?) ---]/ ++[---]. 296 (=Tab. Sulis 91; Kropp 2008: no. 3.2/72) [Inv. No.: 613]: this curse, measuring 15 cm in width, is still folded three times and then doubled over. It preserves the remains of an inscription written in old Roman cursive on the outward facing side, which reads: [S]ecura P++[---]. Tomlin has taken this as part of a personal name.

291 (=Tab. Sulis 86; Kropp 2008: no. 3.2/69) [Inv. No.: RBS 79]: this fragment, measuring (1.7) × (1.6) cm, preserves the remains of three lines, which run from left to right and were written by a trained hand in old Roman cursive. The only legible portion comes from l. 2: [---]res invol[avit or -erit?].

297 (=Tab. Sulis 92) [Inv. No.: 699]: this lead fragment is very corroded and worn down. We can make out the remains of two lines written in old Roman cursive that read: (traces) / TA+[---].

292 (=Tab. Sulis 87; Kropp 2008: no. 3.2/70) [Inv. No.: RBS 80 CS 3]: this fragment, measuring (1.4) × (2.1) cm, preserves the remains of three lines, which run from left to right and were written in old Roman cursive. They read: [---]ili[---]/ [si s]er(v)us[---]/ (traces). The editor has interpreted l. 2 as containing an all-inclusive formula.

298 (=Tab. Sulis 93) [Inv. No.: 480]: fragment of a lead tablet that is still folded. It measures (2.2) × (3.2) cm. It is an opisthograph, but we can only make out some illegible traces of letters written in old Roman cursive. 316

 Britannia 299. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 275–400 CE. [Inv. No.: 618] Material: lead, tin and copper. Measurements: (5.5) × (7.5) cm. Reading:

5



Uricalus Dọ[c]ịlosa ux[or] ṣua Docilis filius suus et Docilina Decentinus frater suus Alogiosa nomina{a} eorum qui iuraverunt {qui iuraverunt} ad fontem deae Suli(s) prid(i)e idus Apriles quicumque illic periuraverit deae Suli facias illum ṣạṇguine suo illud satisfacere ̣

Bibliography: Hassall and Tomlin 1981: 375, 378–79; AE 1982, 661; Tab. Sulis 94; Adams 1992: 10; Kropp 2008: no. 3.2/73; Tomlin 2017c: 12.35; Urbanová 2018: no. 274.

Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a rectangular sheet made of lead, tin and copper. The curse is in excellent condition, though it has lost its left and upper edges. Furthermore, the top-right corner is somewhat corroded, as I was able to confirm during my autopsy (December 2009). The inscription contains eight lines, which run from left to right and were written in new Roman cursive by a trained hand.

The text’s contents stand out. In the editor’s words, ‘[t] his text is unique in being (…) a sanction against perjury which accompanied an oath sworn “at the spring of Sulis” (the only time this phrase occurs). It is the first epigraphic evidence of a belief attested in literary sources, that certain hot springs and seething pools punished perjury’ (Tomlin 1988a: 226). The phrase prid(i)e idus Apriles (l. 6) deserves special attention, since it refers to the moment when the oath was taken; also note ad fontem deae Suli(s) (l. 5), an explicit mention of the spring from which Aquae Sulis takes its name. Finally, ṣạṇguine suo illud satisfacere ̣ (l. 8) alludes to the victim’s punishment, which requires a liar to pay with his or her own blood (for parallels, see 271, 337, 346, etc). Uricalus (l. 1) is a Celtic name derived from the root uri-, while Decentinus (l. 3, see OPEL II, 94) and Alogiosa (l. 3; see OPEL I, 82 for the masculine form) are both Latin cognomina that are not commonly attested.

The text presents the following peculiarities: Suli for Sulis (l. 8) and pride for pridie (l. 6), as well as mistakes like the repetition of the final a in nominaa (l. 4; for Adams 1992: 10, this is better understood as a hypercorrection for aeorum for eorum) and the repetition of the phrase qui iuraverunt (ll. 5–6).

Translation (Tab. Sulis 94): ‘Uricalus, Docilosa his wife, Docilis his son and Docilina, Decentinus his brother, Alogiosa: the names of those who have sworn {who have sworn} at the spring of the goddess Sulis on the 12th of April. Whosoever has perjured himself there you are to make him pay for it to the goddess Sulis in his own blood.’

Image source: Tab. Sulis 94. Courtesy of R.S.O. Tomlin.

300. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 275–400 CE. [Inv. No.: 659] Material: lead, tin and copper. Measurements: 2.2 × 5.9 × 0.7 cm. Reading: Col. II. Col. I. Maria Cunsa Vẹndibedis Docimedis Sedebelia Cunsus Severia{ia}nus 5 Seni{i}la

317

Sylloge of  Defixiones from the Roman West Bibliography: Tomlin 1983: 336, 339; AE 1983, 634; Tab. Sulis 95; Kropp 2008: no. 3.2/74; Urbanová 2018: no. 193.

The circumstances that led to the writing of this curse are unknown. The text consists of nothing more than a list of personal names in the nominative case. The name Cunsa (I, l. 1) is the feminine form of Cunsus (col. II, l. 1), both of which are without known parallels, but appear to be derived from the Celtic root cun(o)- meaning ‘dog’. Another noteworthy name is Vendibelis (col. II, l. 2), which is also without parallel. It appears to be derived from the Celtic root vind(o)- ‘white’. Senila (col. II, l. 5), in contrast, is a well attested name in Britannia and is a variant of Celtic Senilis. Finally, both the Celtic name Docimedis (col. I, l. 2, attested in 210) and the Latin cognomen Severianus (col. II, l. 4) are already attested.

Image source: Tab. Sulis 95. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on an oval sheet, which was formed by pouring an alloy into a stone mould (for parallels, see 301 and 122). The inscription, organized in two columns, contains five lines, which run from left to right and were written in new Roman cursive by the same hand as 301 (these are the only two tablets from the Bath collection that were undoubtedly written by the same hand).

Translation (Tab. Sulis 95): ‘Cunsa, Docimedis, Sedebelia, Maria, Vendibedis, Cunsus, Severianus, Senila.’

301. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 275–400 CE. [Inv. No.: 661] Material: lead, tin and copper. Measurements: 2.2 × 5.8 × 0.7 cm. Reading:

Col. I. Victorinus Talipiẹịṇus Minantius Victorianus



Col. II. Compe- {pe}dita Valaunecus {+}Belia

Bibliography: Tomlin 1983: 336, 341; AE 1983, 637; Tab. Sulis 96; Kropp 2008: no. 3.2/75; Urbanová 2018: no. 194.

The circumstances that led to the writing of this curse are unknown since the text consists of merely a list of personal names in the nominative. Victorinus and Victorianus (col. I, ll. 1 and 4 respectively) are both well-attested Latin cognomina. There are no parallels for Talipieinus (col. I, l. 2), Minantius (col. I, l. 3, probably formed from the participles minans) or Compedita (col. II, ll. 1–2, but cf. Compes and Compedia in Kajanto 1982²: 345). Valaunecus (col. II, l. 3) seems to be a theophoric Celtic name derived from Vellaunus (a deity identified with Mars and Mercury), while Belia (col. II, l. 4) could be a name by itself or part of it (compare with Sedebelia in 300).

Image source: Tab. Sulis 96. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on an oval sheet and organized in two columns. It contains four lines of text, which run from left to right and were written in new Roman cursive by the same hand as 300. This is unique, since this appears to be the only pair of curses from Bath written by the same person.

Translation (Tab. Sulis 96): ‘Victorinus, Talipieinus, Minantius, Victorianus, Compedita, Valaunecus, Belia.’

302. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 275–400 CE. [Inv. No.: 669] Material: lead, tin and copper. Measurements: 4.9 × 12.8 cm. Reading: A Basilia donat in templum Martis ani lum argenteum si ser(v)us si liber mdius fuerit vel aliquid de hoc noverit ut sanguine ẹt liminibus ẹṭ (left margin, descending) Primurudeum 318

 Britannia B omnibus membris configatur vel et iam intestinis excomesis (om)nibus habe(t) is qui anilum involavit vel qui medius fuerit

Bibliography: Tomlin 1983: 336, 338; AE 1983, 633; Tab. Sulis 97; Tomlin 1991: 308; Reynolds 1990: 381; Adams 1992: 8–10; Tomlin 2002: 169; Kropp 2008: no. 3.2/76; Tomlin 2008b: 340; Tomlin 2017c: 14.01; Urbanová 2018: no. 275.

The text presents the following orthographic features: vocalic confusion in liminibus for luminibus (A, l. 4); the omission of syllables as in (om)nibus (B, l. 2). Also note serus for servus (A, l. 2), anilum for anulum (A, ll. 2–3), sanuine for sanguine (A, l. 4), habe for habet (B, l. 2, on this, see Adams 1992: 9–10) and involavit for involaverit (B, l. 3). It is worth noting the phrase medius fuerit (A, l. 3 and B, ll. 3–4), which refers to ‘someone who (...) was in the middle of things, in reference to a theft, would be someone who was there, either unwittingly or by choice, when the crime was committed’ (Adams 1992: 8–9; cf. also Reynolds 1990). Finally, we must point out two lexemes that are rather recherché: configatur (B, l. 1), which can be taken as a synonym for defigatur, and excomesis (B, l. 2), which Tomlin has described as a ‘rare intensive compound of comedo [that] occurs only in late Latin, notably in Gildas De excidio 85.2, animae viscera excomedens’ (1988a: 231).

Image source: Tab. Sulis 97. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a roughly rectangular sheet. It is in excellent condition, despite some marks left from the casting process, several imperfections and signs of having been hammered out before being inscribed. An opisthograph, the inscription contains nine lines of text, which run from left to right except for the final line of side A, which descends down the left edge of the tablet. The text is palaeographically noteworthy, since some letters are written in new Roman cursive and others in old Roman cursive (s, u, etc.).

The name Primurudeum is unparalleled, and ‘recalls the titles borne by the two leading members of a gladiatorial establishment, the summa rudis and secunda rudis’ (Tomlin 1988a: 231). Basilia is a personal name of Greek origin found in the Roman West beginning in the fourth century CE (see OPEL I, 273).

Based on its content, this curse should be grouped among the defixiones in fures. Here, Basilia denounces the theft of a ring, which she has now dedicated to the temple. This deserves some attention, since it is the first time that the stolen object is not given directly to the deity. Furthermore, the temple that ought to receive the stolen ring is the temple of Mars (A, l. 1). This is all the more surprising since the tablet was deposited in the sanctuary of Sulis Minerva (but see 238 where Mars is summoned, and 258 where Mercury is invoked). As the editor has put it, ‘The petitioner may be acknowledging the dominance of Sulis (…), or only reminding her that another god has also been consulted’ (1988a: 231).

Translation (Tomlin 2002: 169): ‘Basilia gives to the temple of Mars her silver ring. If anyone is involved or knows anything about it, whether slave or free, may he be accursed in his blood and eyes and every limb, or even have all his guts eaten away, if he has stolen the ring or is involved.’

303. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 275–400 CE. [Inv. No.: 622] Material: lead, tin and copper. Measurements: 6 × 10.5 cm. Reading: A (vacat) seu gen(tili)s sẹu C h(r)istianus quaecumque utrum vir ụṭrụṃ mulier utrum puer utrum puella utrum s[er]vụs utrum liber mihi Annia[n]5 o Mantuṭ˹i˺n(a)e de bursa mea s(e)x argente[o]s furaveriṭ tu d[o]ṃina dea ab ipso perexi[g] ẹ[-c.1- eo]ṣ si mihi pẹr [f]raudem aliquam in DẸP  RẸG̣[-c.1-]STVM dederit nec sic ipsi dona sed ut sangu inem suum (r)ẹpuṭes qui mihi hoc inrogaverit 319

Sylloge of  Defixiones from the Roman West B Postum[inu]s Pisso Locinna [A]ḷauna Maṭẹrna Gunsula C˹an˺didina Euticius 5 Peregrinus Latinus Senicianus Avitianus Victor 10 Scọ[ti]us Aessicunia Paltucca Calliopis Celerianụs

Bibliography: Tomlin 1982: 404–06; AE 1982, 667; Tab. Sulis 98; Reynolds 1990: 382; Adams 1992: 10–12, 17– 20; Gager 1992: no. 96; Kropp 2008: no. 3.2/77; Tomlin 2017c: 12.97; Urbanová 2018: no. 276.

success of this magical endeavour, it appears that Annianus introduces an apotropaic clause to counter any attempts to stave off the curse: in fact, the phrase pẹr [f]raudem aliquam (A, l. 7) seeks to ‘to make any counter-spell by the thief rebound upon him’ (Tomlin 1988a: 234). Unfortunately, the text has been damaged and we cannot know this for sure.

Image source: Tab. Sulis 98. Courtesy of R.S.O. Tomlin. See also App. IV.3, SD 303.

The lexeme bursa (A, l. 5) is used here with the sense ‘bag’ or ‘coin purse’, which it continued to have in Medieval Latin. Further note the use of the verbs perexigo and exigo: the first is a compound with an intensifying per- (see Adams 1992: 18), while the latter is used as a synonym for the more common involo.

Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a rectangular sheet, which is in good condition despite some signs of corrosion. My autopsy (December 2009) has confirmed that certain areas of the inscribed surface have disappeared, which has led to the loss of several letters (underlined in the reading: see A, ll. 2–3). An opisthograph, the inscription contains 23 lines of text, which run from right to left and from bottom to top (for this uncommon layout, see 131, 134–36, 149, 461 and 524). This complex organization of the text is likely to have caused several errors and omissions, such as genis for gen(til)is (A, l. 1), chistianus for christianus (A, ll. 1–2), sx for sex (A, l. 5), eputes for reputes (A, l. 9) and Cnadidina for Candidina (B, l. 4).

The reference to Christianity in the all-inclusive formula is undoubtedly a notable feature of the text. This is in fact the first known instance in which we find seu gen(tili)s sẹu Ch(r)istianus (A, ll. 1–2) paired together as opposites. Gentilis is the word commonly used in Late Antiquity that Christians used to refer to pagans (on this, see Adams 1992: 11). Given the popularity of this term among Christians, we can assume that the author of the text belonged to that group. Therefore, it is all the more noteworthy that, despite being Christian, Annianus still had such great confidence in the powers of Sulis (for another reference to Christianity in British curses, see the phrase ịn doṃo dei, in 446).

Based on its content, this curse should be grouped among the defixiones against thieves. Here, Annianus denounces the theft of six silver coins to Sulis. After attempting to identify the culprit with an all-inclusive formula (A, ll. 1–4; see the commentary of Adams 1992: 19–20), the defigens demands justice from the goddess, calling upon her with the formula tu, d[o]ṃina dea (A, l. 6). Sulis is asked to convert ((r)ẹpuṭes, A, l. 9) the value of the stolen coins into the victim’s blood, with which the committed offence can be repaid (for a parallel, see 246). With the hope of guaranteeing the

This tablet contains a total of 19 personal names. What is odd about the names is that only the defigens’ name seems to be followed by a matronymic Mantut˹i˺n(a)e (A, ll. 4–5); the name is a reconstruction from the transcription Mantutene. Mantutinae is unparalleled but note Mantutius [CIL VIII, 4710]. Uses of the matronymic, however, are found in a few British curses (235, 360 and 454), where 320

 Britannia it helps identify a curse’s victim. On side B, we find a list of 18 names (Latin, Greek and Celtic) in the nominative (the longest list found on a curse from Bath). The Latin names are all well attested: Postum[inu]s (l. 1, see OPEL III, 155), Materna (l. 3; see OPEL III, 64), Candidina (l. 4, attested here for the only time), Peregrinus (l. 5), Latinus (l. 6), Senicianus (l. 7; see OPEL IV, 66), Avitianus (l. 8, see OPEL I, 230–31), Victor (l. 9), Celerianus (l. 14, see OPEL II, 47). There are two names of Greek origin: Euticius (l. 4; see Solin 2003: 877, 1324, 1476) and Calliopis (l. 13; attested also in Trier, see RICG I, 189). Of the Celtic names, Pisso (l. 1), Locinna (l. 2), Gunsula (l. 3, perhaps a diminutive of Cunsa), Aessicunia (l. 11, feminine name from the Celtic theonym Esus) and Paltucca (l. 12, probably derived from the Celtic cognomen Tucca) are attested here for the first time. Finally, the names Alauna (l.

2, found in Gallia and Britannia) as well as Sco[ti]us (l. 10, sometimes written Scottius) are previously documented Celtic names. Translation (Tab. Sulis 98): ‘Whether pagan or Christian, whether man or woman, whether boy or girl, whether slave or free, whoever has stolen from me, Annianus (son of) Mantutina (?), six silver coins from my purse, you, lady Goddess, are to exact (them) from him. If through some deceit he has given me… and do not give thus to him but reckon as (?) the blood of him who has invoked this upon me. Postumianus, Pisso, Locinna, Alauna, Materna, Gunsula, Candidina, Euticius, Peregrinus, Latinus, Senicianus, Avitianus, Victor, Scotius, Aessicunia, Paltucca, Calliopis, Celerianus.’

304. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 275–400 CE. [Inv. No.: 667] Material: lead and tin. Measurements: 3.9 × 7.5 cm. Reading:

5



ẹxẹcro qui involaver`it´ qui Deomiorix de hos{i}pitio suo perdiderit quicumque rẹṣ deus illụm inveniat sanguine et vitae suae illud redemạt

Bibliography: Tomlin 1983: 336, 339; AE 1983, 635; Tab. Sulis 99; Reynolds 1990: 381; Adams 1992: 3–5; Kropp 2008: no. 3.2/78; Urbanová 2018: no. 277.

As far as diction is concerned, note the use of perdo (l. 3) in the sense of ‘lose because of theft’ (see Adams 1992: 3 and for some parallels, see 210–11, 213 and 267) and the use of deus to refer to Sulis (l. 5). Tomlin has interpreted hospitio as meaning ‘“lodgings”, the sense of “house” or “home” being virtually confined to Christian authors of the fourth century’ (Tomlin 1988a: 235). While Reynolds (1990: 381) has rejected this interpretation, Adams has accepted it and has taken the meaning ‘house’ as a regionalism common in Britannia rather than a vulgarism, which seems right given the attested Romance reflexes. The phrase deus illụm inveniat (ll. 4–5) has a clear parallel in 249, B, ll. 5–7: qui rem ipsam iṇvolavị[t] deuṣ [i]nvenia[t]. To conclude the curse, the defigens demands that the thief compensate him with his blood and life: sanguine et vita{e} sua{e} redemạt (see 447 as a parallel). Deomiorix (l. 2) is Celtic name, which is likely derived from the root de:uo and the suffix -rix.

Image source: Tab. Sulis 99. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a roughly rectangular sheet made of lead and tin. The inscription contains six lines, which run from left to right and were written in new Roman cursive (with the exception of the second n in l. 5) by a trained scribe. These lines run from left to right, except for l. 2, which is written from right to left. Perhaps this is a failed attempt at boustrophedon. Based on its content, this curse should be grouped among the defixiones in fures. Here, Deomiorix denounces a theft to Sulis Minerva, though he does not specify what object was taken from him. The text presents the following orthographic and grammatical features: execro for execror (l. 1), qui for quid (= quod, l. 2; on this see Adams 1992: 3), hosipitio for hospitio (ll. 2–3, cf. Adams 1992: 5), confused case use (cf. the genitive vitae suae, for the ablative vita sua, l. 5) and redemat for redimat (l. 5).

Translation (Tab. Sulis 99): ‘I curse (him) who has stolen, who has robbed Deomiorix from his house. Whoever (stole his) property, the god is to find him. Let him buy it back with (his) blood and his own life.’

321

Sylloge of  Defixiones from the Roman West 305. Aquae Sulis, Bath Provenance: sanctuary. Current Location: unknown. Date: 275–400 CE. Material: lead alloy (?). Measurements: 4.5 × 7.5 cm. Reading: A si puer si puella si vir si femina qui học invọḷạvit non ẹị remit tạtur ṇisi innọc{s}ent5 tiam ALE +++++ ++ (traces)

B nọn illi dimiṭṭa[t]ụr nec somnum nisi ut Ẹuṭic ̣ia mọdium ne bulae modium ṿeni5 ạṭ fumi

Bibliography: DT 105; RIB 2349*; Tab. Sulis 100; Tomlin 1994a; Adams 1992: 18; Kropp 2008: no. 3.2/2; Urbanová 2018: no. 243.

Concerning the phrases non ei remittạtur (A, ll. 3–4) and nọn illi dimiṭṭa[t]ụr (B, l. 1), Tomlin has noted, dimittere and remittere ‘are used in Christian texts in the nonClassical sense of “forgive”’ (1994a: 106). Accordingly, nisi innọcentiam (A, ll. 4–5) makes better sense and could be a variation of a sort of magical time frame, in which the curse should continue to take effect until the thief has proven his/her innocence (or blamelessness, as OLD 2 suggests). The mention of nec somnum (B, l. 2) likely formed part of a non permittas formula in which the culprit’s punishment was laid out (for parallels, see 215 and 240).

Image source: Tomlin 1994: 105, fig. 1. Courtesy of R.S.O. Tomlin. Commentary: this defixio, found alongside 206, was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 206). Nicholson attempted to decipher the text without success. According to Tomlin, ‘[h]e tried to read the text upside down’ (1994a: 102) and subsequently assumed that the text had been written by an illiterate person. Currently, the whereabouts of the tablet are unknown, and we must rely on the few photographs taken by its first editor (Bodleian Library, Oxford MS Eng. misc. d. 71).

Finally, the final formula nisi ut Ẹuṭicịa mọdium nebulae modium ṿeniạṭ fumi (B, ll. 2–5) is unparalleled. Tomlin takes Euticia as the subject of the phrase and has identified her as the curse’s target. Modium nebulae-modium fumi appears to pick up on a proverb, which suggests that veniat should be taken as a form of the verb veneo. Therefore, it appears that Euticia is asked to undertake an unparalleled and impossible punishment: ‘to sell cloud and smoke in measured quantities’ (Tomlin 1994a: 107), which offers the false hope of salvation and accordingly reinforces the curse’s power (for this phrase and magical time frames, see Sánchez Natalías 2019a).

The defixio was written on a roughly rectangular sheet and has been broken along its bottom edge. An opisthograph, the inscription contains 11 lines of text, which run from left to right and were written in a mix of old and new Roman cursive by a trained hand. Thanks to the presence of several formulae, this curse should be grouped among the defixiones against thieves. We find the typical allinclusive formula (A, ll. 1–2) which in this case opens the text (cf. 303). Also note the allusive mention of the purloined object in the phrase qui học invọḷạvit (A, ll. 2–3). For parallels where the object is not named explicitly, see 304 and 358 (in contrast with 267 in which all the stolen cloaks are carefully described).

Translation (Tomlin 1994a: 104): ‘Whether (they be) boy or girl, whether man or woman, forgiveness is not to be given to the person who has stolen this unless (…) innocence. Forgiveness is not to be given to him/her, nor shall he/she sleep, except on condition that Euticia (?) sell a bushel of cloud, a bushel of smoke.’ 322

 Britannia 306. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 275–400 CE. [Inv. No.: 686] Material: lead and tin. Measurements: 5.5 × 6.4 cm. Reading:

si VAPERSVRVSMIIMIII ILE si CERIASIVS si IGEVNS NSER permaneat (vacat)

Bibliography: Tab. Sulis 101; Kropp 2008: no. 3.2/80.

The circumstances that led to the writing of this curse are unknown. The editor has proposed reading the personal name Ceriasus (l. 2; perhaps for Cerialis? See OPEL II, 51). IGEVNS (l. 3) could be taken as an attempt at the adjective ingenuus or the Latin cognomen Ingenuus (see OPEL II, 194).

Image source: Tab. Sulis 101. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a rectangular lead and tin tablet, which had been hammered flat before being inscribed. The whole curse tablet is intact, and the text is complete, but also very corrupt. This precludes us from establishing a satisfactory reading. The inscription contains four lines, which run from left to right and were written in new Roman cursive.

After being inscribed, the tablet was folded once. Translation (Tab. Sulis 101): ‘Whether… let him remain.’

307. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 275–400 CE. [Inv. No.: 597] Material: lead, tin and copper. Measurements: 5.3 × 3.8 cm. Reading:

5



ṇumen furti ṣi sẹ[r](v)ụs si ḷ[ibe]r si puer sị pụẹllạ [---] ++++++[---] +++[---]

Bibliography: Tab. Sulis 102; Kropp 2008: no. 3.2/81.

Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a sheet that was originally

Image source: Tab. Sulis 102. Courtesy of R.S.O. Tomlin. 323

Sylloge of  Defixiones from the Roman West name (ll. 1–2; cf. 220 and especially 221) and the allinclusive formula (ll. 2–5). After l. 6, the tablet is too corroded to be read.

rectangular. It has been broken into three pieces that fit together and constitute one third of the original tablet. It preserves the remains of seven lines of text, which run from left to right and were written in new Roman cursive.

After being inscribed, the tablet was folded.

Despite being so fragmentary, this curse should be grouped among the defixiones in fures, since it contains some typical formulae, including reference to the thief’s

Translation (Tab. Sulis 102): ‘The name (?) of the thief (?), whether slave or free, whether boy or girl…’

308. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 275–400 CE. [Inv. No.: 668] Material: lead, tin and copper. Measurements: 2.7 × 6.6 cm. Reading:

MODSVSIO+IVIVCI deus faci(a)t ani(m)am pe(r)d(e)re sui

Bibliography: Tab. Sulis 103; Adams 1992: 13; Kropp 2008: no. 3.2/82; Urbanová 2018: no. 279.

life at the hands of the deity invoked (ll. 2–3). The text’s layout, which is rather uncommon, appears to have given way to several errors: aniam for animam (l. 2), facit for faciat (l. 2) and pedre for perdere (l. 3). Tomlin has taken the final sui as a vulgarization in which the possessive and demonstrative pronoun have been confused; Adams, however, has noted that ‘The writer of the tablet would seem to have substituted a genitive sui for the possessive adjective suam (referring of course to the subject of the dependent clause). This use of the reflexive genitive is not a vulgarism, but an occasional variant for suus even in literary Latin’ (1992: 13).

Image source: Tab. Sulis 103. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on an irregularly shaped sheet of lead, tin and copper. The inscription contains three lines, which were written in new Roman cursive and run from both left to right (l. 1) and right to left (ll. 2–3). The circumstances that led to the writing of this curse are unknown. The text does contain a phrase related to the victim’s punishment, which states that he ought to lose his

Translation (Tab. Sulis 103): ‘… may the god make (him) lose his life.’

309. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 275–400 CE. [Inv. No.: 123] Material: lead, tin and copper. Measurements: (7.4) × (8.2) cm. Reading: (traces) Lucius (traces) VS (traces) A VASOMO[-c.1-]+ 5 NICINIFII+ [---]VNOM++A S++NVS MAS

324

 Britannia Bibliography: Tab. Sulis 104; Reynolds 1990: 38.

The circumstances that led to the writing of this curse are unknown. Here we have three different texts, two of which are written over one another (ll. 1–3), while the third follows (ll. 4–8). Given both the wide right margin and the endings in –us and ­–a, it seems that the text is a list of names (in such lists, each name often receives its own line; e.g., 208 and 256). For her part, Reynolds (1990: 38) has read Lucius or Iustus (l. 1).

Image source: Tab. Sulis 104. Courtesy of R.S.O. Tomlin. Commentary: this small quadrangular fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207) and conserves eight lines, which run from left to right and were written by at least two different authors who used a mixture of capitals (e.g., the final s in ll. 1–2), old Roman cursive (e.g., the final a in l. 3, va at the beginning of l. 4 and a in ll. 6 and 8) and new Roman cursive (e.g., m, n and s in ll. 4–7).

After being inscribed, the tablet was cut along the left edge and then folded.

310. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 275–400 CE. [Inv. No.: 533] Material: lead and tin. Measurements: (5.2) × (5.3) cm. Reading: A [---]C (traces) [---]DOSCIM (traces) [--- (?) invola]vẹrit si ser [vus si liber] +++ 5 (traces) (traces) (traces)

B [---]+++RI (traces) [---]+++C (traces) [--- (?) fraude]m fecerit (traces) [---]++CNI (traces) 5 [---]DASD++NBR[-c.1-]S+P++D [---] (traces) INTELLEG (traces) (traces)

Bibliography: Tab. Sulis 105; Kropp 2008: no. 3.2/83.

the reconstruction [invola]verit (A, l. 3) and an allinclusive formula (A, ll. 3–4), we would most likely have a defixio against thieves. In this case, it is also worth pointing to the phrase [fraude]m fecerit (B, l. 3, for parallels, see 237, 240, 357, etc.). Doscim- (A, l. 2) could be part of a personal name comparable to Docimedis. INTELLEG (B, l. 6), which appears to come from the verb intellego, is of great note, since it could, in Tomlin’s words, be ‘perhaps a reference to the god’s understanding’ (1988a: 242).

Image source: Tab. Sulis 105. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), has been broken into two pieces that fit together to form a roughly rectangular alloyed sheet of lead and tin. The tablet is broken along its left edge and its surface is wrinkled. An opisthograph, the inscription contains 17 lines of text, which run from left to right and were written in new Roman cursive.

After being inscribed, the tablet was folded four times. Translation (Tab. Sulis 105): ‘…whether slave or free… has done… understand/stood….’

The situation that led to the writing of this curse cannot be fully reconstructed. That said, if we accept 325

Sylloge of  Defixiones from the Roman West 311. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 275–400 CE. [Inv. No.: 620] Material: lead and tin. Measurements: 5.5 × (5.5) cm. Reading:

[---]+T+V+E++[---] [---]S+++N[---] [---]+[---]

Bibliography: Tab. Sulis 106. Courtesy of R.S.O. Tomlin.

well as some of the central portion of the curse. It preserves three lines of text, which run from left to right and were written in new Roman cursive, though the writing is so sloppy that only a few letters can be made out.

Commentary: this defixio, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), was written on a roughly triangular sheet of lead and tin. One of the triangle’s angles has been lost as

After being inscribed, the tablet was folded.

312. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 275–400 CE. [Inv. No.: 619] Material: lead alloy. Measurements: (2.3) × 4.1 cm. Reading: [--- (?)eu]m qui mi[hi] (traces) Bibliography: Tab. Sulis 107; Kropp 2008: no. 3.2/84.

edge of a larger tablet and preserves two lines of text, which run from left to right and were written in new Roman cursive. The editor has proposed the reconstruction [eu]m qui mi[hi]… fraudem fecerit / involaverit (or something similar).

Commentary: this rectangular fragment was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207). It constitutes the upper

313–16. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: 275–400 CE. [Inv. No.: vid. commentary] Material: lead alloy. Measurements: vid. commentary. Reading:

Vid. commentary.

Bibliography: vid. commentary.

314 (=Tab. Sulis 109; Kropp 2008: no. 3.2/86) [Inv.: 399]: two fragments still folded and probably belong to the same tablet ((2.2) × (1.8) cm). They preserve the remains of three lines written in new Roman cursive, which run from left to right and read: [---]VS+[---]/ [---]DVO[---]/ [---]+M+[---].

Commentary: this collection of fragments was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207). They conserve the remains of inscriptions written in new Roman cursive, which have only been briefly described by Tomlin. Given the brief nature of the published account, I provide here a joint entry on these several items in order to give each curse a number in the present catalogue as well as provide bibliographic references (in parentheses) and inventory numbers (in square brackets) and the tablet’s dimensions. A brief commentary is also provided.

315 (=Tab. Sulis 110) [Inv.: 20, 001]: sheet broken into two fragments that fit together, whose surface is irregular and corroded (dimensions are 3 × 4 cm). It is still folded and preserves the remains of four lines, which run from left to right and read: +[---]/ qui[---]/ isti+[---]/ ++NI(or M)++[---].

313 (=Tab. Sulis 108; Kropp 2008: no. 3.2/85) [Inv.: 399]: four fragments still folded. One of them, measuring (2.7) × (3.1) cm, preserves the remains of three lines, which run from left to right and read: Doc ̣im[edis?---]/ dẹ(ae) Su[li---]/ ++N+[---].

316 (=Tab. Sulis 111) [Inv.: 595]: Opisthographic defixio, whose text cannot be deciphered and dimensions are 6 × 9 cm. After being inscribed, the tablet was folded three times. 326

 Britannia 317. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: 666] Material: lead and tin. Measurements: 4.7 × 13.3 cm. Reading:

Vid. commentary.

Bibliography: Tab. Sulis 112.

utterance of the curse and the folding of the tablet (as usual in Bath; on this, see section I.4.1).

Image source: Tab. Sulis 112. Courtesy of R.S.O. Tomlin.

Of the group of pseudo-inscriptions from Bath, the current tablet most closely resembles the work of a literate author. That said, it nevertheless appears to be little more than garbled nonsense. An opisthograph, the inscription contains at least six lines of ‘text’, in which we can make out several letters like m, r and h. On side A, the text takes up as much room as possible.

Commentary: this defixio was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207). The editor has classified it with 318–21 as a group of pseudo-inscriptions, that is, inscriptions that were written by illiterate individuals (for a parallel, see 440). Being illiterate, though, did not prevent them from knowing all the ritual involved in the use a curse tablet: the media employed, the ‘writing’ itself, the probable

After being inscribed, the tablet was folded six times.

318. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: 122] Material: lead and tin. Measurements: 6.8 × 5.9 cm. Reading:

Vid. commentary.

(for the archaeological context, see 207), was written on a roughly triangular lead sheet. An opisthograph, the inscription consists of signs that look like a 7, the ligature ti (side A) and overlapping upside down vs (side B).

Bibliography: Tab. Sulis 113. Image source: Tab. Sulis 113. Courtesy of R.S.O. Tomlin. Commentary: this pseudo-inscription (cf. 317), which was discovered in the British sanctuary of Sulis Minerva 327

Sylloge of  Defixiones from the Roman West 319. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: 674] Material: lead and tin. Measurements: 4.7 × 6 cm. Reading:

Vid. commentary.

(for the archaeological context, see 207), was written on a triangular sheet. Some signs do approximate letter forms, but others seem to be totally random.

Bibliography: Tab. Sulis 114. Image source: Tab. Sulis 114. Courtesy of R.S.O. Tomlin. Commentary: this pseudo-inscription (cf. 317), which was discovered in the British sanctuary of Sulis Minerva

320. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: 694] Material: lead and copper. Measurements: 2.5 × 6 cm. Reading:

Vid. commentary.

Image source: Tab. Sulis 115. Courtesy of R.S.O. Tomlin.

(for the archaeological context, see 207), was written on a roughly semicircular sheet. There are four lines of ‘text’ made of a series of short vertical strokes.

Commentary: this pseudo-inscription (cf. 317), which was discovered in the British sanctuary of Sulis Minerva

After being inscribed, the tablet was pierced with a nail and then folded twice.

Bibliography: Tab. Sulis 115.

321. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: 532] Material: lead, tin and copper. Measurements: 2.7 × (5.2) cm. Reading:

Vid. commentary.

328

 Britannia the final pseudo-inscription from Bath (cf. 317). The inscription is made up of crosses.

Bibliography: Tab. Sulis 116. Commentary: this fragment, which was discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), belonged to the upper-left corner of

After being inscribed, the tablet was folded.

322–27. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: vid. commentary] Material: lead alloy. Measurements: unknown. Reading:

See commentary.

Bibliography: Tab. Sulis 117–22.

tablets cut from metal sheet, but are not inscribed (…) these blanks suggest that tablets were manufacturated for use by others’ (1988a: 253). Three of them (323, 325 and 327) were folded and perhaps were inscribed. Their inventory numbers are 682, 20,006, 617, 663, 662 and 678, respectively.

Commentary: according to Tomlin, these six items, which were discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), ‘resemble curse

328–35. Aquae Sulis, Bath Provenance: sanctuary. Current Location: Roman Bath Museum. Date: unknown. [Inv. No.: vid. commentary] Material: lead alloy. Measurements: unknown. Reading:

Vid. commentary.

Bibliography: Tab. Sulis 123–30.

(334). Another was pierced twice (335). It appears that most of these curses were not inscribed, except for 328 (with traces of capital letters) and 329. Their inventory numbers are 481, 603, 546, 486, 693, 692, 685 and RBS 7913, respectively.

Commentary: these tablets, which were discovered in the British sanctuary of Sulis Minerva (for the archaeological context, see 207), have yet to be opened up. Most were folded, though one item from the group was rolled up

336. Chesterton-on-Fosse Provenance: Roman fort. Current Location: Warwick (County) Museum. Date: 175–275 CE. Material: lead. Measurements: (3.17) × (5.55) cm. Reading:



(in old Roman cursive) [---] (traces) [---] [---]IS+I (traces) [---] [---] (vacat) +IOS (traces) [---] [---D]iogenịs (vacat) dalmatic[um?] (traces; vacat) [---] (traces) [---] [---] (traces) OS (traces) [---] [---] (traces) IVS (traces) [---] [---] (traces) [---]



(in capitals, superimposed over the previous text) [---]SEITHAVS (vacat) [---] dalmaticum [---] [---] (traces) [---]

5

329

Sylloge of  Defixiones from the Roman West Bibliography: Taylor and Collingwood 1921–22: 239; AE 1924, 98; RIB I, 243; Tab. Sulis p. 61, no. 23; Tomlin 1989: 344–45; RIB II, 2504.22; Kropp 2008: no. 3.8/1.

the writing of a text in capitals over another in cursive was ‘perhaps for emphasis, or to renew a worn text, or as a summary of the contents’ (Tomlin 1989: 345). Part of the text in cursive has been reconstructed as [?D]iogenis dalmatic[um?] (l. 4), given that the last word is also found written in capitals in the second inscription. The term SEITHAUS (l. 1) could be a misspelt personal name. Based on repeated mention of a physical item (dalmaticum), it seems quite probable that this curse is a defixio against thieves.

Image source: Tomlin, apud RIB II, 2504.22. Courtesy of R.S.O. Tomlin. Commentary: this rectangular lead fragment was discovered during the excavations of the Roman fort in Chesterton-on-Fosse (Warwickshire) in 1921. The tablet contains two different inscriptions (one superimposed over the other), which run from left to right. The first contains eight lines and was written in old Roman cursive, whereas the second has three lines and was written in capitals. Since both texts seem to be related,

After being inscribed, the tablet was folded in half. Translation (Tomlin 1989: 345): (cursive text) ‘…tunic of Diogenes…’ (capital text) ‘…tunic…’

337. Isca Silurum, Caerleon Provenance: amphitheatre. Current Location: NRLM. Date: 1st–2nd centuries CE. [Inv. No.: 35.119/6.2] Material: lead. Measurements: 7 × 7.5 cm. Reading:

5



dom(i)na Nemesis do tibi palleum et galliculas qui tulit non redimat ni vita sanguine suo

Bibliography: Collingwood 1928: 157; AE 1929, 46; Oxé 1931: 16–19; Preisendanz 1933: 155; Egger 1943: 108–09; RIB I, 323; García Ruiz 1967: no. 21; Solin 1968: no. 20; Mann 1971: 220 and 223; Tab. Sulis p. 61, no. 29; Hornum 1993: no. 28; Gager 1992: no. 100; Tomlin 1999a: 555; Vismara 2000: 153–54, 260; Tremel 2004: no. 91; Kropp 2008: no. 3.6/1; Tomlin 2010: 253; Versnel 2010: 286–89; Tomlin 2017c: 12.57; Urbanová 2018: no. 283.

damage (Tomlin 1999a). The inscription contains eight lines, which run from left to right and were written in old Roman cursive. Based on its content, this curse should be grouped among the defixiones in fures. Here, the theft of a horse blanket and pair of boots is denounced to Nemesis. The stolen items are dedicated to the deity so that the theft becomes a sacrilegious act. The text presents the following orthographic and phonetic features: palleum for pallium (l. 3) and domna for domina (l. 1). After the initial invocation (ll. 1–2) and dedication of the purloined property (ll. 2–4), we find the phrase qui tulit non redimat ni vita sanguine suo (ll. 5–8). This phrase was interpreted literally by Oxé, while subsequent authors like Egger, Hornum and the editors of RIB have taken it as a metaphor that alludes to the colour of the stolen blanket. Nevertheless, we should pause here: comparison with other British defixiones clearly shows that this is common formulaic language that alludes to the need to

Image source: Collingwood 1928: 157, fig 12, no. 10. Commentary: this defixio was discovered in 1927 in the northern part of the arena of the amphitheatre in Caerleon nearby the camp of the Legio Secunda Augusta. The curse was written on a small lead tabula ansata, a common shape in ancient writing practices that was here adapted for magical purposes (see 463 and 479 for parallels). Today, it has lost the left ansa and is generally in a bad state of conservation. This is due to the fact that the tablet was stored in an oak drawer, which caused irreversible

330

 Britannia Translation (Gager 1992: no. 100, modified): ‘Lady Nemesis! I give you this cloak and these shoes. May the person who has taken them not buy back (them) except with his own life and blood.’

pay for the crime with the victim’s blood and life (e.g., 304). Finally, the invocation of Nemesis is noteworthy, compare with 343. After being inscribed, the tablet was pierced twice on the top and bottom edge.

338. Londinium, London Provenance: unknown. Current Location: MOLA. Date: 1st century CE. [Inv. No.: 29.94/6] Material: lead. Measurements: 8 × 12.3 × 0.3 cm. Reading: A T(itus) Egnatius Tyran(n)us deficus est et P(ublius) Cicereius Felix 5 defictus et

B T(itus) Egnatius Tyran(n)us defictus est et P(ublius) Cicereius Felix

Bibliography: Collingwood and Taylor 1928: 213–14; AE 1930, 112; RIB I, 6; García Ruiz 1967: no. 52; Solin 1968: no. 16; Burn 19692: 45; Tab. Sulis p. 60, no. 1; Kropp 2008: no. 3.14/2; Urbanová 2018: no. 199.

The circumstances that led to the writing of this curse are unknown. The text takes aim at two victims, who are identified with their tria nomina (such an onomastic formula is without parallel among all the known British curses). Furthermore, the use of the verb defigo is notable (A, ll. 2 and 5: defictus for defixus; this verb is rare in British curses: for some parallels, see 339 and 345).

Image source: RIB I, 6. Commentary: this defixio was discovered in 1928 while pouring a foundation on Princess Street (London) at the site of the National Provincial Bank. The curse, written on an irregularly shaped lead sheet, is intact and in excellent condition, as I was able to verify during an autopsy (December 2009). An opisthograph, the inscription contains nine lines, which run from left to right and were written in capitals by two different authors. This is evident from the differences found on each side of the sheet, such as the depth of the inscription as well as the size and strokes of certain letters (especially r, though also see t and d).

After being inscribed, the tablet was twice pierced with a nail (once from side A to side B and then the other way as well). The larger of the preserved holes (from side A to B) coincides with the names of both victims. After being pierced, the tablet was cast into a fire, which is now evident from the melting of the edges (for this uncommon practice, see e.g., 493). Translation (RIB I, 6): ‘Titus Egnatius Tyranus is cursed and Publius Cicereius Felix is cursed.’

331

Sylloge of  Defixiones from the Roman West 339. Londinium, London Provenance: unknown. Current Location: British Museum. Date: 2nd century CE. [Inv. No.: 1934, 1105.1] Material: lead. Measurements: 11.5 × 17.2 × 0.1 cm. Reading:

5



Tretia(m) Martia(m) defico et illeus vita(m) et me(n)tem et memoriam ẹt iocinera pulmones interm{x}ix{i}ta fata cogitata memoriam s˹ic˺ no(n) possitt loqui (quae) sicreta si(n)t neque SINITA MERE possit neque[---] [---]CLVDO

Bibliography: Collingwood 1935: 224–26; AE 1936, 87; RIB I, 7; García Ruiz 1967: no. 4; Solin 1968: no. 17; Burn 19692: 45; Mann 1971: 220 and 222; Tab. Sulis p. 60, no. 2; Kropp 2008: no. 3.14/1; Urbanová 2018: no. 198.

The circumstances that led to the writing of this curse are unknown, although the request to render the victim silent could point towards a juridical text. That said, the fragmentary nature of the final portion of the inscription precludes us from drawing any firm conclusion. The text presents the following orthographic and phonetic features: omission of final -m in Tretia (l. 1; possibly for Tertiam, see OPEL IV, 114–15), Martia (l. 1) and vita (l. 2); also note metem for mentem (l. 2); illeus for illius (l. 2), sci for sic (l. 6) and sicreta for secreta (l. 7; see Mann 1971: 220). Note the use of the verb defigo (l. 1), which is rare in British curses (see 338 and 345 for parallels).

Image source: Collingwood 1935: 226, fig. 24. Commentary: this defixio was discovered in 1934 during construction work carried out on Telegraph Street (Moorgate, London). It was written on a reused lead sheet, which, according to Collingwood, had probably been used to cover a wooden beam. The tablet is irregularly shaped and has lost its top-left corner. Furthermore, it now has various fractures along its edges. I was nevertheless able to confirm during an autopsy (December 2009) that the curse tablet has been pretty well preserved, despite the fact that the lower-left corner has been somewhat corroded. The tablet’s surface was hammered flat before being inscribed. The inscription contains nine lines, which run from left to right and were written in old Roman cursive that can be dated on palaeographic grounds to the second century CE.

After being inscribed, the tablet was pierced nine times from side B to side A. The largest of the preserved holes happens to coincide with the victim’s name (also the case in 338) and the word vita. Translation (RIB I, 7, modified): ‘I curse Tertia Martia and her life and mind and memory and liver and lungs mixed up together, and her words, thoughts, and memory; so may she be unable to speak what things are concealed, nor be able…nor…’

340. Londinium, London Provenance: riverbank (Thames River). Current Location: MOLA. Date: unknown. [Inv. No.: 84.351] Material: lead. Measurements: 9.5 × 6.9 cm. Reading: A tibi rogo Metu nus u m vendic as de iste nu mene me ven5 dicas ante q(u)ọ d ven die(s) no vem rogo ṭe Metunus ut u mi vendịcas 10 ante q(u)ọ[d] ven di n˹o˺ve m 332

 Britannia B xuparanti Silviele Satavile xsuparatus Silvico la Avitus Melus5 so datus PERVCITIBI Santinus Mag[-c.2-3-]etụs APIDIMIS Ạntoni[us?] 10 Santus Vasia nus Varaṣius datus

Bibliography: Tomlin 1987b: 360–64; AE 1987, 738; Tab. Sulis p. 61, no. 3; Kropp 2008: no. 3.14/3; Adams 2016: no. 32; Urbanová 2018: no. 287.

nomine (A, ll. 4–5, see Adams 2016: 399–400), qod for quod (A, l. 5, see Adams 2016: 400–401), die for dies (A, l. 11), Silviele for Silviola (B, l. 2), Silvicole for Silvicola (B, ll. 3–4), Santinus for Sanctinus (B, l. 7), Santus for Sanctus (B, l. 10) and Vasianus for Vassianus (l. 10).

Image source: Tomlin 1987b: 361–62, figs. 1–2. Courtesy of R.S.O. Tomlin.

The formula used to ask Neptune to take vengeance on behalf of the defigens, ut me vendicas (A, ll. 2–3; 8–9; on this see Adams 2016: 401), should be compared to those found in 240 and 120. The divine intervention ought to take place within a determined magical deadline: ante dies novem, similar to that found in other curses (267, l. 3 and 450 B, ll. 1–2: ante dies novem; 456, l. 23: ant{a}e nonum diem; 530, l. 14: infra dies noṿẹ(m). See Reynolds 1990: 381–82 and Marco Simón 2010b for the relation to the nundinae; see Sánchez Natalías 2019a for magical deadlines in general).

Commentary: this defixio was discovered in 1984 on the northern bank of the River Thames. The exact circumstances of discovery are unknown. It was written on an irregularly shaped lead sheet, whose surface was hammered out before being inscribed. It is in good condition, as I was able to confirm during an autopsy (December 2009). An opisthograph, the inscription contains 24 lines written in capitals, though every s on side A and all the letters on side B are ‘mirrored’ (i.e., from right to left). Furthermore, the layout of the curse shows a substantial amount of variation: while the text runs from left to right on side A, things are trickier on side B.: ll. 1–5 run from right to left beginning at the top of the tablet; l. 6 runs down along the left edge, l. 7 along the bottom edge and l.8 up the right edge. Ll. 9–12 run from top to bottom with l.9 next to l. 8 and l. 12 closer to l. 6.

The words xuparanti (B, l. 1.) and xsuparatus (B. l. 3) should be interpreted as the Latin cognomen Exsuperantius (see OPEL II, 131). The Latin cognomina are Silviele (perhaps for Silviola B, l. 2; see Kajanto 19822: 168), Silvicole (ll. 3–4; perhaps for Silvicola, see OPEL IV, 82), Avitus (B, l. 4; see OPEL I, 231–33), Santinus (B, l. 7; see Kajanto 19822: 252), Santus (B, l. 10; see Kajanto 19822: 252) and Varasius (B, l. 11, first attested here; see OPEL IV, 147). If we accept the editor’s conjectures, we must add Antonius (B, l. 9) to this list, and perhaps Apicianus (for APIDIMIS (?), B, l. 9) as well. The Celtic cognomina are Satavile (B, l. 2, diminutive of the more common Satto/a), Mag[-c.2-3-]etus (B, l. 8, difficult to reconstruct) and Vasianus (B, l. 10, derived from the Celtic root *vasso-, ‘slave’). Melusso (B, ll. 4–5) is a Greek masculine name (Melissus; see Solin 2003: 1139). Finally, PERVCITIBI (B, l. 6) is of uncertain interpretation, since it could be understood as an incorrect form of the phrase pervici tibi or as a personal name related to Pervica or Pervinca. The second option would make better sense in this context.

The circumstances that led to the writing of this curse are unknown. Here, the defigens summons Metunus to ask for vengeance (A, ll. 4–7: me vendicas ante q(u)ọd ven die(s) novem). This formula, which is repeated again (A, ll. 9–12), is followed by a list of personal names belonging to the victims who are handed over (datus: B, ll. 5 and 12) to the deity. The editor has argued that with the name ‘Metunus’ the defigens was attempting to refer to Neptune. The argument is predicated on a confusion between initial nasals (m for n) and a simplification of the consonant cluster pt resulting in t. According to Tomlin, ‘British Celtic had no -pt- and the same change is found in Latin loan-words taken into Celtic’ (1987: 360; in support of this interpretation, see Adams 2016: 399). In further support of this hypothesis, we can note that Neptune is invoked in other British curses (441, 449 and 451) as well as the aquatic context in which the tablet was discovered. The text presents the following orthographic and phonetic features: vendicas for vindicas (A, ll. 2–3), numene for

Translation (Tomlin 1987b: 360): ‘I ask you, Metunus (i.e., Neptune), that you avenge me on this name (?), (that) you avenge me before nine days come. I ask you, Metunus, that you avenge me before nine days come. (List of names).’ 333

Sylloge of  Defixiones from the Roman West 341. Londinium, London Provenance: pit. Current Location: MOLA. Date: 4th century CE. [Inv. No.: USA 88 [193] ] Material: lead. Measurements: 7.6 × 7.3 × 0.05 cm. Reading:

Martia sive Martina

Bibliography: Hassall 1992: 309–10; AE 1992, 1123; Kropp 2008: no. 3.14/4; Bernardo Stempel and Hainzmann 2009: 77; Seeley and Wardle 2009: 148–49; Urbanová 2018: no. 200.

during an autopsy (December 2009), the curse tablet is in wonderful condition despite several holes found on the upper third of the tablet (these are the result of a natural process). The inscription, located along the bottom edge of the lead tablet, contains two lines, which run from right to left and were written in capitals (measuring 0.3–0.5 cm in height). The circumstances that led to the writing of this curse are unknown. The text only provides the name of the victim, but (amusingly) the conjunction sive makes clear that the writer was not sure of her exact identity, and so the defigens includes two possibilities (for other curses with just the name of the victim, see 368 and 442). Martia and Martina are both well attested Latin cognomina (see OPEL III, 60 and 61 respectively).

Image: see App. IV.3, SD 341. Commentary: this defixio was discovered in 1988 during the excavation carried out in 10–18 Union Street. Upon discovery, the curse was folded and found among the filling of a pit alongside objects from the fourth century CE. Ceramic fragments and a coin from the reign of Constantine II were among the remains (for the archaeological context, see Seeley and Wardle; for the relationship between coins and curses, see Museros Ortiz 2007).

After being inscribed, the tablet was folded.

The curse was written on an almost square lead sheet, which was cut from a larger sheet. As I was able to establish

Translation (Hassall 1992: 310): ‘Martia or Martina.’

342. Londinium, London Provenance: unknown. Current Location: MOLA. Date: second half of the 2nd century CE. [Inv. No.: ONE 94 [8298] ] Material: lead. Measurements: 4.9 × 7 × 0.1 cm. Reading:

Plautius Nobilianus A[u]rel Saturninus Domitia Attiola et ṣị qui ạfuerẹ

Bibliography: Tomlin 2003: 361; AE 2003, 1017; Kropp 2008: no. 3.14/5; Urbanová 2018: no. 201. Image source: Tomlin 2003: 361, fig. 2. Courtesy of R.S.O. Tomlin. See also App. IV.3, SD 342.

during an autopsy in December 2009). The inscription contains four lines, which run from left to right and were written in old Roman cursive dated on palaeographic grounds to the second half of the second century CE. The letters measure between 0.8 and 1.3 cm in height.

Commentary: this defixio was discovered in 1994 during an excavation at 1 Poultry Street. The curse was written on a small, roughly rectangular lead sheet, which is intact despite severe corrosion that has damaged the upper corners and the bottom edge of the tablet (I was able to confirm so much

The circumstances that led to the writing of this curse are unknown. The text targets three victims who are identified by nomen and cognomen, which is somewhat unusual since victims are normally identified solely by cognomen in British curses (for a parallel, see 442). Among the 334

 Britannia After being inscribed, the tablet was folded twice.

names, note the cognomen Attiola (see OPEL I, 211). The text concludes with an unparalleled phrase that appears to function as an all-inclusive formula meant to help the defigens curse all his/her enemies: et ṣị qui ạfuerẹ (l. 4). According to OLD (13), the absum means ‘to be missing’ (from a group), which suits the context well.

Translation (Tomlin 2003: 361, slightly modified): ‘Plautius Nobilianus, Aurelius Saturninus, Domitia Attiola, and any who was missing.’

343. Londinium, London Provenance: amphitheatre. Current Location: MOLA. Date: 2nd–3rd centuries CE. [Inv. No.: GYE ] Material: lead. Measurements: (10.5) × (8.5) × 0.1 cm. Reading: [d]eae Dea[na]e dono c ̣apitûlarem et faṣ [c]iaṃ minus partẹ tertia si quis hoc feci[t] ̣ 5 [s]ị p[u]er sị [p]ụẹlḷạ ṣ[i] [s]er[vus] ṣị [liber] don[o eum] nec p[er] me [vi]v[ere] possit

her two thirds of the stolen goods’ value (ll. 3–4; see 205, 349, etc. for parallels for this type of transaction). The text includes several of the most common phrases found throughout the corpus of British curses, such as si quis hoc feci[t] (l. 4, analogous to the more common qui hoc ̣ involaverit) and the all-inclusive formula (ll. 5–6) used to describe the culprit. The curse concludes with the phrase don[o eum] nec p[er] / me [vi]v[ere] possit (ll. 7–8) that is used to hand the curse’s victim over to the deity invoked (for parallels, see 350, 451, 457, etc.).

Bibliography: Tomlin 2003: 362–63; AE 2003, 1021; Tomlin and Hassall 2005: 496; AE 2005, 885; Kropp 2008: no. 3.14/6; Tomlin 2017c: 12.69; Urbanová 2018: no. 288; Sánchez Natalías 2022. Image source: Tomlin 2003: 362, fig. 3. Courtesy of R.S.O. Tomlin. See also App. IV.3, SD 343. Commentary: this defixio, along with 344, was discovered in the amphitheatre at Guildhall Yard in 1992, specifically among the material used to fill a drain (dated between 160–250 CE) within the arena. The inscription was written on a lead sheet that was originally rectangular but now has lost all of its edges except for part of the bottom one. It has been corroded, which resulted in a hole in the centre of the tablet. The inscription contains eight lines, which run from left to right and were written in capitals.

Finally, it must be mentioned that this defixio provides the first evidence for the worship of Diana in Londinium. That said, this invocation does fit the context of discovery well, since ‘Diana (or Deana), sometimes identified with Nemesis, was particularly associated with gladiatorial games and the amphitheatre’ (Tomlin and Hassall 2005: 496). For the only appearance of Nemesis in a British curse, see 337.

Although the text is quite fragmentary, this curse should be grouped among the defixiones in fures. Here, the theft of various articles of clothing is denounced to Deana (the High Imperial spelling of Diana). In an attempt to encourage the goddess’s intervention, the defigens offers

Translation (Tomlin 2003: 362): ‘I give to the goddess Deana (my) headgear and band less one-third. If anyone has done this, whether boy or girl, whether slave or free, I give him, and through me let him be unable to live.’ 335

Sylloge of  Defixiones from the Roman West 344. Londinium, London Provenance: amphitheatre. Current Location: MOLA. Date: unknown. [Inv. No.: GYE ] Material: lead. Measurements: 4.2 × 6.6 cm. Reading: A VIN MO MṾBOCIṂỊṾI

B ṢṾNXIIA+ AROTAI+

Bibliography: Tomlin 2003: 363; AE 2003, 1022a–b.

An opisthograph, the inscription contains four lines, which run from left to right and were written in irregular cursive letters. The reading and interpretation of the text have proven somewhat difficult. That said, the editor has claimed that the tablet could contain two indigenous onomastic formulae with corresponding patronymics.

Commentary: this defixio was discovered in 1992 during the excavations carried out at the amphitheatre at Guildhall Yard. The defixio was found in the arena, and more specifically in the filling of a drain dated between 250 and 364 CE. The inscription was written on a small, irregularly shaped lead tablet, whose bottom edge was folded and whose central part has been damaged.

After being inscribed, the tablet was pierced four times, a fact that has helped confirm that the inscription is indeed a curse tablet (for parallels, see 338, 339 and 345).

345. Clothall Provenance: necropolis. Current Location: Letchworth Museum. Date: unknown. [Inv. No.: 1930/5789] Material: lead. Measurements: 10.16 × 12.06 cm. Reading:

Vetus quomodo ṣanies signeficatur Tacita deficta

Bibliography: Collingwood and Taylor 1931: 248; Westell 1931: 291–92; AE 1937, 66; Wright 1952: 103; AE 1953, 134; RIB I, 221; García Ruiz 1967: no. 53; Solin 1968: no. 19; Mann 1971: 221; Tab. Sulis p. 61, no. 22; Kropp 2008: no. 3.9/1; Urbanová 2018: no. 195.

of Baldock), whose grave goods consisted of an urn, a globular glass and a plate (see Westell). The archaeological context makes this defixio all the more important, since it is the only British curse that has been discovered in a grave.

Image source: Collingwood, apud Westell 1931: fig. 8.

The curse was written on an irregularly shaped lead sheet, whose bottom and lateral edges are fully intact. The inscription contains four lines, which run from right to left and were written capitals, with several letters ‘mirrored’

Commentary: this defixio was discovered in 1930 in grave 309 of the necropolis in Walls Field (Clothall, east 336

 Britannia (i.e., written from right to left, such as q, r and d) or upside down (see the final a in Tacita).

(see 206, l. 2: sic liquat c ̣om[o] ạqua). For the Latin cognomen Tacita, see OPEL IV, 104.

Nothing is known about the events that led to the composition of this curse. In the text, we find signeficatur for significatur (l. 3; see Mann 1971: 221), and the use of the verb defigo (l. 3; see 338 and 339 for parallels). The similia similibus formula is also noteworthy since it is the second such structure documented in the British curses

After being inscribed, the tablet was pierced with five nails, four of which are still piercing the curse. Translation (Wright 1952:103): ‘Tacita, hereby accursed, is labelled old like putrid gore.’

346. Canonium, Kelvedon Provenance: domestic context. Current Location: Colchester Museum. Date: 3rd–4th centuries CE. Material: lead. Measurements: 5 × 10.5 cm. Reading:

5

quicumque res Vareni involaverit si mulrer si mascel sangu(i)no suo solvat et pecunie quam exesu`erit´ Mercurio dona(t) et Virtuti s(acra)

Bibliography: Wright 1958: 150; AE 1959, 157; Egger 1964: 16–17; García Ruiz 1967: no. 20; Solin 1968: no. 18; Mann 1971: 220 and 224; Tab. Sulis p. 61, no. 17; Gager 1992: no. 97; Versnel 1991: 84–85; Kropp 2008: no. 3.12/1; Cugusi 2014: 400; Urbanová 2018: no. 286.

and Virtus (the sole occurrence of this divinity in these texts) to avenge a theft. The text presents the following orthographic and phonetic features: mulrer for mulier (l. 2), mascel for masculus (l. 2, see Mann 1971: 224), sanguno for sanguino (l. 3), pecunie for pecuniam (l. 4) and dona for donat (l. 5). The curse has an all-inclusive formula used to identify the victim (though note the order in the pair si mul˹i˺er si mascel [l. 2], which is usually found the other way around; cf. 349, 441 and 451). In an attempt to encourage divine justice, the defigens decides that if the stolen property is recovered, half will be dedicated to the deities (l. 4; cf. 205, l. 4: demediam partem). If the property is not recovered, however, the thief ought to pay with his or her own blood (l. 3). Varenus is already attested as both a nomen and cognomen (vid. OPEL IV, 147).

Image source: Wright 1958: 150, fig. 20. Commentary: this defixio was discovered in 1956 inside an oven dated between the third and fourth centuries CE. The oven was located within a Roman settlement in field 204, on the south-eastern side of Orchard Road. The curse was written on a rectangular sheet of lead with irregular edges. The lower-left corner has been lost. The inscription contains five lines, which run from left to right and were written in old Roman cursive. Furthermore, this curse is noteworthy since it contains two iambic senarii (see Cugusi 2014 for a discussion; for a parallel, see 4).

After being inscribed, the tablet was folded. Translation (Gager 1992: no. 97): ‘Whoever stole the property of Varenus, whether woman or man, let him pay with his own blood. From the money which he will pay back, one half is donated to Mercury and Virtue.­’

Based on its content, this curse should be grouped among the defixiones in fures. Here, Varenus invokes Mercury

337

Sylloge of  Defixiones from the Roman West 347. Bravonium, Leintwardine Provenance: bath complex. Current Location: private collection. Date: unknown. [Inv. No.: unknown] Material: lead. Measurements: 5.7 × 9.6 × 0.2 cm. Reading:

5



Carinus Similis Consortius Comes Magloriu[s] Senorix Cunittus Cunittus Cuṇedecan`ẹṣ´ Ceanatis Ṭiberiṇ[us]

Bibliography: Wright 1969: 241; AE 1969/1970, 311a; RIB II, 2504.20; Tab. Sulis p. 61, no. 26; Tomlin 2008b: 346; Kropp 2008: no. 3.13/1; Urbanová 2018: no. 196.

the fractured areas in the upper third of the tablet. The inscription contains seven lines, which run from left to right and were written in capitals (the letter s always ends with a small horizontal stroke, as Wright has noted).

Image source: Wright 1969: 241, fig. 45. Courtesy of The Roman Society (Cambridge University Press).

The circumstances that led to the writing of this curse are unknown, since the text consists of nothing more than a list of the victims’ names, which are both Latin and Celtic. The following Latin cognomina are all well attested: Carinus (l. 1, see OPEL II, 37), Similis (l. 2, see OPEL IV, 83), Comes (l. 4, see OPEL II, 70) and Tiberinus (l. 7, see OPEL IV, 121). Among the Celtic names, note Cunittus (l. 6), Cunedecanes (l. 6) and Maglorius (l. 4), which are all attested here for the first time. Tomlin (2008b: 346, contra Wright) has proposed reading Maglorius (l. 4, instead of Maslorius), given the palaeographic difference between s and g, which seems reasonable. In this case, the name would be derived from the Celtic root *maglos (‘boss’).

Commentary: this defixio was discovered alongside 348 in 1964 during the excavations of the termae of the barracks at ancient Bravonium, which were built around the year 140 CE and then remodelled some twenty years later. These baths were abandoned in the fourth century CE. The tablets were found inside a small drain in the frigidarium, a unique context for depositing a defixio in the Roman West (on this, see section I.6.3 with n. 293). The present tablet was written on a roughly rectangular lead sheet, which has lost its top-left corner. It has been corroded along the lower-left corner and both of

348. Bravonium, Leintwardine Provenance: bath complex. Current Location: private collection. Date: unknown. Material: lead. Measurements: 2.3 × 6.6 cm. Reading:

Enestinus Motius Comitinus

Commentary: this defixio was discovered alongside 347 in 1964 during the excavations of the termae of the barracks at Leintwardine (ancient Bravonium), which were in use between the middle of the second century CE until the fourth century CE. The tablets were found inside a small drain in the frigidarium (on this unique context,

Bibliography: Wright 1969: 241; AE 1969/1970, 311b; RIB II, 2504.21; Tab. Sulis p. 61, no. 27; Kropp 2008: no. 3.13/2; Urbanová 2018: no. 197. Image source: Wright 1969: 241, fig. 45. Courtesy of The Roman Society (Cambridge University Press). 338

 Britannia see section I.6.3 with n. 293). The curse was written on a rectangular lead sheet, whose upper portion is corroded and has lost the top-left corner. The inscription contains three lines, which run from left to right and were written in capitals. The circumstances that led to the writing of this

curse are unknown since the text only consists of a brief list of personal names. Enestinus is unparalleled (but cf. Enestalinus in CIL VI, 32, 525), while Comitinus is found in a British inscription (see RIB III, 3062: Comitinu[s]). Motius is attested in Gallia Belgica (see CAG 8, p. 227).

349. Ratcliffe-on-Soar Provenance: unknown. Current Location: Brewhouse Yard Museum. Date: end of the 2nd–beginning of the 3rd century CE. Material: lead. Measurements: 8.4 × 5.7 cm. Reading: A donatur deo Iovi optimo maximo uṭ ẹxigạṭ ̣ pẹr ṃẹṇṭẹṃ pẹr memoriam per intus 5 pẹr intestinum per cor [p]er medullas per venas pẹr [-c.1-]+[-c.7-]++ạṣ +L+L++ ṣị mascel si femina quiuis

B involavit (dena)rios Caṇi Digni ut in corpore suo in brevi teṃp[or]ẹ pariat donatur 5 deo s(upra)s(crip)to decima pars eius pecuniae qụạṃ [so]ḷṿerit

Bibliography: Turner 1963; Egger 1964: 17–19; AE 1964, 168; Solin 1968: no. 21; Elsdon 1982: 16, 30, Fig. 9; Smith 1983: 934; Tab. Sulis p. 61, no. 24; Versnel 1991: 104, n. 124; Gager 1992: no. 98; Kropp 2008: no. 3.19/1; Urbanová 2018: no. 292; Sánchez Natalías 2022.

cursive, which is dated to the end of the second or beginning of the third century CE on palaeographic grounds. Based on its content, this curse should be grouped among the defixiones in fures. Here, the theft of a sum of money is denounced to Jupiter Optimus Maximus (A, ll. 1–2; for this invocation and for the only possible parallel in the Roman West, see 168). The use of the verb exigo (A, l. 3) finds parallels in the corpus from Bath (e.g., 239, 243, 246, 252) and here means ‘to exact’ (contra Turner 1963), i.e., the god should exact the stolen money through the victim’s body parts. The use of the adverb intus (A, l. 4) deserves mention, since has been interpreted as a synonym for viscera (so Turner 1963: 123–24); as the editor has proposed, taking it as intus would be a

Image source: Turner 1963: pl. IX. Courtesy of The Roman Society (Cambridge University Press). Commentary: this defixio was discovered still folded up in 1960 during agricultural work on the slope of Red Hill (Ratcliffe-on-Soar). The curse was written on a rectangular lead sheet, which is intact despite corrosion and a brownish patina. An opisthograph, the inscription contains 16 lines, which run from left to right and were written in old Roman

339

Sylloge of  Defixiones from the Roman West ll. 5–7), if the stolen property is returned. Compared with other cases, in which a third (see 356) or a half (see 205) of the stolen property is offered to the deity, Canius Dignus comes off as rather stingy.

stretch. Smith (1983: 934), for his part, noted that the word ‘appears as a sort of half-noun with dependent genitive in Apuleius, Met. VIII, 29, intus aedium, “intra aedes”.’ Also note the symbol for denarios (B, l. 1; cf. 455) and the abbreviation for suprascripto (B, l. 5: ssto; cf. 441).

The name Canius Dignus (B, ll. 1–2) belongs to the defigens: both the nomen and his cognomen are already attested (see OPEL II, 31 and III, 100 respectively).

In this text the god invoked is asked that ẹxigạṭ ̣ pẹr ṃẹṇṭẹṃ pẹr memoriam per intus pẹr intestinum per cor [p]er medullas per venas… (A, ll. 3–6), that is that the god take repayment from the victim’s own body. This is noteworthy, because in other curses from Britannia the price to be exacted is limited to the target’s blood (see 249, 299, etc.). The victim is then identified through a typical all-inclusive formula (A, ll. 8–9; cf.; for parallels, see e.g., 346, 441 and 451). The defigens’ request is to be fulfilled in brevi teṃp[or]ẹ (B, l. 3), a phrase which is first found here and is a magical deadline for the debt’s repayment (see 259, 367 and 452; for magical deadlines see Sánchez Natalías 2019a). Finally, in an attempt to urge the god to action, the defigens pledges 10 per cent of the stolen money (B,

After being inscribed, the tablet was folded. Translation (Turner 1963: 123, modified): ‘To the god Jupiter best and greatest there is given that he may exact… through his [i.e., the victim’s] mind, through his memory, his inner parts (?), his intestines, his heart, his marrow, his veins,… whoever it was, whether man or woman who stole away the denarii of Canius Dignus that in his own person in a short time he may balance the account. There is given to the god above named a tenth part of the money when he has (repaid it?).’

350. Ratcliffe-on-Soar Provenance: on the ground. Current Location: Kegworth Village Museum. Date: unknown. Material: lead. Measurements: 5.6 × 11.2 × 0.1 cm. Reading:

5



nomine Camụlorigi(s) et Titocun(a)e molam quam perdederunt in fanum dei dẹṿọvị c ̣uicumquẹ ṇ˹o˺ṃ[e]n involasit mola(m) illam ut samguin(em) suum ṃịṭtat usque die`m´ qụọ moriatur q[ui]c ̣umque iṇvọ[l]a[sit] hụrta moriatur et pa(b)ulatoriam quicumque [illam] involasit et ipse {moriato} mor[i]atur quicumquẹ illam involasit et VERTOGN de (h)ospitio vel vissacio quicumque illam involasit a devo moriotur

Bibliography: Tomlin 1993b: 310–14; AE 1993, 1087; Kropp 2008: no. 3.19/3; Mullen 2013; Adams 2016: no. 36; Urbanová 2018: no. 294.

Commentary: this defixio was discovered on the ground in 1990. The curse was written on an oblong lead sheet, which is intact despite minor corrosion and breakage that has damaged ll. 1–3. The inscription contains eight lines, which were written capitals and run from right to left with ‘mirrored’ letters (cf. 266, 340, 345 and 450).

Image source: Tomlin 1993b: 311, fig. 2. Courtesy of R.S.O. Tomlin. 340

 Britannia following u-vowel has either been lost in syncope between the sounds represented by V (…) and L, or the potential double VV has undergone orthographical simplification to V.’ There is no suggested reconstruction for VERTOGN (l. 7), although Adams has related the beginning of the term with the Celtic word vertragos ‘a fast dog used in hunting’ (2016: 421). Finally, note the lexeme devo (l. 8), which reflects the Celtic word for deus (for parallels, see 205 and 373).

Based on its content, this curse should be grouped among the defixiones in fures. Here, the theft of a molam (l. 1 for mulam), paulatoriam (l. 5) et VERTOGN de hospitiolo vel vissacio (l. 7) is reported. The text presents the following orthographic and phonetic features: Titocuna for Titocunae (l. 1), molam for mulam (ll. 1 and 3, see Adams 2016: 419), cuicumque for quicumque (l. 2), samguin for sanguin(em) (l. 3), moratur for moriatur (l. 4), visaccio for bisaccium (l. 7), ospitio for hospitio (l. 7) and moriotur for moriatur (l. 8). Also note the confusion of num[e]n for nom[e]n (l. 2) and the redundant moriato mor[i]atur (l. 6).

In the phrase nomine Camulorigi(s) et Titocun(a)e (l. 1), the use of nomine is noteworthy. Nomen tends to be used as a way of referring to an unknown victim (e.g., 220 and 221), but in this case it is used to introduce the names of the defigentes (making clear that somebody was writing on their behalf). The curse contains some attested tropes, such as the dedication of the stolen property to the deity invoked (l. 2), the mention of the victim’s blood as an ‘offering’ (l. 3, cf. 249), and the magical deadline by which the curse ought to take effect (ll. 3–4; cf. 250 and 459; for magical deadlines, see Sánchez Natalías 2019a). What stands out, however, is repetition of the death sentence, which is found four times (ll. 3–4, 6 and 8). This level of insistence is unparalleled in the corpus of Latin curses from the Roman West.

Adams has pointed to the form involasit (l. 2), which is ‘a sigmatic aorist subjunctive (optative) usually spelt with a double -ss- (…) This form is typical of archaic Latin legal language, and quicumque involasit may be a formula lifted from it’ (apud Tomlin 1993b: 313). The lexeme hụrta (l. 4) can be taken in two ways: if reconstructed as furta, it would allude to the stolen objects and go with the verb iṇvọ[l]a[sit]. Alternatively, hụrta can be taken as the adverb furto to be construed with moriatur. Adams (2016: 420), for his part, proposes yet another possibility, since so far only one object has been mentioned (molam). In his view, hurta could be related to the Germanic word hurda (‘hurdle, fence’), since some curse tablets contain Germanic words ‘which must have been introduced by Latin speakers coming from the Continent (...). One might speculate that a piece of hurdle fencing had been removed when the mule was taken.’ I think the first option is more plausible, given the Romance parallels for hurta (in Spanish, for instance, hurto (from Latin furtum), means ‘stolen object’). This would offer a nice parallel for the phrase fraudem fecit, which is well attested in this type of curse (see 237, 240, 357, for some parallels). The lexeme paulatoriam (l. 5) also can also be understood in various ways: as the editor has put it, ‘one possibility is that the scribe intended pabulatoriam (“pertaining to fodder”…). But if it is a substantive, then perhaps it means an “(animal) eating fodder”’ (Tomlin 1993b: 314). Following up on this argument, Mullen (2013: 268, with whom Adams agrees 2016: 421), has added, ‘in the context of the B/V alternation seen in this text, the expected B is represented by a V. Presumably the

The names Camulorigis and Titocuna are Celtic. The former is formed from Camulos (the god of war) and the suffix –rix, while the latter is without known parallel, but is probably a female name derived from tito and -cunos/a (‘dog’). After being inscribed, the tablet was folded in half. Translation (Tomlin 1993b: 312, modified): ‘In the name of Camulorigis et Titocuna I have dedicated in the temple of the god the mule which they have lost. Whoever stole that mule, whatever his name, may he let his blood until the day he dies. Whoever stole the objects of theft, may he die; and the fodder basket, whoever stole it, may he die also. Whoever stole it and the (…) from the house or the pair of bags (?), whoever stole it, may he die by the god.’

351. Ratcliffe-on-Soar Provenance: unknown. Current Location: Brewhouse Yard Museum. Date: 4th century CE. [Inv. No.: 1974.337] Material: lead. Measurements: 8.2 × 3.8 × 0.1 cm. Reading:

5

an̂noto de duas ocrias ascia(m) scalpru(m) ma˹ic˺a(m) si m(ulier) au˹t˺ si b(aro) RIANTINE duas partis deo ACCEVM

341

Sylloge of  Defixiones from the Roman West Bibliography: Williams 1964: 53–55; Elsdon 1982: 16, 30; Tomlin 2004a; Tomlin 2004c: 336–37; AE 2004, 856; Kropp 2008: no. 3.19/2; Tomlin 2017c: 11.01; Urbanová 2018: no. 293; Sánchez Natalías 2022.

articles of clothing is reported (ll. 1–3). The text presents the following orthographic and phonetic features: ocrias for ocreas (l. 2), ascia for asciam (l. 2), scalpru for scalprum (ll. 2–3) and aup for aut (l. 3). An̂noto with the preposition de (l. 1) is used to introduce a list of the stolen property. Tomlin has reconstructed macia (l. 3) as ma˹ic˺a(m), which is surely a safe conjecture given the context and the attestation of the diminutive manicilia in other British curses (210 and 367).

Image source: Tomlin 2004c: 337, fig. 2. Courtesy of R.S.O. Tomlin. Commentary: this defixio was discovered in 1963 on Red Hill during an excavation carried out before the construction of a power station. The curse was found, according to Williams, ‘some five inches below the surface in fine soil bearing carbon traces’ (1964: 53) alongside 22 coins, most of which are dated between 367 and 392 CE, and a small bone carving. The curse was written on a roughly oval lead sheet and is intact. The tablet has a purple patina on its surface and the lower-left corner has been corroded. The inscription contains five lines, which run from left to right and were written capitals. In addition to the coins, linguistic characteristics of the inscription point to a date in the late fourth century.

The all-inclusive formula si m(ulier) au˹t˺ si b(aro) (ll. 3–4) is widely attested in the British corpus (e.g., 249, 359 and 449), though the abbreviation used here is quite noteworthy (see 386 where we find SBSM for si baro si mulier). Given the context, Tomlin has cautiously offered the reconstruction involaverit for RIANTIME (l. 4). The phrase duas partes deo (l. 4) is clearly a reference to the dedication of part of the value of the stolen goods to the deity. The final ACCEVM has proven difficult to interpret and Tomlin has suggested the possible reading acce(pt)um. Translation (Tomlin 2004a: 348): ‘I make a note of two gaiters, an axe, a knife, a pair of gloves, whether woman or man… two parts to the god…’

Based on its content, this curse should be grouped among the defixiones in fures. Here, the theft of various tools and several

352. Wanborough Provenance: unknown. Current Location: unknown. Date: 1st–2nd centuries CE. Material: lead. Measurements: 5.5 × 5.5 cm. Reading: [---]deprẹ[co]r ṭe peto +[-c.1-]++++++EVENE++ [---] peto iudiciọ tuo qu[i?]d++++eculaṇṣ++ [---]ṭum ne ˹il˺li permittas bibere nẹc ̣ [do]rmịre nec ambulare neque ullaṃ 5 [---]S geṇṭisve unde ille nascịṭ [ur?]+++d˹ie˺ta ulla nec alumen [tum?] PR+ ve(he)mente (?) loquantur et R+ [---]+++ụgabatur certum sciuṇṭ (vacat) sị (vacat) 10 [---]+MEVERECAME+(?)VE (vacat) [---]+MEOR (vacat)



Bibliography: Rea 1972; AE 1975, 530; Tab. Sulis p. 61, no. 7; Kropp 2008: no. 3.23/1; Urbanová 2018: no. 307. Image source: Wright apud Rea 1972: 364, fig. 22.

contains 11 lines, which run from left to right and were written by a trained hand in old Roman cursive that can be dated to the first or second century on palaeographic grounds.

Commentary: this defixio, discovered in unknown circumstances, was written on a quadrangular lead sheet. Upon discovery, the tablet’s lower-left corner was broken off and the curse was severely corroded (especially along the top and left edges). The inscription

Although the reasons that led to the writing of this curse are unknown, the presence of certain formulae suggests that it belongs among the defixiones against thieves. In particular, this idea is supported by the presence of the phrase peto iudiciọ tuo (l. 2), used to ask for divine justice, 342

 Britannia d˹ie˺ta from EITA (l. 6), which would make sense alongside alumentum. Furthemore, he has conjectured ve(he)ment˹i˺ or ve(he)mente[r] (l. 7). Rea (1972: 366–67) was puzzled by the end of the text, given the plural in loquantur (l. 7) and sciunt (l. 8). There are no possible reconstructions for the final lines.

and the non permittas formula (ll. 3–5), which is found in many British inscriptions (e.g., 205, 259, 365, 366, 443 and 452) and was used to hamper the vital functions and well-being of the victim. In this case, Rea has argued that this part of the curse extends to the thief’s ancestors (neque ullam [partem manere sinas illiu]s gentisve unde ille nascitur…, ll. 4-5), who are also to be deprived of alumen[tum] (ll. 6–7).

Translation (Rea 1972: 366, modified): ‘…I pray, I beg (to deliver?) to your judgment (the man who?)… (I beg)… that you do not permit him to drink nor eat nor sleep nor walk…’

Tomlin for his part, has proposed the reconstruction of the verb deprẹ[co]r (l. 1; based on analogy with 16) and of

353. Old Harlow Provenance: pit or well. Current Location: Harlow Museum. Date: 3rd–4th centuries CE. Material: lead. Measurements: 7.2 × 5.4 cm. Reading: A diọ Ṃ(ercurio) dono ti negotium Et tern(a)e et ipsam nec sit i(n)vidi(a) me(i) 5 Timontneo san guiṇe suo

B dono tibi Mercurius aliaṃ neg[o-] tium NAVIN[-c.2-] 5 ++N[-c.6-] nec ̣ +[-c.3-]+[-c.3-] MIN[-c.3-]ṣang[uine] suo

Bibliography: Hassall 1973: 325–27; AE 1975, 542; Tab. Sulis p. 61, no. 18; Kropp 2008: no. 3.17/1; Urbanová 2018: no. 208.

Commentary: this defixio was discovered in 1970 during the excavation of a pit or a well (the editors’ report does not specify) on the site of the Holbrooks factory, which contained sherds of Roman pottery dated to the third and fourth centuries CE. The curse was written on a rectangular lead sheet, which has lost a good part of the inscribed surface and has been broken along the left and

Image source: Hassall 1973: 326–27, figs. 17–18. Courtesy of M.W.C. Hassall.

343

Sylloge of  Defixiones from the Roman West and perhaps in 168). The phrase sanguiṇe suo (A, ll. 5–6 and B, ll. 7–8; this is normally spelled sangune suo [e.g., 270]), likely alludes to the victim’s punishment. According to Wright and Hassall, NAVIN (B, 1. 4) should be reconstructed as the Latin personal name Navinius (for the name, see OPEL III, 96). Eterna and Timotheus are attested cognomina (see OPEL I, 47 and IV, 122 respectively).

bottom edges. An opisthograph, the inscription contains 14 lines, which run from left to right and were written in capitals (with cursive e made of two vertical strokes). Based on its content, this curse belongs to the realm of erotic magic (side A), which makes it uncommon among the corpus of British curses, which is mostly comprised of defixiones against thieves. The text presents the following orthographic, phonetic and grammatical features: dio for deo (A, l. 1), Etterne for Aeternae (A, ll. 2–3), ividi for invidia (A, l. 4), Timotneo for Timotheo (A, l. 5), aliam for alium (B, l. 3) and Mercurius for Mercurio (B, l. 3). Since invocation of Mercury is indisputable on side B, the editors have rejected the idea that DIOM (A, l. 1) refers to D(eo) I(ovi) O(ptimo) M(aximo) (Jupiter is invoked in 349,

After being inscribed, the tablet was pierced with a nail and folded like an accordion with five folds. Translation (Hassall 1973: 325): ‘To the god Mercury, I entrust to you my affair with Eterna and her own self, and may Timotneus feel no jealousy of me at the risk of his lifeblood. I entrust to you, O Mercury, another transaction…’

354. Uley Provenance: sanctuary. Current Location: private collection. Date: 2nd–3rd centuries CE. Material: lead. Measurements: (1.8) × (3.1) cm. Reading: B [---]AS date+[---] [---]ṣummaṃ DIV[---] ++++++

A ++++++ [--- r]ogọ AB or AD (?) QV[---] [---]OMMVN++[---]



Bibliography: Hassall 1973: 324, no. 2; AE 1975, 538; Tab. Sulis p. 61, no. 12; Tomlin 1993a: 127, no. 6; Kropp 2008: no. 3.22/1.

(Uley). An opisthograph, the inscription contains six lines, which run from left to right. After being inscribed, the tablet was folded, which led to the curve along the fragment’s left side.

Commentary: this rectangular lead fragment was discovered in 1972 during agricultural work on West Hill

355. Uley Provenance: sanctuary. Current Location: British Museum. Date: 2nd–3rd centuries CE. [Inv. No.: 1978, 0102.77] Material: lead. Measurements: 13.5 × 8.5 cm. Reading: A deo Mercurio Cenacus quẹrịtur de Vitalino et Nata lino filio ipsius d(e) 5 iument[o?] quod eị rap tum est eṭ rogat deum Mercurium ut nec ante sa nitatem

344

 Britannia B habeant nissi [[nissi]] repraese[n] taverint mihi [iu] mentum quod rạ5 puerunt et deo devotioneṃ quạ[m] ipse ab his ex postulaverit

Bibliography: Hassall and Tomlin 1979: 340–41; AE 1979, 383; Tab. Sulis p. 61, no. 13; Tomlin 1993a: 119–29, no. 1; Kropp 2008: no. 3.22/2; Urbanová 2018: no. 295.

1993: 33–78; for the discovery of the curse tablets see Woodward 1993: 113). This curse was written on a roughly rectangular lead sheet. The tablet is in good condition, although it has some small fractures along its edges and certain points of its surface are somewhat corroded. An opisthograph, the inscription contains 18 lines, which run from left to right and were written in old Roman cursive.

Image source: Hassall, apud Hassall and Tomlin 1979: 340–41, figs. 20–21. Courtesy of M.W.C. Hassall. Commentary: this defixio and the following (355–40) were discovered during the excavations carried out between 1976 and 1979 in West Hill, Uley. There, archaeologists excavated a religious complex that was first established during the Neolithic. In roughly the second century CE, a Romano-Celtic temple was built in stone on the site and dedicated to Mercury. The temple is surrounded by a series of buildings that scholars have identified as living quarters, guest accommodation and shops. The temple remained in use until the beginning of the fifth century when all the votive objects were scattered and buried. Later, a palaeo-Christian church was built on the site of the old temple with a separate baptistery, all of which was surrounded by an earthen mound. The complex was abandoned in the Middle Ages. The excavations of the complex have brought to light an extraordinary collection of votive offerings, including organic remains (mostly goat, sheep and cockerel bones), cult statues, altars, figurines and 140 curse tablets. Of this collection, only 86 of the lead tablets were inscribed. In the present sylloge, 20 are discussed in depth (for the rest, see the introduction to this chapter). It has been surmised that the defixiones had originally been deposited in the pool located in the temple’s cella, from which they were removed later in Antiquity to be scattered through the sacred area in different heaps that correspond to phase 5 or are ‘unstratified’. Generally speaking, stratigraphic and palaeographic evidence suggests that these tablets were inscribed between the second and fourth centuries CE (for the Celtic-Roman temple, see Woodward and Leach

Based on its content, this curse should be grouped among the defixiones in fures. Here, Cenacus denounces to Mercury the theft of an animal from his flock. The text presents the following orthographic features: nissi for nisi (B, l. 2), rapuerunt for rapuerint (B, ll. 4–5), devotione for devotionem (B, l. 6) and qua for quam (B, l. 6). The verb repraese[n]taverint (B, ll. 2–3) deserves attention: this technical term, which has a legal ring to it, also appears in two other tablets from Uley and means ‘to pay at once’ (OLD 2). Also note devotioneṃ (B, l. 6), which is not used in its Classical meaning, but rather describes the respect that the perpetrators ought to show to Mercury (as used in Christian authors; see Lewis and Short s.v., I, B, 2). The phrase nec ante sanitatem (A, ll. 8–9) recalls the non permittas formulae and is used here to harm the physical well-being of the victims, unless they show devotion to Mercury and return the stolen object to the defigens. The request for property to be returned to its rightful owner is unusual in the British corpus, where it is normal to ask for purloined goods to be deposited in the temple of the deity invoked (e.g., 215, 237 and 356). The Celtic personal name Cenacus (A, l. 2) is first attested here in Britannia. Vitalinus and Natalinus (A, ll. 3–4) are Latin cognomina already attested (vid. OPEL IV, 176 and CIL VII, 3 respectively). 345

Sylloge of  Defixiones from the Roman West from him, and asks the god Mercury that they may have neither health before/unless they return at once to me the draught animal which they have stolen, and to the god the devotion which he has demanded from them himself.’

After being inscribed, the tablet was folded twice. Translation (Tomlin 1993a: 119): ‘Cenacus complains to the god Mercury about Vitalinus and Natalinus his son concerning the draught animal which has been stolen

356. Uley Provenance: sanctuary. Current Location: British Museum. Date: 4th century CE. [Inv. No.: 1978, 0102.78] Material: lead. Measurements: 6 × 8.3 cm. Reading: A commonitorium dẹọ «Mercurio» a Satur nina muliere de lintia mine quod amisit ut il5 le qui hoc ̣ circumvenit non ante laxetur nissi quandọ res s(upra)dictas ad fanum s(upra)d[ic-] tum attulẹrit si vir si ṃu liẹr si servus sị ḷịber

B deo s(upra)dicto terṭiam partem donat ita ût (vacat) exsigat istas res quae (vacat) s(upra)s(crip)ta sunt (vacat) 5 ACA quae pert deo Silvano tertia pars donatur ita ut hoc exsigat sị vir si fẹmina si sẹr[v-] ̣ us si liber[-c.2-]+[-c.7-]+at

Bibliography: Hassall and Tomlin 1979: 343–44; AE 1979, 384; Tab. Sulis p. 61, no. 14; Tomlin 1993a: 120–22, no. 2 Kropp 2008: no. 3.22/3; Tomlin 2017c: 12.36; Urbanová 2018: no. 296; Sánchez Natalías 2022.

define herself as a woman (muliere, A, l. 3; this is without parallel), first calls upon Mercury. That said, the author initially wrote Marti Silvano, perhaps as a mistake or as a reflection of an unusual association between these divinities: in 357, Mercury is called Mars (deo Mrti Mercurio, l. 1) and elsewhere Mars is linked to the Celtic god Cocidius (RIB 602, 993, 2015, 2024) who, in turn, is identified with Silvanus (RIB 1578).

Image source: Hassall, apud Hassall and Tomlin 1979: 344, fig. 22. Courtesy of M.W.C. Hassall. Commentary: this defixio, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), was written on a rectangular sheet of lead. The tablet is in good condition, despite damage to its corners and the areas that fall along the fold lines. An opisthograph, the inscription contains 18 lines, which run from left to right and were written in capitals.

According to Tomlin, commonitorium (A, l. 1) is a ‘quasilegal term, found here for the first time in a curse tablet text (…) It does not occur before the fourth century, when it becomes quite common, sometimes in the special sense of a petition to a superior (…) which would seem to be the sense here’ (1993a: 121). The verb circumvenio (A, l. 5, see Lewis and Short, s. v., II B–C 2) is here used as a synonym for the more common involo and is not attested elsewhere in the collection of British curses. Finally, note the abbreviations s(upra) dictas/-o (A, ll. 7–8 and B, l. 1) and s(supra)s(cripta) (B, l. 4, parallel to deo s(upra)s(crip)to in 349, B, l. 5).

Based on its content, this curse should be grouped among the defixiones against thieves. Here, Saturnina denounces the theft of an article of clothing to Mercury (A, l. 2) and Silvanus (B, l. 5). The defigens, who takes pains to 346

 Britannia The phrase non ante laxetur (A, ll. 5–6; cf. 459: …n[o]n aṇ[t]e eum lasẹt quam) serves as variant of the common non permittas formula (e.g., 259) that is meant to keep the guilty party from being able to rest and be healthy until (s)he deposits the stolen goods in the deity’s temple. Furthermore, Saturnina makes an offering of one third of the stolen property’s value to Mercury and Silvanus (B, ll. 1–2 and l. 6), while she keeps the last third for herself. This is a very common technique to motivate the deity to help recover the stolen object (e.g., 343 and 359). The curse’s victim is referred to with an all-inclusive formula (A, ll. 8–9 and B, ll. 7–8) that is well attested in Britannia (e.g., 215, 237, 271, etc.). Finally, ACA quae (B, l. 5) is difficult to interpret. According to Tomlin, ‘it might be understood as ac a quibus, but the conjunction ac (instead of et) is otherwise unknown in British curse tablets, and

the grammar of a quae …tertia pars is very harsh’ (1993a: 122). After being inscribed, the tablet was folded twice. Translation (Tomlin 1993a: 121): ‘A memorandum to the god Mercury (over Mars Silvanus) from Saturnina a woman, concerning the linen cloth which she has lost. (She asks) that he who has stolen it should not have rest before/unless/until he brings the aforesaid property to the aforesaid temple, whether man or woman, whether slave or free. She gives a third part to the aforesaid god on condition that he exacts this property which has been written above. A third part… what she has lost is given to the god Silvanus on condition that he exact it, whether man or woman, whether slave or free…’

357. Uley Provenance: sanctuary. Current Location: British Museum. Date: 2nd century CE. [Inv. No.: 1978, 0102.79] Material: lead. Measurements: (5.4) × (9.8) cm. Reading:

5

deo Mrti Mercurị[o---] anulus aureụs de hoṣ[pitiolo?---] erịṭ +(?)et pedica ferre[a---] S+ qui fraudem fecị[t---] ̣ R[-c.2-] deụṣ inveniạt

Bibliography: Hassall and Tomlin 1979: 344–45; AE 1979, 385; Tab. Sulis p. 61, no. 15; AE 1991, 1123; Tomlin 1993a: 122–23, no. 3; Kropp 2008: no. 3.22/4; Urbanová 2018: no. 297.

sustulerit. Next, the text mentions a second item that nicely contrasts with the gold ring: pedica ferre[a] (l. 3), which, in Tomlin’s words, is ‘unparalleled and obscure; perhaps a rhetorical conceit, an iron fetter to punish the thief of a gold ring’ (1993a: 123). Although this hypothesis is doubtlessly suggestive, the fetter could also be a second item stolen rather than the reference to a totally unusual punishment. That said, the fragmentary nature of the text precludes any firm conclusion. The curse concludes with two well documented formulae: the first, qui fraudem fecị[t], is ̣ normally part of an all-inclusive formula used to identify the wrongdoer (cf. 365 and 370); the second, deụṣ inveniạt, is used to summon divine justice (see 249 and 304).

Image source: Tomlin 1993a: 123, fig. 105. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), has been broken into two fragments that fit together. The curse, which was originally written on a long strip of lead, has lost its right half and does not have any of its original edges intact. The inscription contains five lines, which run from left to right and were written in capitals.

Finally, we must note the joint invocation of Mars and Mercury. Our only parallel is 355, since everywhere else these deities appear separately (for Mars, see 238 and 302; for Mercury, see 258 and 353).

Despite its fragmentary nature, this curse can be grouped among the defixiones in fures. Here, the defigens denounces to Mars and Mercury the theft of an anulus aureụs (l. 2, for parallels, see 205 and 302), presumably from the victim’s house (de hoṣ[pitiolo], l. 2; for parallels, see 365 and 443). The invocation of both deities is noteworthy, since the sanctuary was dedicated to Mercury alone (but cf. 373). L. 3 begins with a verbal form, perhaps involaverit or

After being inscribed, the tablet was folded. Translation (Tomlin 1993a: 123): ‘To the god Mars/ Mercury (…) gold ring from (…) [house] (…) and iron fetter (…) who did wrong (…) let the god discover.’ 347

Sylloge of  Defixiones from the Roman West 358. Uley Provenance: sanctuary. Current Location: British Museum. Date: 2nd century CE. [Inv. No.: 1978, 0102.80] Material: lead. Measurements: 12.4 × 6.6 × 0.1 cm. Reading:

5



10



15



Biccus dat Mercurio quidquid pe(r)d(id)it si vir si mascel ne m˹eia˺t ne cacet ne loqụaṭụr ne dormiat nẹ vigịlet nec sạḷutem nẹc sanitatem nẹssa in templọ Mercurii perṭulerit ne co(n)scientiam de perferat nessị me intercedeṇte

Bibliography: Tab. Sulis p. 61, no. 16; AE 1988, 840; Tomlin 1993a: 124–26, no. 4; Kropp 2008: no. 3.22/5; Urbanová 2018: no. 358.

presumably would better be taken as ignoscentiam, as a synonym for the Classical indulgentia since ‘the context demands the sense of “forgiveness” or “pardon”’; that said, the word is extremely rare. Also note intercedente (l. 15), which is not found elsewhere in the British curses and which Biccus uses in a legal sense.

Image source: Tomlin 1993a: 124, fig. 107. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), was written on an oblong lead sheet. As I was able to confirm during an autopsy (December 2009), the tablet is intact and in good condition, despite the corrosion found around the edges. The inscription contains 16 lines, which run from left to right and were written in capitals (measuring between 0.4 and 0.7 cm in height) by a trained hand.

The use of the phrase quidquid pe(r)d(id)it (ll. 2–3) to refer to the stolen object without the addition of further details is unusual; perhaps it indicates that Biccus was saying out loud what was stolen, otherwise it could be a signal of his faith in divine omniscience. Next, we find a botched all-inclusive formula, in which we find si vir paired with si mascel (ll. 3–4) instead of contrasting genders (si femina or si mulier). The text also has a fragmentary non permittas formula, which seeks to paralyse the victim’s vital functions (ll. 4–9), though we do not have a direct reference to the victim. Despite these problems, we must note the novelty of formulation: the defigens asks ne m˹eia˺t ne cacet (ll. 4–5, cf. 246), ne loqụaṭụr (ll. 5–6; here Tomlin has compellingly proposed the reconstruction ne taceat; see 339), ne dormiat nẹ vigịlet (ll. 6–7, cf. 352, l. 4) and finally nec sạḷutem nẹc sanitatem (see 446 B, ll. 3–4), until the thief deposits the stolen object in the temple of Mercury (ll. 9–12, cf. 215). Regarding the final phrase nessị me intercedeṇte (ll. 14–16), the editor has suggested that the practitioner ‘did not know how to construct a

Based on its content, this curse should be grouped among the defixiones in fures. Here, Biccus denounces to Mercury a theft. Judging the text, the editor has noted that it ‘contains traces of the Vulgar, spoken Latin (…) [the author...] was handling formulas he only half understood’ (1993a: 124). Tomlin’s appraisal helps explain the following features: pedit for perdidit (l. 3), maiet for meiat (l. 4), nessa for nesi (ll. 9–10), nessi for nesi (l. 14), coscientia for conscientia (ll. 12–13) as well as an erroneous in (l. 10) and de (l. 13). Tomlin has argued that the word co(n)scientiam (ll. 12–13) 348

 Britannia Translation (Tomlin 1993a: 125): ‘Biccus gives Mercury whatever he has lost (that the thief), whether man or male (sic.), may not urinate nor defecate nor speak nor sleep nor stay awake nor [have] well-being or health, unless he bring (it) in the temple of Mercury; nor gain consciousness (sic.) of (it) unless with my intervention.’

conditional clause without a formula to guide him, and may indeed have confused nessi (nisi) with sine or even non’ (1993a: 126). Finally, Biccus is a Celtic name and possibly the variant of Beccus or Becco. After being inscribed, the tablet was folded.

359. Uley Provenance: sanctuary. Current Location: British Museum. Date: 2nd–3rd centuries CE. [Inv. No.: 1978, 0102.81] Material: lead. Measurements: 7.2 × 7 cm. Reading:

5



nọmen furis (vacat) [qu]i frenẹm inṿọlaverit si l[i]ber ṣi servus si baro si ṃụlier deo dona«tor duas» partes AFIMA sua tertia ad sanitatem

Bibliography: Tomlin 1989: 327–28; AE 1989, 486; Tomlin 1993a: 126–27, no. 5; Kropp 2008: no. 3.22/6; Urbanová 2018: no. 299; Sánchez Natalías 2022.

Commentary: this defixio, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), was written on an irregularly shaped lead sheet, which is intact and in good condition, despite some corrosion on the upper-left corner. The tablet was hammered out before being inscribed. The inscription contains eight lines, which run from left to right and were written in old Roman cursive by a trained hand.

who is referred to with an all-inclusive formula (ll. 3–4; for parallels, see 270 and 447), to the deity invoked (ll. 1–5; cf. 215). The mention of duas partes (l. 5) indicates that the defigens was handing over to Mercury two thirds of the bridle’s value (cf. with 351 and 356). After l. 6, the author made various errors while copying the text: for example, Tomlin has taken (though without certainty) AFIMA sua as a bungled attempt at femina sua; while tertia should be taken as referring to the part of the stolen property given to the god (as it happens in other curses, e.g., 205, 356, etc.). We would expect the final ad sanitatem to form part of a non permittas formula used to harm the victim’s physical well-being until the stolen object returned to the temple (cf. 205, 237, etc.).

Based on its content, this curse should be grouped among the defixiones against thieves. Here, the theft of a bridle is reported (frenem for frenum, l. 2). In an attempt to recover it, the defigens dedicates the thief,

Translation (Tomlin 1993a: 126): ‘The name of the thief who has stolen (my) bridle, whether free or slave, whether man or woman, is given to the god (…) two parts from his wife (?), a third to (his) health.’

Image source: Tomlin 1993a: 126, fig. 109. Courtesy of R.S.O. Tomlin.

349

Sylloge of  Defixiones from the Roman West 360. Uley Provenance: sanctuary. Current Location: British Museum. Date: 4th century CE. Material: lead. Measurements: 7.6 × 7 cm. Reading:

5



Lucilia Mellossi AEXSIEVMO Minu(v)assus Senebel[l]en̂ae

Bibliography: Tomlin 1993a: 128, no. 33; Tomlin 1995: 378–79; AE 1995, 987; Kropp 2008: no. 3.22/12; Urbanová 2018: no. 203.

this ending, and the absence of a patronymic, the word is not a personal name; perhaps aestimo (…), but this verb is not found in curse tablets’ (1995: 379).

Image source: Tomlin 1995: 378, fig. 1. Courtesy of R.S.O. Tomlin.

Lucilia is a well attested Latin cognomen (see OPEL III, 35) while her father’s name, Mellossus, is a Latin cognomen only attested in Africa (see Kajanto 19822: 284). Finally, Minuvassus Senebellen̂a[e] is likely the name of Lucilia’s husband. Minuvassus is first attested here and is made of two Celtic elements: minus/minius/minutus (found in Gallia and Britannia) and vassos (‘slave’). The matronymic that follows Minuvassus is without known parallel but is also formed from two Celtic elements: senus and belenos/bellinus (for the use of matronymics in British curses, see 235, 303 and 454).

Commentary: this defixio, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), was written on an irregularly shaped lead sheet. Its surface is uneven and quite damaged (especially along the edges and fold line). The inscription contains six lines, which run from left to right and were written in capitals. The circumstances that led to the writing of this curse are unknown. The text consists of two personal names, the first of which is followed by a patronymic (ll. 1–2) and the second by a matronymic (ll. 4–6). The two onomastic formulae are separated by AEXSIEVMO (l. 3), which is difficult to interpret. According to Tomlin, ‘to judge by

After being inscribed, the tablet was folded twice. Translation (Tomlin 1995: 378): ‘Lucilia (daughter) of Mellossus (…) Minu(v)assus (son) of Senebellena.’

361. Uley Provenance: sanctuary. Current Location: British Museum. Date: 2nd century CE. Material: lead. Measurements: 9.8 × 8.4 × 0.1 cm. Reading:

5



10



deo Mercurịo Docilinus QVAENM Varianus et Peregrina et Sabinianus quị pecori meo dolum ṃạḷụm intụḷerunt et INT+RR[-c.1-] prolocụṇtur rogo te ut eos maxịmọ ḷẹtọ adigas ṇec eis saṇitạ[t]ẹ[m n]e[c] somnum pẹrm[itt]as nịṣị a te quod mi[hi] ạdṃ[i-] niṣ[tr]ạṿerint redeṃ[e]rint

350

 Britannia Bibliography: Tomlin 1989: 329–31; AE 1989, 487; Adams 1992: 7–8; Tomlin 1993a: 128, no. 43; Tomlin 2002: 172; Kropp 2008: no. 3.22/16; Tomlin 2017c: 11.03; Urbanová 2018: no. 300.

a certain vagueness: ‘The “harm” done is not specified (unless in 6–7), but was presumably an ailment blamed on persons known to bear Docilinus a grudge’ (Tomlin 1989: 330).

Image source: Tomlin 1989: 329, fig. 2. Courtesy of R.S.O. Tomlin.

Starting in the seventh line, the defigens asks that the wrongdoers be punished with death: maximo leto adigas (l. 8, for some parallels, see 215, A, ll. 11–12: ṃạximo letum [a]digat and also 530, l. 16: ṇec[et]i[s] eum pes(s) imo leto). Next, we find a non permittas formula that seeks to impair the victims’ health and well-being (cf. 237, 358, etc.). The final formula (ll. 11ff.) is first attested here and, for Tomlin, should be linked to the common idea that the value of a stolen object can be reimbursed in blood (in this case that of a wounded animal).

Commentary: this defixio, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), was written on a roughly oval lead sheet. The tablet was hammered flat before being inscribed. Though intact, the right half is corroded and has breakage that has damaged ll. 9–10. The inscription contains 13 lines, which run from left to right and were written in capitals. Palaeographically speaking, the tablet is almost identical to 215 and perhaps they were even written by the same hand.

Docilinus is a Celtic personal name derived from Docca, while Varianus, Peregrina and Sabinianus are well attested Latin cognomina (see OPEL IV, 147; III, 132 and IV, 39, respectively).

The curse denounces Varianus, Peregrina and Sabinianus for making a herd animal (probably a sheep) sick (pecori meo, ll. 4–5; note the p, which slightly resembles a b). Unfortunately, due to corrosion, there is no reconstruction possible for INT+RR[-c.1-] (l. 6). The word prolocụṇtur (ll. 7–8) is a vulgarization of proloquuntur. The phrase dolum ṃạḷụm intụḷerunt (ll. 5–6; the only parallels are 157, A, l. 5: ...dolum malu(m) fecerunt and 493 ll. 1 and 5: ...dolum malum adm[isit] and ...dolum malum adhibet) comes from the realm of law. The editor notes

After being inscribed, the tablet was folded. Translation (Tomlin 1989: 330): ‘To the god Mercury (from) Docilinus. Varianus and Peregrina and Sabinianus who have brought evil harm on my beast and are (…) I ask that you drive them to the greatest death, and do not allow them health or sleep unless they redeem from you what they have administered to me.’

362. Uley Provenance: sanctuary. Current Location: British Museum. Date: 4th century CE. Material: lead. Measurements: 6.6 × 4.2 cm. Reading:

Ạụnilluṣ Vịc ̣[a]rianạ Coviṭius Ṃini donaṭ 5 Varicillum Minura Atavacụm (traces) (traces)

Bibliography: Tomlin 1993a: 129, no. 49; Tomlin 1995: 376–77; AE 1995, 986; Kropp 2008: no. 3.22/18; Urbanová 2018: no. 206.

Commentary: this defixio, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), was written on a roughly oval lead sheet. Its surface is uneven, corroded along the fold lines and quite damaged in general. The inscription contains nine lines, which run from left to right along previously traced

Image source: Tomlin 1995: 377 fig. 3. Courtesy of R.S.O. Tomlin. 351

Sylloge of  Defixiones from the Roman West lines (for these, cf. 71). The first seven lines were written in capitals and the final two (with much weaker strokes) in a style that is hard to categorize.

no. 3.22/18), an easier solution would be to reconstruct Minura as an accusative (with omission of final -m), while taking donaṭ for donaṭ.

The circumstances that led to the writing of this curse are unknown. The text contains a list of personal names in the nominative (ll. 1–4) and the accusative (ll. 5–7), which is unusual in Britain, where curses containing lists of names are normally in the nominative (for some parallels, see 222, 235, 256, etc.). Here, nominatives and accusatives have been understood as a deliberate choice on the practitioner’s part to distinguish the cursers from the cursed (for a parallel, see 495; in general, see section I.7.2). In line with this interpretation, the victims are handed over to the god with the verb donaṭ (l. 4, in which the vertical stroke of the final t is clearly visible). On the contrary, Urbanová (2018: 289) understands the oscillation between cases as a mistake made by the author of the text. Nevertheless, and as already proposed by Kropp (2008:

Aunillus (l. 1) and Varicillius (l. 5) are Celtic names, whereas Minius/Minnius (l. 4, the patronymic of Covitius) and Minura (l. 6, Holder, apud Tomlin 1995: 377, has pointed to Appian Hisp. 71 where Minurus is a friend of Viriathus in Lusitania; here it is first found in the epigraphic record) are Latin. Here we find the cognomina Covitius (l. 4), Vic[a]riana (l. 2, reconstruction based on Vicarius) and Atavacus (l. 7; perhaps a Celtic name) for the first time. After being inscribed, the tablet was folded twice. Translation (Tomlin 1995: 376, modified in line with Kropp 2008): ‘Aunillus, Vicariana, Covitius (son) of Minius give Varicillus, Minura and Atavacus…’

363. Uley Provenance: sanctuary. Current Location: British Museum. Date: 4th century CE. Material: lead. Measurements: (3.9) × (9.1) cm. Reading:

[-c.6- Mercur]io RID[-c.1-]SONAE[---]`LTELL[---]´ [---]ESVNT sus[pe]cti sunt inter [---] [-c.1-]LLVS[-c.2-]EVSRE[-c.1-]VMINVENETET[---] lami[l]la una et anulli quator[---]

Bibliography: Tomlin 1993a: 129, no. 50; Tomlin 1998: 433–34; AE 1998, 819; Kropp 2008: no. 3.22/19.

curse should be grouped among the defixiones in fures. The text presents the following orthographic features: lamilla for lamella, anuli for anulli and quatuor for quattuor (all in l. 5). According to Tomlin, lamilla is ‘a misspelling of lamella in the sense of “(silver) plate”. Lamina and its diminutive lamella are used in the sense of “silver”, i. e. “money”’ (1998: 434). LTELL[---] (between ll. 1–2), for its part, can be understood as cultellus or a personal name. ESVNT (l. 3) could be part of a compound verb of sum (something like desunt or praesunt). With the unparalleled phrase sus[pe]cti sunt inter (l. 3), we find a reference to the suspected thieves. L. 4 preserves the remains of what looks like a deus inveniat formula (cf. 249 and 357) which alludes to the god’s role in locating stolen property.

Image source: Tomlin 1998: 434, fig. 1. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), was written on a lead tablet that probably was rectangular. It has been broken into three fragments, two of which fit together and belong to what must have been the end of a long strip of lead. The tablet is in poor condition, given that the surface is generally corroded (especially the edges and along the fold lines). The inscription contains four lines, which run from left to right and were written in capitals.

After being inscribed, the tablet was folded at least twice.

Despite the fragmentary nature of the text, the use of the phrase suspecti sunt (l. 3) and the mention of the objects lami[l]la una et anulli quatuor (l. 4) suggests that this

Translation (Tomlin 1998: 433): ‘…one piece of (silver) plate and four rings…’

352

 Britannia 364. Uley Provenance: sanctuary. Current Location: British Museum. Date: 2nd–3rd centuries CE. Material: lead. Measurements: 9.5 × 6 cm. Reading:

deo Mercurio Mint«l»a Rufus donavi eos vel mulier vel PARIVSLIIFASPATEM 5 [ma]teriam sagi (vacat) donavi

Bibliography: Tomlin 1993a: 129, no. 55; Tomlin 1995: 371–73; AE 1995, 984; Kropp 2008: no. 3.22/22; Urbanová 2018: no. 301.

clear: Rufus like Priminus may have used it as a second name or self-identification, or perhaps the petitioner’s name was composite, Mintlarufus (“Redprick”)’ (1995: 371).

Image source: Tomlin 1995: 372, fig. 1. Courtesy of R.S.O. Tomlin.

The text contains a variant of a common all-inclusive formula that is used to describe the wrongdoer who is in turn handed over to the deity (ll. 3–4; note the use of vel instead of the more common si). In l. 4, which has been difficult to interpret, we do not find the expected mascel, baro or vir, but rather read PARIVSLIIFASPATEM. This suggests an error on the author’s part in which he jumbled several formulae, such as the continuation of the allinclusive (PAR for vir, RIVS for the ending of servus and LIIF for a bungled lib(er?)?), with a reference to the deity’s maiestatem or the dedication of part of the stolen property (ASPATEM for the ending of maiestatem or partem?). The curse ends with donavi, which is separated from the rest of the text and centred. This verb emphasizes the dedication of the victim and stolen goods to the god Mercury.

Commentary: this defixio, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), was written on a rectangular lead sheet with three of its corners cut off. Despite some corrosion, a hole in the top third of the tablet (caused by shoddy craftsmanship) and damage along the fold lines, the tablet is nevertheless in good condition. The inscription contains six lines, which run from left to right and were written in old Roman cursive by a trained hand. Based on its content, this curse should be grouped among the defixiones in fures. Here, the theft of a [ma]teriam sagi (l. 5, ‘the material of a cloak’) is reported to Mercury. The curse opens (l. 1) with a common formula that gives the names of the deity invoked and the defigens: Mintla Rufus (ll. 1–2). The interpretation of Mintla is complicated, especially if it is related to Classical Latin mentula, which mean ‘phallus’, or its Vulgar form mentla. Tomlin points to a parallel (Priminus Mintla documented in RIB 631), but continues, ‘[e]xactly how mintla is being used is not

After being inscribed, the tablet was folded five times. Translation (Tomlin 1995: 371): ‘Mintla Rufus to the god Mercury. I have given them, whether woman or [man], (…) the material of a cloak. I have given (them).’

353

Sylloge of  Defixiones from the Roman West 365. Uley Provenance: sanctuary. Current Location: British Museum. Date: 2nd–3rd centuries CE. [Inv. No.: 1978,0102.148] Material: lead. Measurements: 13.1 × 7.6 × c. 0.1 cm. Reading:

5



10



15



deo sanc`to´ Mercurio Honoratus conqueror numini tuo me perdidisse rotas duas et vaccas quattuor et resculas plurimas de hospitiolo meo rogaverim genium numinis `tu{u}i´ ut ei qui mihi fraudem fecerit sanitatem ei non permittas nec iacere nec sedere nec bibere nec manducare si baro si mulier si puer si puella si servus si liber nis{s}i meam rem ad me pertulerit et meam concordiam habuerit iteratis pr{a}ecibus rogo numen tuum ut petitio mea statim pareat me vindicatum esse a maiestate tua

Bibliography: Tomlin 1992: 310–11; AE 1992, 1127; Tomlin 1993a: 129, no. 72; Kropp 2008: no. 3.22/29; Adams 2016: no. 35; Tomlin 2017c: 11.04; Urbanová 2018: no. 303.

non alia minutalia), hospitiolo meo (ll. 4–5, for hospitiolo as a possible diminutive of hospitium, see Adams 2016: 414 and cf. 357 and 443), qui mihi fraudem fecerit (ll. 7–8; analogous phrases in 237, 357, etc.) as well as a non permittas formula (ll. 8–10, cf. 259, 358, 443, etc.) and an all inclusive formula used to describe the victim (ll. 10–12, see 237 and 356; for the Germanic word baro, see Adams 2016: 415). Honoratus addresses Mercury in a very respectful tone and hence uses verbal forms like rogaverim (l. 6), and the phrase genium numinis (ll. 6–7; cf. 452).

Image source: Tomlin 1992: 311, fig. 1. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), was written on rectangular lead sheet. The tablet is intact and in excellent condition, even though it has a thin brownish patina on its surface. The inscription contains 17 lines, which run from left to right and were written in old Roman cursive by a trained hand.

What is perhaps most notable about this curse, however, is what the defigens asks for: he not only wishes to recover his stolen property and punish the thief, but also wants to be reconciled with him (ll. 13–14, …meam concordiam habuerit); this does not find many parallels in the curses from the Roman West (but cf. with 358, ll. 12–14: co(n)scientiam perferat).

Based on its content, this curse should be grouped among the defixiones against thieves. Here, Honoratus denounces to Mercury the theft of rotas duas et vacas quattuor et resculas primulas (ll. 3–4, ‘two wheels and four cows and many small belongings’). The text presents the following orthographic and lexical features: resculas (l. 4, diminutive of res), tuui for tui (l. 7), nissi for nisi (l. 12) and praecibus for precibus (l. 14). This curse contains a good deal of familiar phrases and formulae found elsewhere in the corpus of British curses: et resculas plurimas (ll. 3–4; for a parallel catch all, see 459, l. 2: nec

The text concludes with the extraordinary phrase iteratis pr{a}ecibus rogo numen tuum ut petitio mea statim pareat me vindicatum esse a maiestate tua (ll. 14–17; for comparisons, see 120, 240). Here the practitioner repeats his petitio (cf. petio in 214, l. 1) and asks for immediate divine intervention. Note the use of the verb pareo (l. 16), not as a mistake for pario (as proposed by Tomlin 1992), but rather ‘as a quasilegalism (...) Paret = “is clear, evident” is particularly 354

 Britannia I would ask the genius of your divinity that you do not allow health to the person who has done me wrong, nor allow him to lie or sit or drink or eat, whether he is man or woman, whether boy or girl, whether slave or free, unless he brings my property to me and is reconciled with me. With renewed prayers I ask your divinity that my petition may be immediately fulfilled (and that it become obvious) that I have been avenged by your majesty.’

common in legal formulae and language (see OLD s.v. 6, TLL X.1.373.38ff.)’ (Adams 2016: 416). Honoratus is a well attested Latin cognomen (see OPEL II, 184). After being inscribed, the tablet was folded six times. Translation (Tomlin 1992: 311, modified following Adams 2016: 314–15): ‘Honoratus to the holy god Mercury. I complain to your divinity that I have lost two wheels and four cows and many small belongings from my house.

366. Uley Provenance: sanctuary. Current Location: British Museum. Date: 2nd–3rd centuries CE. Material: lead. Measurements: 7.5 × 7.9 cm. Reading: A [deo] sanc ̣to Mẹrc ̣ụriọ q[ue]rọr tibi de illis qui mihi mạle cogitant et male faciunt supra ED[-c.3-]S iumen[t -c.2-3-] 5 si servus si liber si m[ascel] ṣị [fe]ṃina ut ṇọn illis per mittas nec ṣtare nec sedere nec bibere

B nec manducarẹ nẹc hạ[s] ịrạs redemere possit nessi sanguine suo AENE+ (traces)

Bibliography: Tomlin 1993a: 130, no. 76; Tomlin 1995: 373–76; AE 1995, 985; Kropp 2008: no. 3.22/32; Urbanová 2018: no. 304.

inscription contains 12 lines, which run from left to right and were written in old Roman cursive dated to between the second and third centuries on palaeographical grounds.

Image source: Tomlin 1995: 374–75, figs. 2a and 2b. Courtesy of R.S.O. Tomlin.

The inscription uses the phrase qui mihi mạle cogitant et male faciunt (ll. 3–4, ‘those who are badly disposed towards me (and) who are acting badly’) to curse the defigens’ enemies. This phrase is unparalleled in the corpus of curse tablets, but it would serve as an all-inclusive formula. While mihi male cogitant is unparalleled in curses, the editor has pointed out that it is an idiom in Classical Latin (e.g., Cicero, ad. Fam. VIII, 12, 1: male eum de me cogitare).

Commentary: this defixio, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), was written on a roughly trapezoidal lead tablet. The tablet is in poor condition, since it has been corroded on its surface to the point that there are now two holes in the central portion and many of the edges have been lost as well as several lines of text (A and B l. 4, especially along the fold lines). An opisthograph, the

While this curse takes aim at a broad group of enemies, it nevertheless employs formulae associated with the 355

Sylloge of  Defixiones from the Roman West After being inscribed, the tablet was folded twice.

defixiones in fures, such as an all-inclusive formula (ll. 5–6; cf. 215, 237, 459, etc.) as well as a non permittas structure used to impair the health and well-being of the victims (specifically that they be kept from sitting, standing, eating and drinking; A, ll. 6–8, B, ll. 1; cf. 246, 443, 365). Furthermore, the curse concludes with the phrase nẹc hạ[s] ịrạs redemere possit nessi sanguine suo (B, ll. 1–3). While the first part of the phrase is unparalleled, the nessi sanguine suo is very common (e.g., 270 and 353; for texts that preclude the possibility of escaping punishment, see 493).

Translation (Tomlin 1995: 373): ‘To the holy god Mercury. I complain to you about those who are badly disposed towards me (and) who are acting badly over (? …), whether slave or free, whether male or female. Do not allow them to stand or sit, to drink or eat, or to buy off these provocations (?) unless with their own blood …’

367. Uley Provenance: sanctuary. Current Location: British Museum. Date: 3rd century CE. [Inv. No.: 1978,0102.156] Material: lead. Measurements: 4.2 × 7.2 cm. Reading: A carta qu(a)e Merc ̣urio dọna tur ut manẹc ̣ilis qui per[i]erunt ụltiọnem requirat qui iḷlọṣ unvalavi{i}t ụṭ iḷli saṇgu(in)em [e]t sanita5 tem ṭọlla[t] qui ipsos manicili[o]s tulit [u]ṭ quantoci{c ̣ị}us ilḷi pareat quod deum Mercurium r[o]gamus [-c.2-]++VRA

B Q[-c.1-]OS+NC+V[-c.2-3-]LAT (vacat)

Bibliography: Tomlin 1993a: 130, no. 80; Tomlin 1996: 439–41; AE 1996, 936; Kropp 2008: no. 3.22/36; Adams 2016: no. 34; Urbanová 2018: no. 306.

presents the following orthographic features (all found on side A): que for quae (l. 1), manecilis for manicilius (l. 2), invalaviit for involavit (l. 4), saṇguem for saṇguinem (ll. 4–5), quantoc ̣ịcius for quantocius, (l. 6) and pareat for pariat (l. 7; but see Adams 2016: 411 for a different interpretation). Note the use of the verb pereo instead of the more common perdo (A, 1. 2; cf. 446); tollo is used in the Late-Latin sense of ‘to have stolen’ (A, l. 5; cf. 337) and pario in the sense of ‘to accomplish’ (A, l. 6).

Image source: Tomlin 1996: 440, fig. 1. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), was written on a rectangular lead sheet. The tablet is in good condition, although there is some corrosion on its surface and several damaged areas along the fold lines. An opisthograph, the inscription contains eight lines, which run from left to right and were written in an old Roman cursive that can be dated to the third century CE.

The opening phrase carta qu(a)e Merc ̣urio dọnatur (A, l. 1) is notewhorthy for the use of the lexeme carta to refer to the curse tablet (for parallels, see 96, A, ll. 1–2: quomodo h(a) ec chârta, 441, l. 13: carta s(upra) s(cripta), and perhaps 213, A, l. 6: carta picta perṣc ̣[ripta]). To identify the thief, the curse does not employ the common all-inclusive formula, but rather a series of phrases that repeatedly allude to the theft: ut manẹc ̣ilis qui per[i]erunt (A, l. 2), qui iḷlọṣ unvalavi{i}t (A, ll. 3–4), qui ipsos manicili[o]s tulit (A, l. 5; as Adams has noted, ‘repetitious specification

Based on its content, this curse should be grouped among the defixiones in fures. The theft of a pair of gloves (manicilios; for this term, see Adams 2016: 409–10 and 205 for a parallel) is reported to Mercury. The text 356

 Britannia of an object under discussion (...) is a feature of technical, including legal, Latin, and the redundancy suits the style of curses, which have features both of religious and legal language’ [2016: 410–11]) and q[u]o ṣ[a]nc ̣[t]u[m tol]lat (B, l. 1, if we accept the editor’s reconstruction). The text is structured around a series of requests directed to Mercury that begin with ut. With these petitions, the author of the text asks for divine intervention. After soliciting the god’s vengeance (ụltiọnem requirat, A, l. 3), the defigens asks the god to impair the thief’s mental and physical health (ụṭ iḷli saṇgu(in)em [e]t sanitatem ṭọlla[t], A, ll. 4–5; cf. 246, among others). All of this is placed within a magical time frame that urges that the situation be resolved quickly: [u]ṭ

quantoci{c ̣ị}us ilḷi pareat quod deum Mercurium rogamus (A, ll. 6–7; see Adams 2016: 411 and cf. with 259; for magical time frames, see Sánchez Natalías 2019a). After being inscribed, the tablet was folded five times. Translation (Tomlin 1996: 439): ‘The sheet (of lead) which is given to Mercury, that he exact vengeance for the gloves which have been lost; that he take blood and health from the person who has stolen them; that he provide what we ask the god Mercury […] as quickly as possible for the person who has taken these gloves.’

368. Uley Provenance: sanctuary. Current Location: British Museum. Date: 4th century CE. Material: lead. Measurements: (10.5) × 8.5 × 1.5 cm. Reading:

Petroni{i}us

Bibliography: Tomlin 1993a: 130, no. 86; Tomlin 1993b: 310–11; AE 1993, 1078; Kropp 2008: no. 3.22/39; Urbanová 2018: no. 205.

The circumstances that led to the writing of this curse are unknown. The text consists solely of the victim’s name: Petroni{i}us (for Petronius; see OPEL, III, 135). This error could be due to a lack of attention or confusion with the preceding n on the part of the author. Among the British curses, it is not common to have such a curt and simple text with no information except for the name of the victim, though there are several parallels (e.g., 341 or 442).

Image source: Tomlin 1993b: 311, fig. 1. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), was written on what must have been a rectangular lead sheet. The tablet has lost the top-left and bottom-right corners and its surface is somewhat uneven. It contains one line of text that runs from left to right and was written in capitals.

A hole next to the e indicates that the tablet was pierced with a nail after being inscribed.

369. Uley Provenance: sanctuary. Current Location: British Museum. Date: 75–125 CE. Material: lead. Measurements: 12 × 7.1 cm. Reading:

Mercurio -------res id est lanam 357

Sylloge of  Defixiones from the Roman West Bibliography: Tomlin 1993a: 129, no. 58; Kropp 2008: no. 3.22/24; Urbanová 2018: no. 302.

remain unknown about this tablet. The inscription contains nine lines written in old Roman cursive (the reading of only two lines has been published). The letter forms are quite similar to those found in the tablet from 337 and have been dated to the first or second century CE, which makes it the oldest curse from Uley. The mention of res id est lanam suggests that this is a defixio against thieves.

Commentary: this defixio, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), was written on a lead sheet. Due to the lack of a full editio princeps, there are some basic things that

370. Uley Provenance: sanctuary. Current Location: British Museum. Date: 3rd century CE. Material: lead. Measurements: 6.3 × 14.4 cm. Reading: A tibi commendo [---] qui mihi fraudem fecit de denar(ii)s ills quos [mih]i debeat B seminudi edentuli tremuli podagrici sine cuiusque hominis mis{s}ericordia in fanum et thesaurum poten{ten}tiss[imi] dei Bibliography: Tomlin 1993a: 130, no. 78; Tomlin 1999a: 556 and 558; Tomlin 2002: 169; Kropp 2008: no. 3.22/34; Urbanová 2018: no. 305.

practitioner who wrote the curse is noteworthy: the text has six ways of saying ‘hand over’ (commendo, dono, obdo, offero, destino and deputo), the final four of which are found in a single sequence. Additionally, there are two words of ‘sickness’ (morbus and valitudo). The text also takes aim at the victim’s family who would be rendered ‘half-naked, toothless, tremulous, gouty, beyond human pity’ (Tomlin: 2002: 169; for other ‘hereditary’ curses, see 246, 57, etc.).

Commentary: this defixio, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), was written on a lead sheet. Due to the lack of a full editio princeps, there are some basic things that remain unknown about this tablet (e.g., there is not a fully published reading and we do not know the number of lines). Despite the fragmentary nature of the text, this curse can be grouped among the defixiones in fures. Here, the practitioner targets an individual whom he had lent money, but who has stolen that money instead of repaying the defigens (apparently 100,000 denarii according to Tomlin). As noted by the editor, the lexical richness of the

After being inscribed, the tablet was folded. Translation (Tomlin 2002: 169): ‘…the man who has cheated me of the denarii he owed me. I give, I offer, I destine, I depute one hundred thousand denarii to the god Mercury, that he may bring them to the temple and treasury of the most mighty god.’

371. Uley Provenance: sanctuary. Current Location: British Museum. Date: 4th century CE. Material: lead. Measurements: 3.8 × 4.7 cm. Reading: A deo sancto Ṃercurịọ Cạriṇ[us? tibi ex?]ẹc ̣ ro de furṭo {VO} quod 5 mihi factum est Pri manus ṇẹc ̣ ei per mitt[at -c.3-4-]S Mercurius {ụṣ} nec (traces)

358

 Britannia B nec mas [[++++++++++]] (traces) nẹc solẹṃ nec ̣ lụṇ[am] nec CONIVV++ infantis 5 (traces) neum san(g)uine sụo conpliat veṇdica[tionem?]

Bibliography: Tomlin 1993a: 129, no. 68; Kropp 2008: no. 3.22/26; Tomlin 2015: 398–99; AE 2015, 727; Tomlin 2017a; AE 2017, 874.

is well attested in the corpus of curses from Britannia: the first is laid out in a non permittas formula and seeks to harm the health and physical well-being of the victim (for parallels, see e.g., 352 and 358); here, however, we also find an interesting variant in this common structure: nẹc solẹṃ nec ̣ḷụṇam (B, l. 3). This appears to refer to the types of light that the victim should not see, and presumably should be understood as a rhetorical way of saying ‘do not let him see any light at all’ (during the day or night). Although an entire phrase cannot be reconstructed, the presence of the lexemes infantis and CONIVV++ (for coniugis?) in B, l. 4 suggests that this may be a ‘hereditary’ curse (for an example, see 215). Also note san(g)uine sụo coṇpliat veṇdica[tionem] (B, ll. 6–7), which references the payment of the stolen object with the victim’s blood and is here clearly tied to the idea of vengeance (for parallels, see 249, 299, 337, etc.). Finally, Carinus and Primanus are well attested Latin cognomina (see OPEL II, 37 and III, 158, respectively).

Image source: Tomlin 2015: 399, fig. 16. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), was written on a roughly rectangular lead sheet. While the tablet is mostly intact, it has lost several parts of the top and bottom edges and its surface is quite damaged, especially along the fold lines (side A, ll. 2, 5, 6 and B, ll. 1–2 and 6). Furthermore, there are two holes on level with A, l. 5 and B, ll. 2–3. An opisthograph, the inscription contains 15 lines written in a new Roman cursive that is dated to the fourth century CE on palaeographical grounds. Based on its content, this curse should be grouped among the defixiones against thieves. Here, Carin[us] (or Carin[ianus]) denounces a theft carried out by Primanus. The text presents the following orthographic features: sanuine for sanguine (B, l. 6), conpliat for compleat (B, l. 6) and vendica[tionem] for vindica[tionem] (B, l. 7). The author of the text appears to have been confused about how to properly use the passive voice in the phrase de furṭo quod mihi factum est Primanus (A, ll. 3–5; for a similar formula, see 120, ll. 5–6). Primanus’ punishment

After being inscribed, the tablet was folded twice. Translation (Tomlin 2015: 398): ‘To the holy god Mercury. I (?), Carinus (?), implore you concerning the theft which has been done to me (by) Primanus. And Mercury is neither to permit him (…) nor (…) nor (…) neither sun nor moon, neither (…) of an infant (…) fulfil vengeance with his blood.’

359

Sylloge of  Defixiones from the Roman West 372. Uley Provenance: sanctuary. Current Location: British Museum. Date: 175–275 CE. Material: lead. Measurements: 8.8 × 5.3 cm. Reading: G̣enitus Mercurio [g]enio qui mihi [fru]dum fecerit aṇ5 ụlụṃ involaverit sa[nct(um)?] genium si [-c.2-]+[-c.1-]+[-c.1-]+ ạ[n]cilla si liber si [---] [mu]ḷieris si puer si [pue]ḷḷạ +[---]+permitt (?) eụṃ [-c.2-]++ 10 [---]++ERNAT pudoreṃ[-c.3-] [---]+[-c.2-]IAT pede (traces) [---]ạṃb{l}lat (traces) nec ̣ manducat nec sedit nec magiat ṇ[i-] 15 si ad templum tuum repraese ntaverit OP TIBEIVS

Bibliography: Tomlin 1993a: 128, no. 34; Kropp 2008: no. 3.22/13; Tomlin 2016: 396–97; AE 2016, 978.

After the invocation and identification of Mercury as genio (ll. 2 and 6; surely the author of the text intended to invoke the genius of the god), the thief is referred to with an allinclusive formula (ll. 6–8) consisting of two contrasting pairs. Next, the text has a non permittas structure in which the practitioner appears to use an imperative form instead of the usual prohibitive subjunctive. In this section of the curse, the defigens asks for the vital functions of the victim, such as eating, urinating, defecating and walking (if we accept the editor’s reconstruction ạmb{l}lat [l. 12], which would make good sense after pede in the previous line; see 246 and 259 for parallels). The text concludes with the phrase ṇ[i]si ad templum tuum repraesentaverit (ll. 14–17), which serves as a magical time frame that states that the curse is to remain in effect until the stolen objects are returned to the sanctuary.

Image source: Tomlin 2016: 397, fig. 11. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), was written on an irregularly shaped lead sheet. Though intact, the tablet has been severely corroded along the lateral sides and bottom edge, which has resulted in the loss of the text in ll. 3–12. The inscription contains 18 lines written in old Roman cursive that has been dated between the second and third centuries CE. Despite the fragmentary nature of the text, this curse can be grouped among the defixiones against thieves. Here, Genitus (a cognomen first attested here, unless the reading should be C̣unitus(?)) denounces to Mercury the theft of a ring. The text presents the following orthographic and phonetic features: sedit for sedeat (l. 14) and magiat for meiat (l. 14; there seems to be a confusion between the synonyms mingo and meio). Also note the word frudum (l. 4), which seems to be the result of confusion between fraudem and furtum (for a parallel, see 456).

After being inscribed, the tablet was folded four times. Translation (Tomlin 2016: 396, modified): ‘Genitus (or: Cunitus?) to Mercury the genius. (He) who has done me (?) wrong, has stolen (my (?)) ring (…) holy genius that whether (free woman (?)) or slave-girl, whether free man (…) of a woman, whether boy or girl (…) you do not permit him (…) shame (…) with his foot (…) nor to eat nor to sit (at stool (?)) nor to (?) urinate, unless he pay (…) at your temple.’

360

 Britannia 373. Uley Provenance: sanctuary. Current Location: British Museum. Date: 2nd–3rd centuries CE. Material: lead. Measurements: 9.2 × 12.5 cm. Reading: A devo Marti [---]lirus qu{a}eritur si ser(v)us si liber qui vas apium invalavit [v]ere si COMODIA erat ne illi permittatur 5 nec bib[er]e nec mandure nec {nec} somnum nec sanitate nesi ipsum vas ad locum suum reversetur et congortiam Mercuri agat SEPET deum V++ illi qui [---] 10 feci ut illi s me pariat qui +++

B [d]evo Mar ti VS Propitio

Bibliography: Tomlin 1993a: 128, no. 24; Kropp 2008: no. 3.22/10; Tomlin 2017b: 462–64; AE 2017, 875.

Tegullirus; see AE 1987, 767b) denounces to Mars and Mercury the theft of a vas apium (A, l. 3), that is, a beehive. The text presents the following orthographic and phonetic features: quaeritur for queritur (A, l. 2), serus for servus (A, l. 2), invalavit for involavit (A, l. 3), nesi for nisi (A, l. 6) and congortiam for concordiam (A, l. 8; with the change from c to g and d to t as well as the first o written as a small loop; cf. the o in somnum, A, l. 6). The word devo (B, l. 1 and A, l. 1) is noteworthy, since it could be borrowed from Celtic deivo meaning ‘god’ (for a parallel, see 205). The phrase vas apium (A, l. 3), which literally means ‘container of bees’, is noteworthy since this is the first (and currently only) attestation of apiculture in Roman Britannia. COMODIA (B, l. 4) has defied easy interpretation and Tomlin has taken it as ‘a bungled clause to the effect of quocumque modo erat’ (2017: 462; similar confusions with quomodo are common (e.g., 206). Finally, the verb reverto (A, l. 7) is common in Late Latin.

Image source: Tomlin 2017b: 463, figs. 9 and 10. Courtesy of R.S.O. Tomlin. Commentary: this defixio, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), was written on an irregularly shaped lead tablet. The curse has been broken into two pieces that fit together and has suffered damage along the fold lines. Furthermore, the top, bottom and right edges have been damaged. An opisthograph, the inscription contains two lines on side A and one line of capitals followed by 10 of old Roman cursive on side B. A trained but careless hand wrote all the lines. Based on its content, this curse should be grouped among the defixiones in fures. Here, [---]lirus (perhaps

361

Sylloge of  Defixiones from the Roman West After inscribing the tablet, the practitioner folded it twice. When this was done, (s)he wrote the last two lines of the text, can be taken as side the tablet’s main face (B, ll. 1–2), like addresses in archaic and classical Greek letters (cf. Sarri 2017: 87; for a parallel, see 529). Finally, Mars is evoked once again (here with the epithet propitius at B, l. 2) and there is an abbreviation that Tomlin has interpreted as v(otum) s(olvit). Such a phrase would be unparalleled in the corpus of Latin curses.

The text, presented to Mars as a complaint (qu{a}eritur, A, l. 2; for a parallel, see 355), is structured around a series of formulae that are well attested in British curses. It begins with a short all-inclusive formula to describe the thief (A, l. 2), which is followed by a non permittas formula used to harm the vital functions and well-being of the victim (A, ll. 4–6). The curse is to remain in vigour until vas ad locum suum reversetur (A, l. 7, ‘unless the said hive be returned to its proper place’). This detail is noteworthy since normally defigentes ask that stolen property be brought to the sanctuary in which the curse was also deposited (see 237, 269, etc.). An unusual clause adds how to bring the curse’s effect to an end: the victim must re-establish cogortiam (i.e., concordia) with Mercury, the patron god of the sanctuary, who is here invoked instead of Mars (for the pairing of these two gods, see 357).

Translation (Tomlin 2017: 462, modified): ‘To the god Mars the Propitious (…) complains to the god Mars: whether slave or free, (he) who has stolen (my) beehive (…) if it was (…) let him not be permitted to drink or eat, nor (to have) sleep or health, unless the said hive be returned to its proper place and (he) gain the goodwill of Mercury (…) the god (…) to him who did (this) that to him if he furnishes me (…).’

374–88. Uley Provenance: sanctuary. Current Location: British Museum. Date: unknown. Material: lead. Measurements: vid. commentary. Reading:

Vid. commentary.

Bibliography: Tomlin 1993a: 127–30, nos. 7, 19, 29–31, 38, 44, 56, 59–61, 66, 75 and 83.

377 (= no. 30): 4.9 cm wide. This strip of lead broken into two parts that fit together conserves the remains of five lines of capitals that are almost illegible.

Commentary: this collection of curses, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), has been briefly published by Tomlin, who offers ‘a summary catalogue (…) of all the other inscribed tablets of Uley, which have now been examined, but whose text has not been fully deciphered and studied’ (1993a: 127). Accordingly, given the dearth of published information, here I present a joint file for these curses. This file gives each item a number in the present catalogue, the reference to Tomlin’s publication (with further bibliography if it exists) in parentheses, followed by the tablet’s dimension and a brief commentary. Here the tablets are organized palaeographically and the present entry deals with those written in capitals.

378 (= no. 31): (3) × (6.5) cm. This lead fragment is very corroded and has the traces of writing (perhaps an e). 379 (= no. 38): (4) × (7) cm. These four fragments of lead are very corroded and probably fit together. They conserve the remains of very irregularly sized capitals. 380 (= no. 44): (5) × 4.8 cm. Three lead fragments that are very deteriorated and appear to have capitals. 381 (= no. 56): 7.3 × 5.7 cm. This tablet is very corroded and conserves the remains of eight lines of capitals that are illegible except on the left side. 382 (= no. 59): 4.4 × 5.7 cm. This opistographic oval tablet is inscribed with capitals: side A reads Q, while side B conserves four lines of capitals that follow the curse’s curve (perhaps it was originally written in a concentric circle?). According to Tomlin, the text is not Latin. Perhaps it is Celtic? (See 219, 223, 407 and 453).

374 (= no. 7): 11.5 × 5.8 cm. An opisthographic lead sheet with 16 lines on side A and two on side B. 375 (= no. 19): 8 × 5.4 cm. This tablet preserves 10 lines of capitals and is very corroded. 376 (= no. 29): (7) × (15.5) cm. Three fragments of a lead sheet that appear to fit together and have the remains of some capital letters.

383 (= no. 60): (7.4) × (8.3) cm. Five lead fragments, the biggest of which has three lines of text written in capitals (ll. 1 and 3 are illegible). The four remaining fragments (one with some text) do not fit together.

362

 Britannia 384 (= no. 61): 9.4 × 8.3 cm. The tablet conserves eight lines of capitals that are extremely corroded.

money (d[e] arca) and grain (frumenta). The defigens asks that the stolen goods be returned to the temple and uses an abbreviated all-inclusive formula sbsmspsp for si baro si mulier si puer si puella (see 351 for another abbreviation).

385 (= no. 66): 5.6 × 6.2 cm. This curse conserves five lines of capitals that are extremely irregular in size. According to the editor, this is ‘perhaps a pseudo-inscription’ (if so, see 317–21, 440).

387 (= no. 83; Kropp 2008: no. 3.22/37): 11.1 × 5.4. This strip of lead was discovered on the surface of the ground. It’s very corroded and conserves the remains of 12 lines of capitals which include the verb [in]volavit.

386 (= no. 75; Kropp 2008: no. 3.22/31): 5.9 × 11 cm. This tablet, though intact, has been greatly damaged by corrosion and the folds. It conserves the remains of 10 lines of capitals and appears to be a defixio in fures. The deity is offered the stolen property, which appears to be

388 (= no. 85): 7.7 × 6.9 cm. This lead tablet is inscribed with 10 lines of capitals (perhaps a list of personal names?). The curse is intact but very poorly conserved.

389–432. Uley Provenance: sanctuary. Current Location: British Museum. Date: 175–275 CE. Material: lead. Measurements: vid. commentary. Reading:

Vid. commentary.

Bibliography: Tomlin 1993a: 127–30, nos. 8, 9, 11–18, 20–25, 27, 28, 32, 34–36, 39–42, 45–48, 54, 57, 62–64, 67, 69–71, 73, 74, 79, 81, 82, 84 and 87.

393 (= no. 13): 4.9 × 6.7 cm. This tablet is very corroded and contains some six lines of text. 394 (= no. 14): 6.2 × 18.5 cm. This lead strip has been very corroded and conserves six lines of text.

Commentary: this collection of curses, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), has been briefly published by Tomlin, who offers ‘a summary catalogue (…) given below of all the other inscribed tablets of Uley, which have now been examined, but whose text has not been fully deciphered and studied’ (1993a: 127). Accordingly, given the dearth of published information, here I present a joint file of the curses. This file gives each item a number in the present catalogue, the reference to Tomlin’s publication (with further bibliography if it exists) in parentheses, followed by the tablet’s dimension and a brief commentary. Here the texts are organized palaeographically and the present entry deals with those written in old Roman cursive.

395 (= no. 15): (3.8) × (3.7) cm. This lead fragment is very corroded and belongs to what was the top-left corner of a larger tablet. 396 (= no. 16): 3 × 13 cm. This rectangular lead sheet has one rounded edge (the other edge is lost). It contains a oneline inscription. It has not been definitively determined what kind of text this is (Tomlin has wondered whether it is actually a label). 397 (= no. 17): 8.2 × 10.3 cm. This lead sheet is inscribed but so corroded that it cannot be deciphered.

389 (= no. 8): 16 × 12 cm. This tablet is broken in six fragments that are extremely corroded.

398 (= no. 18): (5.2) × (4.6) cm. This lead fragment originally belonged to a rectangular tablet. An opisthograph, the inscription contains 14 lines of text that are now illegible due to corrosion.

390 (= no. 9; Kropp 2008: no. 3.22/7): 7.9 × 9.2 cm. This opisthographic tablet contains 12 lines of text. Side A reads Divo Mercurio and appears to have a reference to a stolen sheep (ovem).

399 (= no. 20; Kropp 2008: no. 3.22/9; Urbanová 2018: no. 206): 13 × 9 cm. This opisthographic tablet is so corroded that side A is illegible. Side B, in contrast, begins with the Celtic personal name Cunovinna and references a lost object.

391 (= no. 11): These 11 fragments, five of which fit together, are still rolled up. Some fragments have text. 392 (= no. 12; Kropp 2008: no. 3.22/8): (12) × (7.5) cm. These three lead fragments fit together. The inscription, which is very deteriorated, preserves the formula si puer si puella. After being inscribed, the tablet was folded and pierced with a nail.

400 (= no. 21): 8.1 × 8.3 cm. This opisthographic tablet has a symbol on side A (cf. 259, 448), while side B conserves 12 lines of text, among which a symbol for denarii and a number can be read.

363

Sylloge of  Defixiones from the Roman West 401 (= no. 22): 9 × 9 cm. This tablet is very corroded and converses the remains of six lines of text.

415 (= no. 47; Kropp 2008: no. 3.22/17): 5.1 × 8 cm. This strip of lead is broken into two fragments that fit together. An opisthograph, the inscription contains 12 lines among which the word conqu[e]ror can be read (A, l. 2; see 259 and 264 for parallels).

402 (= no. 23): These 10 fragments belong to an opisthographic tablet (the largest piece measures (4.5) × c. (7) cm) and are now very deteriorated. Only several letters can be deciphered.

416 (= no. 48): 4 × 7 cm. This tablet is broken into two pieces that are in poor condition. An opisthograph, the inscription contains 13 lines of nearly illegible text.

403 (= no. 25): 6.2 × 10 cm. This lead strip is broken into four pieces that fit together. The curse is still rolled up and very corroded. It appears to be completely illegible.

417 (= no. 54; Kropp 2008: no. 3.22/21): (3) × (3) cm. This lead fragment preserves the remains of four lines, among which invola- (l. 2) can be made out.

404 (= no. 27): (4.1) × (3.4) cm. This is an extremely corroded lead fragment that conserves six lines of text.

418 (= no. 57; Kropp 2008: no. 3.22/23): (5.4) × (7.1) cm. This lower-left corner of a tablet conserves six lines of text that end with the verb habeat. It comes form an archaeological context dated to the second century CE.

405 (= no. 28; Kropp 2008: no. 3.22/11): 16.6 × 7.7 cm. This lead strip is very corroded and contains a long inscription that today is nearly illegible. The text begins with the invocation d[eo] Mercurio.

419 (= no. 62; Kropp 2008: no. 3.22/25): 12.8 × 8.9 cm. This lead sheet comes form an archaeological context dated to the second century CE. An opisthograph, the inscription contains 18 lines of text (one found on side A and 17 on side B). The curse has been corroded. Nevertheless, we can make out that Senovarus, son of Senovirus, asks Mercury to help recover various stolen objects, including a pallium.

406 (= no. 32): (3.9) × (6.2) cm. This lead fragment belonged to the left side of a curse tablet. It conserves the remains of five lines of text that are almost illegible. 407 (= no. 35): 10.9 × 5.9 cm. This lead sheet is inscribed with 16 lines of text and is in good condition. According to Tomlin, the inscription is not in Latin. Perhaps this is another Celtic defixio (see 219, 223, 382 and 453).

420 (= no. 63). These four lead fragments conserve the remains of a text in which only several letters are legible.

408 (= no. 36): 5 × 6.6 cm. This opisthographic tablet conserves six hardly legible lines on side A, on top of which another nine lines running diagonally from right to left have been written. Side B contains six lines that are very deteriorated and practically illegible.

421 (= no. 64): 3.4 × 9.1 cm. This long strip of lead is opisthographic and inscribed by two different hands. Side A conserves two or three lines, while side B contains two lines found along the bottom edge of the tablet.

409 (= no. 39): 10.1 × 9.2 cm. This opisthographic tablet conserves 19 lines, two of which were written upside down on the top part of side B.

422 (= no. 67): 3 × 9.1 cm. This long strip of lead conserves two lines found at the end of the tablet.

410 (= no. 40; Kropp 2008: no. 3.22/14): 11.1 × 5.2 cm. This corroded tablet contains 15 lines, which are shallowly inscribed. We can read the invocation to deo Me[r]{p23} curio.

423 (= no. 69): 11.2 × 7.2 cm. This irregularly shaped lead sheet has the remains of several letters. 424 (= no. 70; Kropp 2008: no. 3.22/27): 15.9 × 5.8 cm. This opisthographic lead strip conserves the remains of 11 lines. Side A, nearly illegible, has at least nine lines of old Roman cursive and four lines of capitals found in the right that read dovo Mercurio (dovo for devo, lexeme which is borrowed from Celtic deivo meaning ‘god’, see 205 for parallels). On side B there are two lines of capitals among which Severino / dona can be made out.

411 (= no. 41; Kropp 2008: no. 3.22/15): 6.1 × 8.9 cm. This corroded opisthographic tablet conserves 10 lines. On side A we can read si quis. 412 (= no. 42). Several fragments make up the remains of an opisthographic tablet, which has been severely damaged and cracked and whose text is practically illegible. 413 (= no. 45): 4.8 × 8.4 cm. This deteriorated tablet conserves the remains of six lines. The endings –us and –a suggest that the text consisted of a list of personal names.

425 (= no. 71; Kropp 2008: no. 3.22/28): 4.6 × 5.6 cm. This opisthographic lead sheet has an irregular shape. It contains a text that is hardly visible and begins with the invocation [deo?] Mercurio.

414 (= no. 46): (3.3) × (6.1) cm. This corroded tablet is broken into four pieces that fit together. They conserve an inscription of six lines of text that is hardly legible.

426 (= no. 73): (1.8) × (1.5) cm. This lead fragment conserves the remains of two lines.

364

 Britannia 427 (= no. 74; Kropp 2008: no. 3.22/30): (5.3) × (8) cm. This fragment of lead belonged to the upper edge of a tablet and has been severely corroded. It conserves the remains of six lines that begin with the invocation deo Merc[u]ri[o].

430 (= no. 82). These six lead fragments, two of which fit together, conserve the remains of an inscription.

428 (= no. 79; Kropp 2008: no. 3.22/35): 8.4 × 9 cm. This lead tablet is very deteriorated and hardly legible. It contains nine lines that begin with the invocation deo Mercurio.

431 (= no.84; Kropp 2008: no. 3.22/38): 13.4 × 8.4 cm. This opisthographic lead tablet has 10 lines of capitals on side A, in which the theft of two pewter plates is reported to Mars. Side B contains 12 lines of old Roman cursive, in which the defigens asks the deity to punish the wrongdoer and offers him half of the value of the stolen items in exchange for their return.

429 (= no. 81): 4.2 × 8.8 cm. This opisthographic lead sheet is very corroded and conserves 18 lines, in which only several letters can be deciphered.

432 (= no. 87): 8.2 × 8.7 cm. This lead sheet is extremely corroded and conserves around six lines that are hardly legible.

433–39. Uley Provenance: sanctuary. Current Location: British Museum. Date: 275–400 CE. Material: lead. Measurements: vid. commentary. Reading:

Vid. commentary.

Bibliography: Tomlin 1993a: 128–30, no. 26, 37, 51–53, 65, 68 and 77.

read: DEVENDI. The editor has taken this as the ending of a Celtic personal name.

Commentary: this collection of curses, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), has been briefly published by Tomlin, who offers ‘a summary catalogue (…) given below of all the other inscribed tablets of Uley, which have now been examined, but whose text has not been fully deciphered and studied’ (1993a: 127). Accordingly, given the dearth of published information, here I present a joint file of the curses. This file gives each item a number in the present catalogue, the reference to Tomlin’s publication (with further bibliography if it exists) in parentheses, followed by the tablet’s dimension and a brief commentary. Here the texts are organized palaeographically and the present entry deals with those written in new Roman cursive.

436 (= no. 52): 8.6 × 9.5 cm. This tablet, which has been greatly damaged by corrosion, conserves 12 lines. According to Tomlin (2002: 175), the tablet was inscribed in Latin but using the Greek alphabet (this would be the only case of the Greek alphabet being used in a British curse). There are, however, North African parallels for this practice: DT 267, 269, 270 and 304. Translation (Tomlin 2002: 175): ‘I have given the man who stole my linen and my cloak and my two silver coins, whether boy or girl, whether male slave or female, whether man or woman, whether soldier or civilian. Take away his marrow, his blood, his soul, unless he brings them back to your temple.’ 437 (= no. 53; Kropp 2008: no. 3.22/20). These are two fragments of a lead sheet, the first of which is not inscribed and measures (5.5) × (8) cm. The second fragment is broken into three parts that fit together and measures (5.2) × (15) cm. It conserves the remains of an inscription in which si femina can be made out. If this is, as it seems, part of an all-inclusive formula to assist in identifying the wrongdoer, this text should be classified as a curse against thieves.

433 (= no. 26): 8.4 × 9.7 cm. This opisthographic tablet contains the remains of random strokes and perhaps a d on side A. Side B contains seven lines, which were written in a mix of capitals and new Roman cursive. The curse’s surface is quite damaged, since the tablet was apparently stabbed various times with a pointy object. 434 (= no.37): 4 × 7 cm. This opisthographic tablet is broken into four fragments that fit together. After being inscribed, the tablet was hammered on side A, which has led to such extensive damage that the text is now illegible.

438 (= no. 65): (5.5) × (3.2) cm. This lead fragment originally belonged to the left side of a larger tablet. It conserves the remains of five lines. 439 (= no.77; Kropp 2008: no. 3.22/33): 5 × 5.8 cm. This opistographic tablet is so corroded that the text from side A is illegible, while side B conserves the remains of three lines that read quos a.

435 (= no. 51): 16.5 × 7.5 cm. This irregularly shaped lead tablet has been severely corroded. It conserves the remains of about seven lines, of which only the final line can be 365

Sylloge of  Defixiones from the Roman West 440. Uley Provenance: sanctuary. Current Location: British Museum. Date: unknown. Material: lead. Measurements: 6 × 9 cm. Reading:

Vid. commentary.

Bibliography: Tomlin 1993a: 127, no. 10.

read and was inscribed on a fragment of lead of uneven surface. According to Tomlin, the tablet was ‘[i]nscribed in strange angular letters, perhaps a pseudo-inscription’ (1993a: 127), comparable to 382 and probably some tablets found in the sanctuary of Sulis Minerva in Bath (see 317–21).

Commentary: this curse, which was discovered in the sanctuary of Mercury in Uley (for the archaeological context, see 355), was published among those that Tomlin treated in a rather cursory fashion. The text is difficult to

441. Venta Icenorum, Caistor St Edmund Provenance: riverbank. Current Location: private collection. Date: unknown. Material: lead. Measurements: 6.5 × 10.5 cm. Reading:

a NASE[-c.1-]+[-c.1-] eve(h)it+ VROC+ SIV+ fascia(m) et armi[lla-] s cap(t)olare ṣpectr[um(?)] 5 cufia(m) duas ọc ̣rias x vas a stagnea si mascel si me mina si puer si pulla duas ocri˹as˺ si vulleris fạc ̣tae sangu(ine) suo ut (i)llu(m) requerat {at} Nepṭus 10 e(t) amic ̣ṭus e(t) cufia (et) arm(i)lla(e) + (traces) (denarii) xv cape(t)olare tunc ̣ sanguin(e) (traces) (vacat) fasciaṃ ṭeṇet fure carta s(upra) s(cripta) (vacat) ratio

Bibliography: Hassall 1982: 408–09; AE 1982, 669; Kropp 2008: no. 3.7/1; Urbanová 2018: no. 284.

for femina (ll. 6–7), pulla for puella (l. 7), ocrisa for ocrias, vulleris for volueris (both in l. 8) and e for et (l. 10). The lexeme capeolare (for capetolare, ll. 4 and 11) is related to capitulum, as we learn from Isidore (Orig. XIX, 31: quod vulgo capitulare dicunt), and refers to a head garment worn by women. Ṣpectrum (l. 4) should be taken as a synonym for speculum, even though this usage is not otherwise attested until the Middle Ages. Finally, cufia(m) (l. 5) is derived from cofia.

Image source: Hassall 1982: 409, fig. 34. Courtesy of M.W.C. Hassall. Commentary: this defixio was discovered in 1981 on the ground during a walk along the banks of the Tas River nearby the western part of the Roman city. The man who happened upon the tablet, Mr T. Sapwell, now has the curse in his private possession. It was found still rolled up alongside a tear-shaped lead label (5 × 2.2 cm) which was pierced with a nail. The inscription contains 13 lines, which run from left to right and were written in capitals.

Even though the defigens knew the name of the thief (cf. l. 2), (s)he nevertheless uses an all-inclusive formula to identify him: si mascel, si memina (for femina); si puer, si pulla (ll. 6–7), which is very well attested among British curses (e.g., 305 and 451) The practitioner calls upon Neptune for justice, offering him the purloined objects so that he will ‘charge’ the culprit their value in blood (ll. 8–9, see 249, 271, 304 and 366 for a few parallels). Finally, note the final formula carta s(upra) s(cripta) (l. 13), where we find the abbreviation s s for super scripta (for parallels, see 349 and 356) and also have an example of the curse being referred to as a carta (cf. 96 and 367).

Based on its content, this curse should be grouped among the defixiones against thieves. Here, the theft of a series of objects and articles of clothing is denounced. The text contains numerous errors such as the omission of letters and the repetition of others. These mistakes include: eveit for evehit (l. 2), fascia for fasciam (l. 3), capolare for capetolare (ll. 4 and 11), cufia for cufiam (l. 5), memina 366

 Britannia The names in this curse are themselves notable: There is neither a parallel for the defigens (Nase…, l. 1; cf. Nasennianus) nor for that of the thief (Vroc…, l. 2; probably derived from the root *brocco, ‘badger’).

headdress, a pair of leggings, ten pewter vessels, whether he be man or woman, boy or girl. If you (Neptune) want (literally: shall have wished for) the pair of leggings, they shall become yours at the price of his blood, so that he, Neptune, shall seek him out, and a cloak and head-dress and bracelets, fifteen denarii, the cap. Then the thief holds onto the wreath at the cost of his blood in accordance with the transaction on the above written sheet.’

After being inscribed, the tablet was rolled up. Translation (Hassall 1982: 408): ‘Vroc…sius (?) carries off from Nase… a wreath, bracelets, a cap, a mirror (?), a

442. Thetford Provenance: unknown. Current Location: unknown. Date: 4th century CE. Material: lead. Measurements: 4.6 × 5.3 × 0.5 cm. Reading:

OVA Peminius Noṿalis [--- defix]ụs est Pem[inius?]

Bibliography: Hassall 1982: 410; Kropp 2008: no. 3.21/1; Urbanová 2018: no. 202.

items of gold (among them an amulet) and 33 of silver, all of which were made by the same metal smith and date to the fourth century CE. The curse was written of an irregularly shaped lead sheet, whose text contains three lines of capitals, which run from right to left. The circumstances that led to the writing of this curse directed against Peminius Novalis are unknown. Peminius is unparalleled, while Novalis is an attested Latin cognomen (see OPEL III, 105).

Image source: Hassall 1982: 410, fig. 35. Courtesy of M.W.C. Hassall. Commentary: this defixio was discovered in 1979 ‘in association with’ the discovery of a treasure made up of 39

443. Pagans Hill (Chew Stoke, Somerset) Provenance: unknown. Current Location: unknown. Date: first half of the 3rd century CE. Material: lead. Measurements: (10.1) × 9.5 cm. Reading:

5



10



++[-c.17-]RI[-c.8-] [-c.14-]ṃịṭr[-c.1-2-]pio[-c.5-] ịṇ (denari)is (tribus) milibus cuiuṣ [de]mediam partem tibi ut ita illuṃ [e]xigaṣ a Ṿạssicillo [-c.1-]pẹc ̣omini filio et uxore sua quoniam [-c.2-]rtussum quod illi dẹ hospitiolọ ṃ[eo] [pec]uḷaverint nec illis [p]ermittaṣ sanit[a-] [tem] nec bibere nec ma[n]d[u]care nec dormi[re] [nec nat]ọṣ sanos habẹ[a]ṇṭ nessi hanc rem [meam] ad fanum tuum [at]tulerint · iteratis [pre]c ̣[i]bụṣ te rogo ut [ab ip]sis nominibus [inimicorum] meorụṃ học ̣ [pertu]ṣṣụṃ recipi [---] pẹrven[ia]t

367

Sylloge of  Defixiones from the Roman West Bibliography: Tomlin 1984: 336, 339 and 351–52; AE 1984, 623; Tab. Sulis p. 61, no. 9; Hassall and Tomlin 1991: 309; Kropp 2008: no. 3.18/1; Tomlin 2017b: 462; Urbanová 2018: no. 291; Sánchez Natalías 2022.

stolen object’s value to the deity (ll. 3–4: [de]mediam partem; cf. 205) and the non permittas formula (ll. 7–9) with which the defigens seeks to harm the thief’s vital functions and impede his well-being. The curse’s punishment, which is also extended to the culprit’s family (see 215 for a parallel), is not to be lifted until the stolen goods are placed in the temple of the deity invoked (ll. 9–10). Finally, note the phrase hospitiolọ ṃ[eo] (l. 6; cf. 217 and 357), which the editors have taken as a synonym for ‘house’.

Image source: Tomlin 1984: 339, fig. 36. Courtesy of R.S.O. Tomlin. Commentary: a metal detectorist discovered this defixio as well as the following two curses in 1983. The antiquities firm Fox and Co. in Yeovil later acquired them. The tablet, originally rectangular in all probability, has lost its two top corners, the lowerleft corner as well as the right edge. The surface has been largely corroded and dented, especially the upper and central portions, where the tablet was folded. The inscription contains 13 lines, which run from left to right and were written in an old Roman cursive that can be dated to the first half of the third century CE on palaeographic grounds.

The last lines of the inscription (ll. 10–13) have been severely damaged but appear to contain a final invocation of the deity to ask for restitution. The text has been restored as [at]ulerint iteratis [pre]c ̣[i]bụṣ (ll. 10–11), on the basis of a literary parallel (Cyprian, De mortalitate 18: quid precibus frequenter iteratis rogamus). Also note the phrase [ab ip]sis nominibus [inimicorum] meorum (ll. 11–12; for a parallel, see 468), which alludes to the victims, even though their names are known. By reiterating their role in the crime, the defigens seeks to guarantee that the deity exacts the necessary punishment. The editor has conjectured that the deity invoked is Mercury and proposed the reconstruction [deo sancto Mercu]ri[o] (l. 1). Such a reconstruction is highly speculative.

Based on its content, this curse should be grouped among the defixiones in fures. Here, Vassicillius, son of […]comius (ll. 4–5) and his wife are accused of stealing three thousand denarii (for a parallel amount, see 455). The identification of the thief is uncommon especially with the addition of a patronymic (see 305, 355 and 446). The text presents the following orthographic features: [de]mediam for [de]midiam (l. 3) and nessi for nisi (l. 9). As Hassall and Tomlin have pointed out (1991: 309), the verb peculor (l. 7) is deponent, though here it is conjugated with active morphology (for a parallel, see adsellare in 246, l. 5). Note the use of ma[n]d[u]care as a ‘Vulgar’ synonym for edere (l. 8, for parallels, see 246, 365 and 366) as well as the abbreviation for denarius (l. 3; cf. 349). Tomlin has proposed the reconstruction [pe]rtussum (ll. 6 and 14), which is ‘a variant spelling of percussum, conjecturally derived from the supine of percutio in its sense (Oxford Latin Dictionary s. v. 5) “to strike, stamp (coinage)”. Percussum, in the sense of “stamped (coinage)”, which is unattested, would at least make sense in this context’ (1984: 352).

The name Vassicillus (diminutive from vassos, from which Vassius and Vassillus are also derived) is first attested here as is Pecominus. After being inscribed, the tablet was folded. Translation (Tomlin 1984: 336, modified after Hassall and Tomlin 1991: 309): ‘… in three thousand denarii, of which (I give) you half part on condition that you exact it from Vassicillus the son of (…)comius and from his wife, since the coin(?) which they have stolen from my house and you are not to permit them health nor to drink nor to eat nor to sleep (nor) to have healthy (children) unless they bring this (my) property to your temple. With repeated (prayers) I ask you that this (coin?) may come to be recovered (from the very) names of my (enemies).’

This tablet includes some of the most common formulae found in British curses: the dedication of part of the

368

 Britannia 444. Pagans Hill (Chew Stoke, Somerset) Provenance: unknown. Current Location: unknown. Date: unknown. Material: lead. Measurements: (7.5) × (4.6) cm. Reading:

5



10



[---]++ GNO queṃ [---]tụạ DRODIT [---]+[---] [------] [------] [---]T[---] [---]Q[---] [------] octies novem E sit omni gen[ere la-] borum fatigatu[s] EEXORIT[-c.1-]S++++ O++[---]

Bibliography: Tomlin 1984: 340 and 352; Tab. Sulis p. 61, no. 10; Tomlin 2002: 172; Kropp 2008: no. 3.18/2.

this curse. We can read the phrase octies novem (l. 10), which may form part of a magical time frame (see Sánchez Natalías 2019a for a general discussion). Next, and on analogy with a passage from Livy (XL, 22, 15) the editor has proposed the reconstruction of the phrase sit omni gen[ere la]borum fatigatu[s] (ll. 11–13), which would be completely unparalleled in the context of the curse tablets from Britannia. Still, it seems like an appropriate punishment for the victim of the curse.

Image source: Tomlin 1984: 340, fig. 37. Courtesy of R.S.O. Tomlin. Commentary: a metal detectorist discovered this defixio as well as 443 and 445 in 1983. The antiquities firm Fox and Co. in Yeovil later acquired them. It was written on a sheet that is currently broken into three pieces, two of which are so corroded that they cannot be read. The inscription contains the remains of 14 lines, which run from left to right and were written in capitals.

After being inscribed, the tablet was folded. Translation (Tomlin 1984: 336): ‘… eight times nine… let him be wearied with every kind of hardship…’

The fragmentary nature of the text prevents us from determining the motive that provoked the defigens to write

445. Pagans Hill (Chew Stoke, Somerset) Provenance: unknown. Current Location: unknown. Date: unknown. Material: lead. Measurements: (4.9) × (4.3) cm. Reading:

5

[---]COND+TIN+[---] [--- (?)quic]umqu[e] quomin[us ---] [---]frau[d]e sua ul[la---] [---]VS donav[i or it?---] [--- (?)d]eus[---]

369

Sylloge of  Defixiones from the Roman West Bibliography: Tomlin 1984: 341 and 352–53; Tab. Sulis p. 61, no. 11; Kropp 2008: no. 3.18/3.

in capitals. The circumstances that led to the writing of this curse are unknown, though the mention of frau[d]e (l. 3) is very suggestive. Despite the fragmentary nature of the text, Tomlin has proposed several reconstructions such as [qui]bus, [nomini]bus and a personal name following donav[i or -it] (l. 4). For the last line, he proposes [r]eus, [m]eus or more likely [d]eus, which would be part of a final invocation of the deity meant to help favour divine intervention.

Image source: Tomlin 1984: 341, fig. 38. Courtesy of R.S.O. Tomlin. Commentary: a metal detectorist discovered this defixio as well as 443 and 444 in 1983. The antiquities firm Fox and Co. in Yeovil later acquired them. The fragment is broken into two pieces that fit together and contain the remains of five lines, which run from left to right and were written

After being inscribed, the tablet was folded.

446. Eccles Provenance: domestic context. Current Location: unkown. Date: 4th century CE. Material: lead. Measurements: 7.1 × 8.9 cm. Reading: A SSSTI

B donatio diebus quo perit Butu resque qu(a)e +VM+ nec ante sa ṇetate ṇẹc ṣạlute 5 ṇesị qua(m) ịn doṃo dei ++C ṣanetatẹ ịn dọ [mo dei (?)]SAN++AN[-c.3-]P[-c.1-2-] [-c.1-2-]+++ sum +[---]

Bibliography: Detsicas 1971: 29; Tomlin 1985; Tomlin 1986: 428–31; AE 1986, 463; Tab. Sulis p. 61, no. 4; Tomlin 1996: 457; Kropp 2008: 3.1/1; Urbanová 2018: no. 241.

which were written in new Roman cursive that can be dated to the fourth century CE on palaeographical grounds. On side A, the text runs from left to right, while on side B it follows the boustrophedon pattern (for the only parallel for this from Britannia, see 304).

Image source: Tomlin 1985: 20–21, figs. 1–2. Courtesy of R.S.O. Tomlin.

Given the presence of certain common formulae, this curse should be grouped among the defixiones in fures. The text presents the following orthographic features: diebus for diibus (l. 1), que for quae (l. 2), sanetate for sanitate (ll. 3–4) and nesi for nisi (l. 5). Side A only has one line of text that reads SSSTI with a horizontal stroke running through the SSS. At first, this symbol was taken as an abbreviation for s(upra) s(cripto), but here the third S would not bear any meaning. Accordingly, this hypothesis has been rejected.

Commentary: this defixio was discovered in 1970 among the rubble covering the floor of room 121 of the Roman Villa in Eccles (for the archaeological context, see Detsicas 1971). The curse was written on a rectangular lead sheet, which is almost preserved intact despite several breakages along the bottom edge and the presence of corrosion. An opisthograph, the inscription contains nine lines of text, 370

 Britannia As Tomlin has put it, ‘although it is otherwise unknown in Britain, SSS with horizontal line should probably be seen as the “Chnoubis” magical symbol found in GraecoEgyptian magical texts’ (1996: 457). This makes perfect sense if we bear in mind that is precisely during the Late Antiquity when the Graeco-Egyptian techniques (i.e., the use of voces magicae, iconography, magical symbols, etc.) spread throughout the Roman West (see section I.4.2).

phrase in Christian authors in the sense of ecclesia, whether meaning the Church (or Christian community) or a church’ (1985: 25). That said and despite Eccles’ suggestive toponym, the use of the plural diebus (l. 1, for diibus) complicates any argument that the defigens was a monotheistic Christian. Any Christian connection, then, must remain a hypothesis. Butu (with the simplification of t) could be an attempt at the attested cognomen Buttus (see OPEL I, 330 and CIL IV, 3340, 14).

Side B contains the proper curse. Here, the salus and sanitas of Butu, the wrongdoer (who is surprisingly identified by the defigens; for parallels see 259, 355, etc.) are targeted. After dedicating the stolen goods to the deity, the text includes a clause stipulating that the curse is to remain in effect until Butu returns the stolen goods in domo dei (B, l. 5 and perhaps ll. 6–7; for parallels for this modus operandi, see 205, 215 and 356). The phrase in domo dei is first attested here in the corpus of curses from the Roman West and in Tomlin’s words is ‘a common

After being inscribed, the tablet was folded seven or eight times. Translation (Tomlin 1985: 22): ‘A gift to the gods (?) by which Butu has perished, and the property which… neither health nor safety before unless in the house of God… health in the house of God (repeated ?)…’

447. Brean Down (nr Weston-super-Mare, Somerset) Provenance: unkonwn. Current Location: Woodspring Museum. Date: 4th century CE. Material: lead. Measurements: (5.7) × 8.5 cm. Reading:

5



++[-c.2-]++NC[-c.5-]+[-c.1-]++B+[---] caricula quae[-c.3-]+[-c.2- si s]er(v)u[s i] liber si ba[ro] sị [muli]ẹr qui VE+A ut [d]eus [-c.2-] domin`e´ A[-c.3-] EAMAE+DISAM facias sic [i]lla dim[a]t sa(n)guinẹ ṣuo eṭ +ANTISS[-c.2-]+ si bar[o] si mu`lier´ [-c.1-]+++N++[-c.1-]B[-c.3-]

Bibliography: Tomlin 1986: 433–35; Tab. Sulis p. 61, no. 8; Kropp 2008: no. 3.4/1; Urbanová 2018: no. 281.

contra Tomlin). The final -e, which originally could have been a c corrected into e, is located between ll. 3 and 4. The phrase deus domine is somehow unnatural (we would expect domine deus), and unparalleled among curse tablets (but cf. CIL VI, 9657). The text presents the following orthographic features: serus for servus and saguinẹ for sanguinẹ (l. 6). The stolen object, a caricula (l. 2), has been puzzled over. The editor has followed Meyer-Lübke and connected the lexeme to ‘*carricum (a load), from romance derivatives; cognate with the latter, apparently, are the Celtic-derived words carrus/ carrum (a wagon) and carruca (a carriage). Is car(r) iculum the diminutive? And, if so, what has been stolen?’ (Tomlin 1986: 435). For the lexeme baro (ll. 3 and 7) as a synonym for vir, see 249 and 270 as parallels and the analysis of Adams 1992: 15–17.

Image source: Tomlin 1986: 434, fig. 4. Courtesy of R.S.O. Tomlin. Commentary: this defixio was discovered in 1980 by a metal detectorist nearby the Roman temple at Brean Down (dated between 340 and 390 CE and dedicated to an unknown deity). The curse was written on a rectangular lead sheet, which has been broken along its upper edge and has been severely corroded. The inscription contains eight lines, which run from left to right and were written in a new Roman cursive that can be dated to the fourth century CE on palaeographic grounds. Based on its content, this curse should be grouped among the defixiones against thieves. Here, the theft of a caricula (l. 2) is denounced to a male deity ([d]eus [-c.2-] domine or perhaps [d]eus [et?] domine, ll. 3–4,

Note the presence of an all-inclusive formula (ll. 2–3 and 7) which is used to help identify the wrongdoer, whose punishment is sic [i]lla dim[a]t sa(n)guinẹ ṣuo (l. 6). 371

Sylloge of  Defixiones from the Roman West Translation (Tomlin 1986: 435, modified): ‘[… I give to you] the caricula which [I have lost. Whether] slave or free, whether man or [woman] who… God, lord, (?)… you are to make redeem (?) them thus with his own blood… whether man or woman….’

This punishment is well attested in the province and entails payment for the crime with the perpetrator’s own blood (for parallels, see 304, 337, etc.). After being inscribed, the tablet was folded twice.

448. Puckeridge Braughing Provenance: unknown. Current Location: unknown. Date: 2nd–3rd centuries CE. Material: lead. Measurements: 5.5 × 14.7 cm. Reading: A (crux)

B (traces) EM (vacat) (traces) +++A+SRAFVSA [---]+L++S+LLA++++MEQ++RTNDOEFA [---]M (denarii) +RISI+E+SOSAF+HNEIS (vacat) 5 (vacat, crux, vacat)

Bibliography: Tomlin 1986: 436–37; Tomlin 1988b: 68– 69; Kropp 2008: no. 3.20/1.

Establishing a text is difficult for various reasons: the letter forms themselves (it is hard to distinguish a, r and m as well as e and f), certain underscores (B, ll. 2–3), certain strokes (see the end of B, l. 4 where strokes may belong to a previous inscription that ran from bottom to top) and the symbols that appear on both sides of the tablet. Despite these difficulties, the editor has proposed the reading RAFVSA (B, l. 2), which would be the end of a personal name, and has interpreted the second symbol (B, l. 4) as denoting denarii (for parallels, see 397 and 455). For the symbol found on both sides, cf. 259.

Image source: Tomlin 1986: 436–37, figs. 5–6. Courtesy of R.S.O. Tomlin. Commentary: this defixio was found unstratified in 1972 during the excavations carried out in Ermine Street. The curse was written on a lead strip, whose right side is rounded. The tablet is wholly intact, but its surface has been scratched and worn down. Furthermore, the edges are quite corroded, which has led to the loss of part of l. 1 and beginning of ll. 2–4. The inscription contains five lines on side B, which run from left to right and were written in old Roman cursive (dated on palaeographic grounds between 100 and 250 CE). As far as the quality of the preserved text goes, Tomlin has pointed out that ‘[i]ndividual letters, but no continuous text, can be recognised’ (1988b: 68).

After being inscribed, the tablet was pierced twice as can be inferred from the two holes present in the centre of the curse. Next, the tablet was folded five times and then pierced three more times along the lower edge. In Britannia pierced tablets are relatively rare (cf. with 338, 339 and 345).

372

 Britannia 449. Brandon Provenance: sanctuary. Current Location: Moyses Hall Museum. Date: 4th century CE. Material: lead. Measurements: 4 × 5.7 cm. Reading:

5



SERADVASORISDVAS s ser(v)us si ancl(l)a si li liberta si mlier si baro popia(m) fer(re)a(m) EAENEC furtum fecere domino Neptuno cor˹u˺lo pare(n)taṭor

Bibliography: Tab. Sulis p. 61, no. 21; Tomlin 1994b: 293–95; AE 1994, 1112; Kropp 2008: no. 3.3/1; Urbanová 2018: no. 280.

object, there is virtually no room for all that, even if we count on some transcription errors (Tomlin 1994b: 295). This curse contains an all-inclusive formula used to identify the thief (ll. 2–4). Here, it is noteworthy that the first adjoining pairs contrast gender, not social status (l. 2: s(i) ser(v)us si ancl(l)a instead of the more usual si servus si liber).

Image source: Tomlin 1994b: 294, fig. 1. Courtesy of R.S.O. Tomlin. Commentary: this defixio was discovered in 1979 by a metal detectorist in the mud surrounding the Little Ouse River on the south side of the Roman site at Hockwold-cum-Wilton. The curse was written on a rectangular lead sheet and is fully intact. The inscription contains seven lines, which run from left to right and were written in new Roman cursive. On palaeographic grounds, the tablet can be dated to the fourth century CE.

Popia fera, for popia(m) fer(re)a(m) (ll. 5–6) is understood as the stolen object, a pannum ferri (perhaps an ‘iron pan’, see 271 for a parallel). Furthermore, it is worth noting the phrase furtum fecere (ll. 5–6) to allude to the theft instead of the more common involaverit. The phrase cor˹u˺lo pare(n)taṭor (l. 7) also merits mention, since the guilty is offered to the deity as a sacrificial victim (for this type of offering, see 268, 269, etc.). The use of the word corulo could be related to a specific ritual ejecution, which is described by Tacitus: ignauos et inbellis et corpore infamis caeno ac palude, iniecta insuper crate, mergunt (Germ. XII, 1; on this punishment, cf. Marco Simón 2020). Finally, we must note that the invocation of Neptune (also in 340, 441 and 451) perfectly fits the context in which the tablet was deposited (presumably in the River Little Ouse).

Based on its content, this curse should be grouped among the defixiones against thieves. Here, the theft of an iron pan is denounced to Neptune. The text presents the following orthographic and phonetic features: serus for servus (l. 2), popia for popiam (l. 4), fera for ferream (l. 4), fecere for fecerit (ll. 6–7) and paretator for parentatur (l. 7). Tomlin attributes other errors to visual confusion on the scribe’s part: ancela for ancilla (l. 2), mlier for mulier (l. 3) and corlilo for corulo (l. 7; in this hand the similarity between v and li is quite noticeable and hence led to confusion). The first line of the text has been a matter of some speculation: while it ought to have the defigens’ name and perhaps the stolen

Translation (Tomlin 1994b: 293): ‘(Whoever)… whether male slave o female slave, whether freedman or freedwoman, whether woman or man… has committed the theft of an iron pan (?), he is sacrificed (?) to the Lord Neptune with hazel (?).’

450. Weeting-with-Broomhill, Norfolk Provenance: unknown. Current Location: private collection. Date: unknown. Material: lead. Measurements: 6 × 5.5 × 0.1 cm. Reading: A s sevus si [l]ib[e](r) q[u-] i furavit suṣṭulị t ṇẹ ẹị dimitte ṃạḷẹfic(i)um dm 5 tu vindi[c]a[s]

373

Sylloge of  Defixiones from the Roman West B ante dies {e} novẹ(m) ṣi pa ga[n]us si ̣ mil[e]s [qui] 5 su[s]tu[l]it

Bibliography: Tab. Sulis p. 61, no. 20; Hassall and Tomlin 1994: 296–97; AE 1994, 1113a–b; Kropp 2008: no. 3.5/1; Tomlin 2017c: 9.01; Urbanová 2018: no. 282.

Commentary: this defixio was discovered by a metal detectorist in the early 1970s at the site Hockwold-cumWilton and has been edited based on a drawing by T. Gregory from the Norfolk Archaeological Unit in 1979. The curse was written on a small, almost quadrangular lead sheet that is intact though slightly corroded. An opisthograph, the inscription contains 10 lines, which run from right to left and were written in capitals and were mostly ‘mirrored’ (i.e., from right to left; for this layout, see 266, 340, 345 and 350).

explain, ‘is the technical and legal term for “civilian” and is often contrasted with miles, e.g. Vegetius, II, 23, si doctrina cesset armorum, nihil paganus distat a milite’ (1994: 297). This lexeme is found in the curse’s all-inclusive formula which is split up: s sevus si [l]ib[e](r) (A, l. 1) and si paga[n]us si miles (B, ll. 2–4), which is used to identify ̣ the wrongdoer. The second part is unparalleled in the corpus of British curses. Also note the phrase ṇẹ ẹị dimitte (A, l. 3), with which the practitioner asks the god not to let the thief off the hook for his evil deed (ṃạḷẹfic(i)um, A, l. 4); instead, the deity ought to take vengeance (vindi[c]as, A, l. 4) before nine days to come to a close (ante dies novẹ(m), B, ll. 1–2). This magical time frame of nine days is found in other curses (such as 267, 340, 456, 530, etc. For a discussion, see Marco Simón 2010b and Sánchez Natalías 2019a, for magical time frames in general).

Based on its content, this curse should be grouped among the defixiones in fures. The text uses the verbs furavit and sustulit (A, ll. 2–3; B, l. 5) in place of the more common involavit. Also note the use of paganus, which, the editors

Translation (Hassall and Tomlin 1994: 296): ‘(Whoever) has stolen (it), taken (it), whether slave or free, do not forgive him his evil-doing until you punish him within nine days, whether civilian or soldier, (whoever) has taken (it).’

Image source: T. Gregory apud Hassall and Tomlin 1994: 297, fig. 2.

451. Hamble Provenance: unknown. Current Location: Fareham Museum. Date: 4th century CE. Material: lead. Measurements: 12.8 × 8.4 cm. Reading:

5



10



15



domine Neptune tb dno (h)ominẹm qụị ld˹um˺ invọḷạṿ[it] Mụconi et argenṭị[olo]s sex ide(o) dono nomi(n)a qui decepit si mascel si femina si puuer si puuella ideo dono tibi Niske eṭ Neptuno vitam valitudinem sanguem eiụṣ qui conscius fueris eiụṣ deceptionis animus qui hoc involavit et qui conscius fueriṭ ut eum decipias furem qui hoc involavit sanguem eiius consumas et decipias doṃiṇ[e] Nẹ[p]tune 374

 Britannia Bibliography: Tab. Sulis p. 61, no. 6; Tomlin 1997: 455– 58; AE 1997, 977; Tomlin 1999a: 561–62; Kropp 2008: no. 3.11/1; Adams 2016: no. 33; Tomlin 2017c: 12.75; Urbanová 2018: no. 285.

Throughout the text, we find various transcription errors: tb for tibi, dno for dono (both in l. 1; cf. l. 8), ldmu for solidum (l. 2), ide for ideo (l. 5) and sanguem for sanguinem (ll. 10 and 16). The text possesses the expected all-inclusive formula used to identify the thief (ll. 6–8; cf. 441); the victim is also alluded to with the expressions (h)ominẹm qụị (...) invọḷạṿ[it] (ll. 2–3), and simply qui hoc involavit (in ll. 13 and 16), nomi(n)a qui decepit (l. 6, where decepit is a synonym for involavit; for decipio see Adams 2016: 405–06). Also note the phrase qui conscius fuerit, which refers to any potential witness to the crime (ll. 11 and 14; cf. 302, B, ll. 3–4: qui medius fuerit, ‘where the same sense of “guilty knowledge” is required’ [Tomlin 1997: 458]).

Image source: Tomlin 1997: 456, fig. 1. Courtesy of R.S.O. Tomlin. Commentary: this defixio was discovered by a metal detectorist on the shore of the Hamble estuary (Southampton Water) in 1982. Upon discovery the defixio was still rolled up. It was written on a rectangular lead tablet. Both right corners have been corroded and there is a gash at ll. 10–12. Nevertheless, the inscription is well preserved and contains 19 lines, which run from left to right and were written in new Roman cursive.

Also noteworthy is the various aspects of the victim that are targeted: nomi(n)a (l. 5, see 221), vitam, valitudinem, sanguem (ll. 9–10, 16; cf. 249) and animus (l. 12). Finally, the name Muconius (ll. 3–4) is a Celtic name attested here for the first time.

Based on its content, this curse should be grouped among the defixiones in fures. Here, the theft of a sum of money is denounced (ll. 2–4). As is normal, the thief (l. 2) is dedicated to the deities invoked (cf. 215). In this curse, the defigens names Neptune three times and once calls upon Niskus (1. 8), a male god first found here. Tomlin has interpreted Niskus in two different ways: first, as a male god related to the Niskas, which are found in the curses from 150–55 (Tomlin 1997: 457); second, as ‘a Celticized misspelling of Neptunus (…) Since Celtic Latin had no pt, Latin pt was replaced by British (…) Thus Latin Neptunus might be written *Neksunus or even, by analogy with “Nept(un)us” at Caistor St. Edmund (…)*Neksus’ (1999a: 562, n. 52).

After being inscribed, the tablet was rolled up. Translation (Tomlin 1997: 455): ‘Lord Neptune, I give you the man who has stolen the solidus and six argentioli of Muconius. So I give the names who took them away, whether male or female, whether boy or girl. So I give you, Niskus, and to Neptune the life, health, blood of him who has been privy to that taking-away. The thief who stole this, may you consume his blood and take it away, Lord Neptune.’

452. Marlborough Downs Provenance: unknown. Current Location: private collection. Date: unknown. Material: lead. Measurements: 4.9 × 7.2 × 0.01 cm. Reading: (vacat) do (vacat) deo Marti A+VNCISEA id [est?---] ++IV (or P?) eculium (m?)eum et secur[im---] TIDISSIE E +++ illum iumeṇ[tu]ṃ 5 rogạṭ genium tuum dom[ine] ut quampr[imu]ṃ res[ideant?] nec eant per annos novem n[on eis] (traces) permittas nec sedere [nec ---] (traces) [---]+++MIMBRIC+[---]

Bibliography: Tomlin 1999b: 378–79; AE 1999, 975; Kropp 2008: no. 3.16/1; Urbanová 2018: no. 290.

Commentary: a metal detectorist discovered this defixio in 1998 alongside some Roman coins and pottery. It was written on an oblong lead sheet with rounded corners. The tablet’s surface has been corroded to the extent that even though it was originally inscribed on both sides, today

Image source: Tomlin 1999b: 379, fig. 2. Courtesy of R.S.O. Tomlin. 375

Sylloge of  Defixiones from the Roman West Mercury in Uley (see 357 and 373). The text also contains a non permittas formula (ll. 7–8), through which the defigens attempts to harm the physical well-being of the culprit (for nec sedere, see 259, 365 and 366) for a period of nine years (l. 7), a magical time frame which is first found here (nine days, however, is a common time; for magical deadlines, see Sánchez Natalías 2019a).

only two thirds of side A are legible. Nine lines have been preserved, which run from left to right and were written in old Roman cursive. Based on its content, this curse should be grouped among the defixiones in fures. Here, the theft of eculium eum et secur[im] (l. 3, ‘my pony and an axe’) is denounced to Mars. Tomlin has reconstructed the phrase eculium (m)eum (l. 3) as equuleum (m)eum (‘my pony’), a reading that is buttressed by the presence of illum iumeṇ[tu]ṃ (‘beast of burden’) in the following line. The mention of genium in the curse’s invocation of Mars (rogạṭ genium tuum, dom[ine], l. 5) parallels other phrases used to summon the ‘divine majesty’ (e.g., 365). Mars is also invoked in the sanctuaries of Sulis Minerva in Bath (see 238 and 302) and

After being inscribed, the tablet was folded three times. Translation (Tomlin 1999b: 378, modified): “I give to the god Mars… my pony (?)...and an axe ...beast of burden... asks your Genius, Lord, that they (stop) as soon as possible and do not go for nine years. Do not allow (them) to sit (or to…)”.

453. Dodford, Northamptonshire Provenance: unknown. Current Location: unknown. Date: 2nd–3rd centuries CE. Material: lead. Measurements: 8.9 × 7.2 × 0.2 cm. Reading: [---]ṃṇeụi (traces) cl+++ṇịcm +pluminono+telo[-c.2-]at [---]ṣụ[-c.1-]s++silomo++ c ̣ui++rlioṃị+q 5 oploulnsllm++++na pocciapụoịịco[-c.2-]tcs ṃarinan (traces) rt+ ṃasus +msaso si ṣ[-c.3-]isnsus 10 [-c.3-]ns

Bibliography: Tomlin 2009: 347; AE 2009, 733.

reconstructing individual words has proven a challenge (except for the personal name Marina, which can be read in l. 7; see OPEL III, 58). According to the editor, ‘the text, as a whole is not Latin, and some sequences are positively un-Latin. However, it is not a reversed text, and there are not evident anagrams or encipherment. Since the letter-forms are literate, it cannot be a pseudo-inscription either. So perhaps it is spoken Celtic transliterated’ (2009: 347). Perhaps we do have a new Celtic defixio such as 219 and 223 (and perhaps 379 and 406)?

Image source: Tomlin 2009: 347, fig. 48. Courtesy of R.S.O. Tomlin. Commentary: still folded, this defixio was discovered by a metal detectorist in 2005 in the area around Daventry. The curse does not have any of its original edges and has been severely corroded. The inscription contains 10 lines, which run from left to right and were written in old Roman cursive. Based on its palaeographic features, the text has been dated to 150 to 250 CE. Not only is it difficult to establish a reading of the text, but even

After being inscribed, the tablet was folded.

376

 Britannia 454. Calleva Atrebatum, Silchester (Hampshire) Provenance: unknown. Current Location: Winchester City Museum. Date: 4th century CE. Material: lead. Measurements: 10.3 × 6 × 0.1 cm. Reading:

5



Nimincillus ˹Quintinus˺ Iuṇctinus Dcillinae Lon˹g˺inus VSCANIMIHS +NIS+IC+++++ eu(m) qui invalaverit deus det mala(m) plagam

Bibliography: Tomlin 2009: 323–24; AE, 2009, 690; Solin 2012: 230; AE 2012, 841.

Quintinus have been totally scrambled: TNVSIQNIV. Perhaps the difficulties inherent in the text’s right to left layout led the author of the text to botch the name. Docilliana (l. 2) can be taken as a matronymic of Iunctinus (for parallels, see 235 and 303). Finally, note the transcription SVNIONOL for Longinus, where we find o instead of expected g, perhaps due to a visual confusion on the author’s part. There is no reconstruction for the final two names in the list, given that the tablet has been damaged. Solin, for his part, has maintained that, with the exception of Docillina and Longinus, the reading of the rest of the names is highly hypothetical. Nevertheless, he has not offered an alternative reading.

Image source: Tomlin 2009: 323, fig. 13. Courtesy of R.S.O. Tomlin. Commentary: this defixio was discovered around 1901 in unknown circumstances and then acquired by the Winchester City Museum. The curse was written on a rectangular sheet, whose bottom half has rounded corners. The defixio is fully intact, though corroded (especially at ll. 3–4). The tablet also has several small fractures in the top third of the left side as well as the bottom third of the right side. The inscription contains seven lines, which run from right to left and were written in new Roman cursive. The size of the letters varies greatly.

Given that after being inscribed the tablet was pierced twice from the back, Tomlin has suggested the possibility that the curse was hung up on a wall with the text hidden from view. At a later date, according to this theory, it was taken down and folded twice so that the text would be visible. That said, there is no evidence for this sort of procedure elsewhere. Accordingly, it seems more likely that we have a case of reuse: a lead tablet had been nailed up for some unknown purpose before being repurposed for writing a curse (cf. 185, 339 and 520 for other cases of reused tablets). Thereafter, the tablet was folded up.

Based on its content, this curse should be grouped among the defixiones against thieves. Here we find a list of names in the nominative case (ll. 1–4) followed by the phrase qui invalaverit deus det mala(m) plagam (ll. 5–7). This phrase dealing with terrible punishment is without known parallel in the Roman West (for plaga, see OLD, II, D ‘slaughter, destruction (late Lat): percussit eos plaga magna (Vulg. 1, Reg. 23, 5)’). The text presents the following orthographic and phonetic features: eu for eum (l. 4), invalaverit for involaverit (ll. 5–6), and plaga for plagam (l. 7).

Translation (Tomlin 2009: 324): ‘Nimincillus, Quintinus, Iunctinus (son) of Docillina, Longinus, [name], [name]. Him who has stolen, let the god give a nasty blow.’

The names Nimincillus and Iunctinus (ll. 1–2) are first attested here. In l. 1 note how the letters in the name 377

Sylloge of  Defixiones from the Roman West 455. Farley Heath, Surrey Provenance: sanctuary. Current Location: unknown. Date: end of the 3rd century CE. Material: lead. Measurements: 1.7 × 12.5 × 0.2 cm. Reading: A deo DAVIIS // [---] // [---] // ++ Seni{s}lis Senni ++MDV // [---] // [---] // +++ALLIARSIS ATCAN+++ // [---] // [---] // [---] B (denariis) IIII milibus // [---] // [---] // [Aur]elius Se[---] [---] // [---] // [---] // [---]VS[---] [------]

Bibliography: Tomlin 2004c: 335–36; AE 2004, 843; Kropp 2008: no. 3.10/1.

the amount of money reported stolen constitutes a sum that would have been a fortune before the end of the third century CE.

Image source: Surrey County Archaeological Unit, apud Tomlin 2004c: 336, fig. 1.

As already hinted at, this curse should be grouped among the defixiones against thieves since it denounces the theft of four thousand denarii to an unknown local deity. Note the symbol used for denarii (B, l. 1; for a parallel, see 441). The names Seni{s}lis Senni (A, ll. 1–2) and Aurelius Se[---] (B, ll. 1–2) likely belong to the practitioners.

Commentary: this defixio was discovered in 1995 during a clean up at an archaeological site that had been looted. The tablet was located among the remains of a rural Romano-Celtic temple, specifically in the area of the temenos. The curse was written on a thin lead sheet, which is very corroded and still folded up. An opisthograph, the inscription contains, at least, six lines of text, which run from left to right and were written in an old Roman cursive that is difficult to date. Although the cursive e would suggest an earlier date,

After being inscribed, the tablet was folded twice. Translation (Tomlin 2004c: 335): ‘To the god… Senilis (? son) of Sennus… At four thousand of denarii… Aurelius Se[…]’

378

 Britannia 456. Ratae Corieltauvorum, Leicester Provenance: domestic space. Current Location: University of Leicester. Date: 2nd–3rd centuries CE. Material: lead. Measurements: 20.1 × 7.8–7.5 cm. Reading:

d{a}eo Maglo ˹do˺ e{u}um qui frudum fecit de padoio ˹do˺ el{a}eum qui furtum (fecit) de padaoium {sa(g)um} qui sa(g)um Servandi invola5 vit S[il]vester Ri(g)omandus Ṣ[e]nilis Venustinus Vorvena Calaṃịnus 10 Felicianus Ruf{a}edo Vendicina Ingenuinus Iuventinus 15 Alocus Cennosus G̣e«r»manus Senedo Cunoven«d»us 20 Regalis Ni(g)ella [[Sẹ[nic]ịạṇus+]] od ant{a}e nonum diem illum tollat 25 qui sa(g)um involavit Servandi

Bibliography: Tomlin 2008a: 207–15; AE 2008, 792; Tomlin 2009: 327–28; AE 2009, 739; Adams 2016: no. 37.

(l. 2), Rufaedo for Rufedo (l. 11) and antae for ante (l. 23). We must also note the inversion of od for do (ll. 1, 2 and 23), a verb which is already attested in British curses (see 452). Adams, for his part, has offered a different interpretation for the od in l. 23: by comparison with 340, A, ll. 7–9 (rogo tẹ Metunus ut u mi vendịcas...), he has read it as ut, a conjunction that ‘is expected at this point (...) The spelling o for ŭ is common in late Latin, and ut not infrequently has the form ud’ (Adams 2016: 424f., with previous references). Another notable feature of this text is the repeated omission of intervocalic g, as we can observe in the words sa(g)um (ll. 4 and 25), Ri(g) omandus (l. 6) and Ni(g)ella (l. 21). Based on this tendency to drop intervocalic g, Tomlin rightly reads padoio (l. 2) and padaoium (l. 3) as forms of paedagogium, that is the rooms belonging to the slaves in a villa or palace. On this phenomenon, see Adams 2016: 427–427.

Image source: Tomlin 2009: 328, fig. 18. Courtesy of R.S.O. Tomlin. Commentary: this defixio was discovered alongside 457 in 2005 during the excavations in the north-eastern neighbourhood of ancient Ratae Corieltauvorum, specifically in the patio of a mansio. The curse was written on a rectangular lead sheet, which is intact despite being corroded (especially at ll. 6–7 and 22). The inscription contains 26 lines, which run from left to right and were written in old Roman cursive, which can be dated on palaeographic grounds between 150 and 250 CE. Based on its content, this curse should be grouped among the defixiones in fures. Here, Servandus denounces the theft of an article of clothing to the god Maglus. The text presents the following orthographic and phonetic features: daeo for deo (l. 1; on this, see Adams 2016: 424), euum for eum (l. 1), frudum for fraudem (l. 1), elaeum for eleum

After the invocation of Maglus (l. 1) to whom the victims are handed over, the text presents a long list containing the names of all the suspects (ll. 6–22). This list ends with the temporal phrase ant{a}e nonum diem (l. 23), which 379

Sylloge of  Defixiones from the Roman West attested. S[e]nilis (l. 7), Vendicina (l. 12), Regalis (l. 20) and Se[nic]ianus (l. 22) are Latin cognomina that are built on Celtic elements. The text also includes several names first found here: Calaminus (l. 9), Alocus (l. 15), both of which are likely derived from Greek names), as well as Ruf{a}edo (l. 11), Ri(g)omandus (l. 2), Vorvena (l. 8) and Cennosus (l. 16)

constitutes a magical deadline within which the deity ought to expose the guilty party (for parallel phrases, see 267, 340, 450 and 530 and Marco Simón 2010b; for a discussion of magical deadlines in general, see Sánchez Natalías 2019a). As far as the deity invoked is concerned, Tomlin has pointed out that ‘Maglo: Celtic *maglos (prince) is a frequent name-element and might be a divine title here: compare the god Apollo Cunomaglus, to whom the Nettleton Shrub altar is dedicated. But more likely it is the name of a god hitherto unattested, at least in Britain’ (2008a: 208).

Translation (Tomlin 2008a: 208, modified following Adams 2016: 424): ‘I give to the god Maglus him who did wrong from the slave-quarters; I give him who (did) theft (the cloak) from the slave-quarters; who stole the cloak of Servandus. Silvester, Ri(g)omandus, Senilis, Venustinus, Vorvena, Calaminus, Felicianus, Rufaedo, Vendicina, Ingenuinus, Iuventinus, Alocus, Cennosus, Germanus, Senedo, Cunovendus, Regalis, Ni(g)ella, Senicianus (deleted). May he destroy him before the ninth day who stole the cloak of Servandus.’

The Latin cognomina in the text are S[il]vester (l. 6; see OPEL IV, 82), Venustinus (l. 6; see OPEL IV, 155), Felicianus (l. 10; see OPEL II, 137), Ingenuinus (l. 13; see OPEL II, 193), Iuventinus (l. 14; see OPEL II, 211), Ge«r»manus (l. 17; see OPEL II, 166) and Ni(g)ella (l. 21; see OPEL III, 101). The Celtic cognomina are Senedo (l. 18) and Cunovendus (l. 19), both of which are already

457. Ratae Corieltauvorum, Leicester Provenance: domestic space. Current Location: University of Leicester. Date: 2nd–3rd centuries CE. Material: lead. Measurements: 6.9 × 12.3 cm. Reading:

5



{V (traces) abhẹreọ ilis res ir} qụ[i a]rgentios Sabiniani fura- {R} vẹrunt id est Similis Cupitus Loc ̣hiṭa hos deus siderabit in hoc septisọnio et petọ ut vitam suam perdant ante dies septem

Bibliography: Tomlin 2008a: 215–18; AE 2008, 793; Tomlin 2009: 329; AE 2009, 740.

Based on its content, this curse should be grouped among the defixiones against thieves. Here, Sabinianus denounces the theft of a sum of money and curses the perpetrators, calling them out by name (l. 3; for parallels, see 355, 361 and 443). Sabinianus continues, siderabit in hoc septisọnio et petọ ut vitam suam perdant ante dies septem (ll. 4–6). The verb sideratio alludes to the sudden stroke that the victims ought to endure before the passing of seven days. This magical deadline is without known parallels and may have something to do with the deity invoked and the word septisonium (ll. 4–5; see Sánchez Natalías 2019a on magical time frames and deadlines). According to the editor, this lexeme denotes a type of monumental façade decorated with the seven deities who gave their names to the days of the week. Therefore, ‘The unspecified deus in line 4 is one of these “stars”, seven deities appropriate to sideratio and a time-limit for one week’ (Tomlin 2009: 329). On septizonium, see CIL VIII, 2657 (= 18105= ILS 5626) and VIII, 14372 (= ILS 5076); also AE 1964, 182, and s.v. in OLD.

Image source: Tomlin 2009: 329, fig. 19. Courtesy of R.S.O. Tomlin. Commentary: this defixio was discovered alongside 456 in 2005 during the excavations of a domestic patio in the north-eastern neighbourhood of ancient Ratae Corieltauvorum. The curse, which is intact and in good condition, was written on an irregularly shaped lead sheet. The inscription contains six lines, which run from left to right and were written in old Roman cursive, which can be dated on palaeographic grounds between 150 and 250 CE. Of the six lines, ll. 1–2 have been written on top of each other as if the defigens was not satisfied with his first attempt in l. 1 and hence began a second time. This hypothesis is strengthened by the fact that ll. 2–6 stand alone without l. 1 (see the reading above). 380

 Britannia Translation (Tomlin 2008a: 216): ‘Those who have stolen the silver coins of Sabinianus, that is Similis, Cupitus, Lochita, a god will strike down in this septisonium, and I ask that they lose their life before seven days.’

Finally, Sabinianus (l. 2), Similis and Cupitus (both in l. 3) are Latin cognomina that are well attested (see Kajanto 1982²: 154, 289 and 296, respectively), while Loc ̣hiṭa (l. 3) is probably the feminine form of the Greek cognomen Lochus (Solin 2003: 1389 and 1395).

458. East Farleigh, Kent Provenance: sanctuary (?). Current Location: unknown. Date: unknown. Material: lead. Measurements: 6.8 × 10 cm. Reading:

5



Col. I [---]AN or NV[---] Ati˹den˺[us (?)] Sacratu[s] [---]+[-c.1-]RICV[---] [---]ΛḶILE[---] [---]OTIS [(?)Atr]ectus

Col. II CVNDAC[-c.1-]V+ Cunoariṭụṣ [---]MΛVIR[-c.1-]VS Oneratus Memorinus Constitu[tus (?)] Constan˹ti˺us

Bibliography: Tomlin 2012: 8–9; AE 2012, 843; Tomlin 2016: 414–15; AE 2016, 901.

or upside down, it can be difficult to distinguish A (without the horizontal stroke) from upside down V. In addition to this complicated, if not haphazard, layout, the text contains a series of transcription errors (e.g., col. I, l. 2: Atined[---] for Atiden[us], and col. II, l. 7: Constanitus for Constantius).

Image source: Tomlin 2016: 415, fig. 32. Courtesy of R.S.O. Tomlin. Commentary: this defixio was discovered in 2009 during the excavations carried out in a Roman building located in East Farleigh (Kent, near Maidstone). The building dates to the third century CE and has been identified as a sanctuary (though with reservations). The tablet has suffered from corrosion and is broken into various fragments, two of which fit together. The inscription contains 14 lines, which have been arranged in two columns and written in capitals. The layout is rather complicated, since some lines are written upside down (col. I, ll. 1–3 and 5), while some run from right to left (col. I, l. 4) and others from left to right (col. I, ll. 6–7 and col. II, ll. 2 and 5–7). Yet others run from right to left and have letters ‘mirrored’ (i.e., from right to left; col. II, ll. 1 and 4). Since the capital letters are at times mirrored

The circumstances that led to the writing of this curse are unknown and the text contains nothing more than a list of personal names in the nominative. The Celtic cognomina are Atined[---] (col. I, l. 2, perhaps for Atiden[us]), [S]acir[u]s (col. I, l. 4), [Atr]ectu[s] (col. I, l. 7), Cundac[us] (col II, l. 1) and Cunoariṭụṣ (col. II, l. 2, derived from Cunoarus). The Latin ones are Sacratu[s] (col. I, l. 3; see OPEL IV, 42), Oneratus (col. II, l. 4; see Kajanto 1982²: 353), Memorinus (col. II, l. 5, derived from Memor; see Kajanto 1982²: 255), Constitu[tus] (col. II, l. 6; see Kajanto 1982²: 350) and Constanitus (col. II, l. 7, probably for Constantius, see OPEL II, 73). After being inscribed, the tablet was rolled up. 381

Sylloge of  Defixiones from the Roman West 459. Locus Incertus Provenance: unknown. Current Location: private collection. Date: 2nd–3rd centuries CE. Material: lead. Measurements: 10.4 × 13.7 cm. Reading:

[don]ạtur deo Merc[urio si] q[u]ịs involaverit C++++LAM [r?]ịcịnnụm nec non alia minutalia Ṭocitaṃị (?) si baro si mụlier si pueḷ[l]a 5 si puer si ingenuus si servus n[o]n aṇ[t]e eum lasẹt quam mimbra PI maṇụ di em mortịs concrutiat ẹ[u]ṃ qu{q}[i] ṣẹ curam [-c.1-]nnoris involavit EA+++++ AEAPR nec non et qu{qu}i res pịctor[i]ạ[s] (?) 10 involaverit

Bibliography: Tomlin 1991: 293–95; AE 1991, 1167; Kropp 2008: no. 3.24/1; Urbanová 2018: no. 308.

alia minutalia (ll. 2–3) is noteworthy since it semantically parallels the phrase et resculas primulas attested in 365. Next, we find an all-inclusive formula (ll. 4–5) with a few peculiarities: the usual order of the pair si puer si puella has been reversed and we find the new pair si ingenuus si servus instead of the expected si servus si liber.

Image source: Tomlin 1991: 294, fig. 1. Courtesy of R.S.O. Tomlin. Commentary: we do not know where or when this defixio was discovered. A private collector from Bristol acquired it in 1985 during a public auction. According to Tomlin, the defixio probably comes from Gloucestershire or Avon. The curse was written on an irregularly shaped tablet of lead, which was flattened out before being inscribed. The curse has lost its upper corners. The inscription contains 10 lines, which run from left to right and were written in old Roman cursive that can be dated on palaeographic grounds between the middle of the second and end of the third centuries CE.

The phrase n[o]n … eum laset (ll. 5–6; cf. 356) is used to keep the victim from resting; furthermore, the text singles out the thief’s mimbra and manu (l. 6, which should be understood as membra and manus). Between these two, Tomlin has read PI, and reconstructed pi but has noted that ‘a passive infinitive (“to be torn”) has no obvious place here’ (1991: 295). It would make better sense to read another body part (perhaps pedes?, but there is not room for such a restoration). The tablet also contains a magical time frame: (ad) diem mortịs (ll. 6–7; for analogous phrases, see 250 and 350; for magical time frames in general, see Sánchez Natalías 2019a). The Celtic name Tocitamus (l. 4) is first attested here.

Based on its content, this curse should be grouped among the defixiones in fures. Here, the defigens denounces to Mercury the theft of several items: a c[-c.4-]ḷam (l. 2, no restoration has proven possible), a [-c.1-]ịcịnnụm (l. 3, perhaps for ricinum?, ‘a hood’), alia minutalia (l. 3, ‘other sundries’) as well as a securam (ll. 7–8, ‘an axe’) and res [p]ịctor[i]ạ[s] (l. 9, ‘the writing things’). The text presents the following orthographic, morphological and phonetic features: laset for laxet (l. 6), mimbra for membra (l. 6), manu for manus (l. 6), concrutiat for concruciat (l. 7), securam for securim (ll. 7–8, wrongly declined as a first declension noun) and other transcription errors like quq[i] and ququi for qui (ll. 7 and 9, respectively).

After being inscribed, the tablet was pierced with four small conical nails. Given that the holes are alongside the edges and that the tablet was not folded, Tomlin has conjectured that ‘it would seem to have been nailed up for display’ (1991: 293; for parallels, see 95 and 129 and section I.5.4). Translation (Tomlin 1991: 293, modified): ‘Given to the god Mercury, whoever has stolen (…), a hood (?) and other sundries (…) whether man or woman, whether girl or boy, whether free-born or slave. May (the god) not allow him rest before (…) limbs (? …) hands (? …) day of death (…) may (the god) torment him who has stolen the axe (?) of (…) and who has stolen the writing (?) things.’

As the editor has pointed out, ‘the scribe seems to have been copying formulas ungrammatically and without comprehension’ (1991: 295). That said, the phrase nec non 382

 Britannia 460. Locus Incertus Provenance: unknown. Current Location: private collection. Date: 150–275 CE. Material: lead. Measurements: (5.3) × (6.9) cm. Reading:

5



[------] ạmiṣi orọ ṭuam ṃ[aie-] ṣtaṭem ut {f}furem (si) ṣẹrṿ[us] si ạ[nc]il(l)a si [p]uer si [puella] ext[i]nguas [-c.2-] uṭ ilḷi S[-c.4-] CIAS perduc ̣i [r]em ra[ptam] [---]VM ẹt[---] [------] ++

Bibliography: Tab. Sulis p. 61, no. 30; Tomlin 1988c: 488–90; AE 1988, 837; Kropp 2008: no. 3.24/2; Urbanová 2018: no. 309.

ei raptum est… and B, ll. 3–5: …mihi [iu]mentum quod rạpuerunt…) After being inscribed, the tablet was folded three times.

Image source: Tomlin 1988 c: 490, fig. 2. Courtesy of R.S.O. Tomlin.

Translation (Tomlin 1988c: 488): ‘(…) I beg to your majesty that you destroy this thief, whether slave woman (?) or boy or [girl], (…?) that you [force] him to produce the stolen property.’

Commentary: this defixio was discovered in unknown circumstances around 1983, when B.A. Seaby Ltd put it up for sale (Coin and Medal Bulletin, p. 253), and then the curse entered a private collection. The tablet is largely intact, although it was broken into two fragments that fit together; furthermore, it has suffered damage in the areas around the fold lines. The inscription contains nine lines, which run from left to right and were written in old Roman cursive. Based on its content, this curse should be grouped among the defixiones against thieves. Here, the defigens denounces a theft and asks for the victim to be punished. While semantically close to consumas (cf. 259 and 451), we must note that ext[i]nguas (l. 5) is not documented elsewhere in the corpus of British curses. The initial invocation, orọ ṭuam ṃ[aie]ṣtaṭem (ll. 2–3), can be compared to that found in 120, ll. 3–4: per tuam maiestatem te rogo oro obsecro. Next, we find an all-inclusive formula used to discover the thief (ll. 3–4; cf. 449). Starting in l. 5, the defigens asks the deity to punish the wrongdoer, using the already cited verb ext[i]nguas, which should be followed by ut and was perhaps preceded by ita, ‘its force would be to qualify extinguas (“destroy him, so long as the property is returned”)’, (Tomlin 1988c: 490). Finally, the editor has reconstructed the phrase perduc ̣i [r]em ra[ptam] to refer to the stolen property (l. 6; for the participle, cf. 355, A, ll. 4–6: …d(e) iument[o?] quod 383

5 Germania [Germania Superior] Groß Gerau: Scholz, Kropp and Blumer (2006: 181) mention the existence of three defixiones under restoration.

The total number of defixiones from Germania discussed in the following pages is 56 items, all of which are dated to the High Empire. The overwhelming majority of them come from Germania Superior, where 48 tablets have been found, in large part thanks to the important collections from Cruciniacum/Bad Kreuznach and Mogontiacum/Mainz. Outside this area, three curses have been discovered in Germania Inferior, while another five have been found in Agri Decumates. All of this attests the important role played by the Roman military in the diffusion of this magical praxis. In addition to the defixiones analysed in the following pages, we must add several others which are currently under study and await full publication. [Germania Inferior] Leiden-Roomburg: a defixio from this area has recently been acquired by the Dutch National Museum of Antiquities in Leiden. The tablet, which probably comes from the necropolis associated with the Roman fort of Matilo, was discovered in 2008 and was written on a rectangular sheet of lead. The text contains two lines made up of magical symbols and some letters. The tablet dates to c. 200 CE. After being inscribed, the lead sheet was rolled up. Currently, the curse is being studied (see the brief note in Matilo Magazine 7 [February 2010, p. 3]). I thank R. Polak for information about this tablet (per litt.). [Germania Inferior] Neuss: Goedecker-Ciolek, Pause and Scholz (2016) have published two tablets that do not bear inscriptions from ancient Novaesium (Germania Inferior). These curses are currently housed in the Clemens Sels Museum. The first (Inv. no. 1989.48) was discovered in 1980 during the excavations carried out at 20 Büchel Street, where various tombs from the Roman period have been uncovered. Although we have no stratigraphical information about the tablet in question, the editors have argued that it was deposited in the area of the necropolis. The lead sheet is irregularly shaped (approx. 6 × 4 cm) and was still folded and covered with a greyish patina upon its discovery. The second tablet (Inv. no. 1989.427) was discovered in 1971 during an emergency excavation on Kölner Street, where the remains of a legionary camp from the first century CE were found. Upon its discovery, the lead tablet (7.5 × 11 × 0.3 cm) was found folded and pierced with two iron nails. After being restored, it was noted that this large and curved sheet of lead had one smooth edge, whereas another was serrated. This has led to the piece’s identification as a sealing ring for a water pipe that was later repurposed with magical aims in mind. Although it bears no inscription, the tablet contains 14 marks that have damaged the surface of the tablet.

385

Sylloge of  Defixiones from the Roman West Germania Inferior 461. Fectio, Vechten Provenance: unknown. Current Location: PUGKW. Date: unknown. Material: lead. Measurements: 3.24 × 9 cm. Reading:

Smertuccus Amaionis · f(ilius)

Bibliography: Braun 1857: 21–24; Brambach 1867: no. 57; CIL XIII, 8822; Riese 1914: no. 3906; Byvanck 1935: no. 294.

303). Some of the letters are written upside down and others are ‘mirrored’ (i.e., from rigth to left). The circumstances that led to the writing of this curse are unknown. Here the victim is identified with a matronymic. Smertuccus is a Celtic cognomen, which has only been attested in one other inscription from Mainz/Weisenau (Finke 1929: no. 216), while Amaiona is a feminine cognomen found on an epitaph from Rome (see ICUR 9, 23787).

Image source: Brambach 1867: no. 57. Commentary: This defixio was discovered in 1846 or 1847 in the city of Vechten (ancient Fectio) under unknown circumstances. The curse was written on a strip of lead, which has been excellently preserved. The inscription contains two lines of capitals, which run from right to left and bottom to top (for parallels, see 131, 134–36, 149 and

After being inscribed, the tablet was folded in half. Translation: ‘Smertuccus, son of Amaiona.’

462. Gelduba, Krefeld-Gellep Provenance: necropolis. Current Location: Burg Linn Museum. Date: 4th–5th centuries CE. Material: lead. Measurements: 6.4 × 7.2 × 0.1–0.2 cm. Reading:



Theudocsius Lupicinus Iustinianus Leontius Terentianus Aelario Hermoginis Mastidius como oc perversum est sic no[n] POS[-c.3-]O addi nilum cui possit



(descending along the right edge of the tablet): conmo hoc perversum scriptum est sic illos dei spernent

5



10

386

 Germania Bibliography: Galsterer 2000: 32–39; Reuter and Scholz 2004: 43–44; Blänsdorf 2014.

The circumstances that led to the writing of this curse are unknown. The text mentions eight individuals, whose names all appear in the nominative. Noteworthy linguistic and orthographic features include the following: como and conmo for quomodo (ll. 9 and 13), oc for hoc (l. 9) and nilum for nihilum (l. 11). The layout used in the list of personal names (i.e., writing the names backwards from right to left) provides material for the similia similibus formulae found in the second part of the text (for parallels, see 464). The first of these formulae (ll. 9–12) connects the inversion of the cognomina to the victims’ inability to do something (we cannot say what exactly due to a break in the text). The second analogy (ll. 13–14) establishes a comparison between the layout of the personal names and the spurn that the deities ought to show the victims. This formula, sic illos dei spernent, has no parallels in the Roman West. As far as names are concerned, it is worth noting the spelling of Theudocsius (l. 1, for Theodosius, see Solin 2003: 74), Aelario (l. 6, for Hilarius, see Kajanto 19822: 261), Hermoginis (l. 7, for Hermogenes, see Solin 2003: 57 and 1484) and also Mastidius (l. 8, without any known parallels). The rest of the cognomina are all well attested.

Image source: U. Weichhart apud Blänsdorf 2014: 182. Courtesy of J. Blänsdorf. Commentary: this defixio was discovered in 1988 in the necropolis in Krefeld-Gellep, specifically inside of grave 5486, which has been dated to the second half of the first century CE. The tablet was written on a roughly rectangular sheet of lead and is in a good state of conservation. The surface, which is slightly uneven, is covered with a whitish patina resulting from oxidation. Furthermore, we can see signs of damage to the upper-left corner and to the lower third of the tablet. This damage has led to the loss of text corresponding to l. 10. The inscription contains 14 lines of text written in new Roman cursive by a trained hand. The text has been dated on palaeographic grounds between the fourth and fifth centuries CE and the letters measure between 0.3 and 0.5 cm. The curse is structured around a list of personal names in the nominative followed by two cursing formulae, the second of which is written in descending fashion along the right edge of the tablet. We must note that the list of personal names (ll. 1–8) was written from right to left, while the rest of the text was written from left to right.

Translation: ‘(List of names) Just as this text is written backwards, so nothing can (?) be added so (they) cannot (?)... Just as this text is written backwards, so the gods will spurn them.’

463. Bodegraven Provenance: Roman fort. Current Location: Museum Het Valkhof. Date: half of the 1st–-2nd century CE. [Inv. No.: 2006.401] Material: lead. Measurements: 6.5–7.5 × 17 × 0.2 cm. Reading: 5



Col. I Clodium CE++++EM Flav(u)m Placidum Campanum Calticium Atrectun Prolum



Col. II Caltium Poebum {S} Scanitum Iulium Pastorem Lo[g]um Telesinu Cẹruscum Ingenu(u)

387



Col. III Saturninum Alcimum Escingium et si qui afuerint sic ext[r]iti [si]nṭ quomodo h[oc] e[s]t

Sylloge of  Defixiones from the Roman West Bibliography: Haalebos and Polak 2007; AE 2007, 1029; Swinkels 2018.

Poebum for Phoebum (II, l. 2), Telesinu for Telesinum (II, l. 15), Ceruscum for Cheruscum (II, l. 7), Ingenu for Ingenuum (II, l. 9) and Escingium for Exingus (III, l. 3). The following Latin cognomina are rather rare: Logus (II, l. 6, see OPEL III, 30), Telesinus (II, l. 7; see OPEL IV, 111) and Alcimus (III, l. 2; see OPEL I, 72). The text also contains the Celtic cognomina Atrectus and Escingius (I, l. 8 and III, l. 3, respectively) as well as the Greek P(h)oebus (II, l. 2). As Swinkels has noted, the name C(h)eruscus (II, l. 7) is noteworthy for being derived from the ethnonym Cherusken, an old Germanic people.

Image source: Museum Het Valkhof, Nijmegen (NL). Commentary: this defixio was discovered in 2000 in an old garage in Oud-Bodegraafseweg (Bodegraven), which is very close to the castella in Woerden and Zwammerdam. R. Polak (per litt.) has noted, ‘the findspot is probably located in an auxiliary fort built in the pre-Flavian period, in A.D. 61 at the latest (dendrochronological date of the northern gate, but probably of a rebuilding rather than the first building phase).’ More recently, L. Swinkels has argued, ‘a new evaluation of all the available data for Bodegraven now suggests that the findspot lies within the fort, between the barracks of the praetentura sinistra’ (per litt.). Even if we categorize this context as ‘domestic’, this military context is rare in the Roman West (for a possible parallel, see the news about the curse from Neuss at the beginning of this chapter). The curse was written on a lead sheet in the form of a tabula ansata. This shape is quite rare among the corpus of curses from the Roman West, but it is common in writing practices. In this case, the common form has been adopted in a magical context (for parallels, see 337, 479, etc.).

The list of personal names is followed by an all-inclusive formula (III, ll. 3–4), which is used to curse all of the defigens’ adversaries including those not mentioned explicitly in the text (for some parallels, see 342 and 302). Finally, there is a similia similibus formula (III, ll. 5–7) built on the verb extero (the verb is unparalleled in the corpus of Latin defixiones). According to Swinkels’ analysis, the text constructs a persuasive analogy between the way in which the victims’ names are erased and their being wiped off the face of the planet. He goes on to conjecture that the defigens had written the names of the victim elsewhere (perhaps in the sand or on a wax tablet) in order to then wipe them out as part of a sympathetic magical ritual. That said, OLD identifies a third meaning of extero: ‘to crush’. Accordingly, another (and perhaps more plausible) interpretation would be that the defigens literally crushed the inscribed names by folding the tablet or even stepping on it.

The tablet has been very well preserved, though its surface is uneven. The text contains eight lines, which were written in old Roman cursive (but note the repeated capital d) measuring 0.6 cm in height. The lines run from left to right and are organized in three columns. In his re-edition of the tablet, Swinkels dates the curse between the second half of the first century CE and the beginning of the second. This proposal coincides with the text’s palaeographic characteristics and with the other artefacts found in the same area (see Swinkels 2018: 212–13). We do not know the motives that led to the writing of this curse, which consists of a list of 20 personal names in the accusative followed by a cursing formula.

Haalebos and Polak have noted that the names in this curse are well documented in Hispania, which has led them to suggest that this defixio could have been directed against a group of soldiers serving in Germania and stationed on the limes. Indeed, the I Flavia Hispanorum was stationed in Vechten, II Hispanorum peditata  in Utrecht and  II Hispanorum equitata in Maurik. After being inscribed, the tablet was folded.

Noteworthy orthographic and phonetic features are the following: Flaum for Flavum (I, l. 3), Atrectun for Atrectum (I, l. 7), Prolum for Prolum (I, l. 8),

Translation: ‘(List of names) ...and any who were missing. May they be crushed just as this (tablet) is (crushed).’

464. Colonia Claudia Ara Agrippinensium, Cologne Provenance: necropolis. Current Location: RGM Köln. Date: mid-1st century CE. Material: lead. Measurements: 2.7 × 8.2 cm. Reading: A Vaeraca (or Uxeraca) sic res tua pervese agas «c»omodo hoc perverse scriptu(m) est

388

 Germania

B quidquid exopt[a] nobịs in caput tuum eveniat

Bibliography: Blänsdorf 2008: 53; Blänsdorf, Kropp and Scholz 2010; AE 2010, 1023; Blänsdorf 2010b: 150, no. 3; Faraone and Kropp 2010: 382; DTM, p.182; Urbanová 2018: no. 100.

letters) are backwards so too may things be backwards for her’, trans. Faraone and Kropp 2010: 382, n. 5). In addition to this parallel, we must add the formula found on the tablet from Waldmössingen (cf. 479) in which the defigens asks the deities invoked ut illum aut illam aversum faciant di(i) sicut hoc est aversum (‘may the gods render him/her averted, just as this (text) is averted’).

Image source: Blänsdorf et al. 2010: 273. Commentary: this defixio was discovered in 1985 during the excavation of a necropolis on the north-eastern side of ancient Colonia, located between what are now Friesenstraße and Spiesergasse streets. The curse was written on a dark brown strip of lead. Although the tablet’s surface is uneven and somewhat oxidized, it is still whole and in a fairly good state of conservation. An opisthograph, the inscription contains six lines of text, which run from right to left and were written in capitals (letters range from 0.3 to 0.45 in height).

Noteworthy linguistic and orthographic features include the following: the author originally wrote the classical quomodo only to go back and change it to comodo (A, l. 2), the omission of the final m in scriptu(m) (A, l. 3) and the omission of r in pervese (A, l. 2; this error probably does not reflect a phonetic practice). The text’s final phrase in caput tuum eveniat (B, ll. 5–6) clearly reflects several literary curses found in both prose and poetry (for example, Ovid Her., XX, 127 and Ars am. I, 340). This lexical coincidence may suggest that the author of this curse was familiar with highbrow literary culture. As far as the text’s names are concerned, Vaeraca, which has also been read as Uxeraca, has no known parallels. According to the editors, the ending –aca­is characteristic of Celtic personal name (derived from the root uxse-, if we accept the second reading of the name).

The circumstances that led to the writing of this curse are unknown. The centrepiece of the curse consists of a similia similibus formula, in which the defigens enjoins that the target’s affairs shall go awry just as the letters are disordered: pervese agas comodo hoc perverse scriptu(m) est (A, ll. 2–3). Faraone and Kropp have highlighted how ‘this kind of analogy (…) sets up a symmetrical relationship between the ritual manipulation of the text and the intended effects on the victim’ (2010: 382). As the editors have pointed out, this formula should be compared to one found on the Greek defixio SGD 40 (‘Just as these things (i.e. the

Translation (Faraone and Kropp 2010: 383 [side A] and my own): ‘Vaeraca, in this way may you undertake your affairs backwards, just as this text is written backwards. Whatever you wish for us, may come down on your head.’

465. Ulpia Noviomagus, Nijmegen Provenance: unknown. Current Location: Museum Het Valkhof. Date: unknown. [Inv. No.: Mws1.00026] Material: lead. Measurements: 5 × 10.3 cm. Reading:

Vitlia(m) PV[-c.2-3-]+ · adnuntio tibi ad istas manes

389

Sylloge of  Defixiones from the Roman West Bibliography: Weiß-König 2018.

this curse are unknown. The text contains a female name, Vitlia(m) (l. 1; an accusative with the loss of final m), followed by the phrase adnuntio tibi ad istas manes (ll. 1–2). The verb adnuntio is lexically significant, since it does not appear elsewhere in the corpus of Latin defixiones from the Roman West. Here it is used to introduce (and hand over) the victim to the Manes. Here tibi presumably refers to a deity (perhaps Dis Pater), who is not named explicitly in the text. Vitilia is an attested Latin nomen (see OPEL IV, 177).

Image source: Weiß-König 2018: 43, fig. 5.7. Courtesy of R. Reijnen (Radboud University of Nijmegen). Commentary: this defixio was discovered in the city of Nijmegen as a ‘stray find’ (Weiß-König per litt.). It was inscribed on an irregularly shaped sheet of lead. The text, written in capitals, contains two lines, which run from right to left and bottom to top (for such a layout, see 131, 134– 36, 149, 303 and 525). Note that some of the letters have been ‘mirrored’ (i.e., written from right to left, such as s in l. 2), while others have been written upside down (such as m in l. 2). The circumstances that led to the writing of

After being inscribed, the tablet was folded once. Translation: ‘I report Vitilia Pu... to the Manes.’

Germania Superior 466. Cruciniacum, Bad Kreuznach Provenance: necropolis. Current Location: unknown. Date: 1st century CE. Material: lead. Measurements: 2.5 × 6 cm. Reading: A Fructus

B Gracilis

Bibliography: Weckerling 1887: no. 3; Wünsch 1897: 28; DT 94; CIL XIII, 2, 1, 7555, 1; Weckerling 1904: no. 4; Jeanneret 1917: no. 94; Preisendanz 1933: 153; Kropp 2008: no. 5.1.4/1; DTM, p. 186; Urbanová 2018: no. 72.

was written on a strip of lead, which at some point was broken into two pieces that fit together. An opisthograph, the tablet contains two lines of old Roman cursive which run from left to right. The circumstances that led to the writing of this curse are unknown. These lines contain the name of the victim, Fructus (see Kajanto 19822: 352), followed by his patronymic. The reasons that led to the writing of this curse are not detailed, but the appearance of Fructus Gracilis in 467 leads us to the conclusion that this text formed part of a group of juridical defixiones. After being inscribed, the tablet was folded in half.

Image source: CIL XIII, 2, 1, 7555, 1. Commentary: this defixio was discovered c. 1885 in the Roman necropolis of ancient Cruciniacum, specifically in area 54, section 158/25 (see Geib and Petry, apud Oxé 1926: 144). The curse, which was found alongside 467,

467. Cruciniacum, Bad Kreuznach Provenance: necropolis. Current Location: Stadt Worms Museum. Date: 1st century CE. [Inv. No.: R 6154] Material: lead. Measurements: 5.5 × 4.5 × 0.1 cm. Reading: A Fructus Gra cilis et Aurum Aditorium dẹ[fe-] 5 ro in fẹ̣ ris 390

 Germania

B sic non possit respo[nde]re qua(e)s[tionibus]

Bibliography: Weckerling 1887: no. 4; Wünsch 1897: 28; DT 95; Adami 1903: no. 4; CIL XIII, 2, 1, 7555, 2; Weckerling 1904: no. 5; Jeanneret 1917: no. 95; Preisendanz 1933: 153; Kropp 2008: no. 5.1.4/2; DTM, p. 187; Urbanová 2018: no. 73.

An opisthograph, the defixio contains 10 lines of text which run from left to right and were written in old Roman cursive. Based on its content and the reconstruction non possit respo[nde]re qua(e)s[tionibus] (B, ll. 1–3; according CIL), this curse belongs among the juridical defixiones. On side A we find the names of the curse’s victims, which appear in both the nominative (Fructus Gracilis, also attested in 466) and the accusative (Aurum Aditorium, whom Wünsch believed was the same as the Auditorius found in 470, ll. 4 and 6). For lists of names that mix nominatives and accusatives, see section I.7.2. Finally, note that the phrase dẹ[fe]ro infẹris (A, ll. 4–7, proposed ̣ by Adami) would have a clear parallel in 470 (l. 7).

Image source: CIL XIII, 2, 1, 7555, 2. Commentary: this defixio was discovered c. 1885 in the Roman necropolis of ancient Cruciniacum, specifically in area 54, section 158/25 (see Geib and Petry, apud Oxé 1926: 144). The curse, which was found alongside 466, was written on an irregularly shaped sheet of lead, which has been broken into two pieces that fit together. My autopsy of the defixio (March 2018) has confirmed that the lower portion of the tablet has been lost. For this reason, we only have the upper fragment of a larger text (A–B, ll. 1–3).

Translation (Urbanová 2018: 288, modified): ‘I hand over Fructus Gracilis and Aureus Adiutorius to the Underworld so that they cannot answer the questions.’

468. Cruciniacum, Bad Kreuznach Provenance: necropolis (?). Current Location: Stadt Worms Museum. Date: 1st century CE. [Inv. No.: 4679 a] Material: lead. Measurements: 10.4 × 7.6 × 0.1 cm. Reading: A inimicorûm nomina`ad Infe-´`ros´ Optatus Silonis Faustus Ornạtus (?) Ṭẹrentius Attisso 5 Atticinus Ammonis Latinus Valeri Adiutor Iuli Terṭius Domiti Mansuetus Senodaṭiûm 10 Montanus materiarius Aninius Victor Quartio Severi Sintọ Valentis Lutumarus lanius 15 Similis Crescentis Lucanus Silonis Communis Mercatoris Publius offector Aemilius Silv̂anûs 20

(From bottom to top on the right side) Cossus Matuini

391

Sylloge of  Defixiones from the Roman West

B inimicorum nomina ad Inferos (Diagonally from left to right) ++ICLV̂ M

Bibliography: Weckerling 1887: no. 1; Wünsch 1897: 28; Adami 1903: no. 1; DT 96; CIL XIII, 2, 1, 7553; Weckerling 1904: no. 1; Diehl 1910: no. 858; Jeanneret 1917: no. 96; Preisendanz 1933: 153; Kropp 2008: no. 5.1.4/3; Urbanová 2018: no. 74.

the following cursing formula inimicorum nomina ad Inferos (A, l. 1). Since there was not enough space for the full formula, it is squeezed in the right-hand margin at the level of ll. 2–3. Based on the content, we can determine that this is a juridical defixio. In addition to the formula inimicorum nomina ad inferos (A, l. 1, B, ll. 1–2, with clear parallels to 469, 472 and 476), the text presents a list of the victims’ names (contra Urbanová 2018: no. 74). Among the cursed, we must note that Optatus Silonis, Terentius Attiso, Atticinus Ammonis and Cossus also appear in 473 (ll. 3–4, 7–8, 7 and 4, respectively), while Sinto Valentis is mentioned in 470 (l. 1) and Mansuetus in 471 (l. 1). This last victim as well as two others (Montanus materarius and Lutumarus lanius at B, ll. 10 and 14) are further identified by their professions. Among the cognomina found in this curse, note Atticinus (A, l. 5, see OPEL I, 209), Sintọ (A, l. 13; see OPEL IV, 84), Lutumarus (A, l. 14, see OPEL III, 40) and Cossus (A, l. 20, see OPEL II, 80). After the curse was inscribed on side A, the tablet was folded twice. At this point, the formula inimicorum nomina ad Inferos was added to resume the curse’s general idea.

Image source: Weckerling 1904: 93, fig. 1. Commentary: this defixio was discovered in the Roman necropolis of ancient Cruciniacum in 1885, specifically in area 54, section 158/25 (see Geib and Petry, apud Oxé 1926: 144). The tablet was found with 470 inside of a clay cinerary urn, which also held two bronze coins from the reign of Vespasian. The tablets were both rolled up together (for a parallel for this type of deposit, see 101–02 and 125–26). This curse was written on a roughly rectangular lead sheet that had been repurposed (according to Wünsch, ICLVM of A, l. 4 is a vestige of a previous inscription). The tablet has irregular margins and has lost its upperleft corner. Despite this minor damage, the defixio is well preserved, as I confirmed during autopsy (March 2018). An opisthograph, the curse has 24 lines that run from left to right and were written in capitals. The text is arranged in the following manner: first, the defigens wrote the names of his/her enemies (A, ll. 2–20), after which (s)he added

Translation: ‘The names of the enemies (have been handed over) to the underworld (a list of names follows).’

469. Cruciniacum, Bad Kreuznach Provenance: necropolis. Current Location: unknown. Date: 1st century CE. Material: lead. Measurements: 5.5 × 8 cm. Reading: A Data nomina ad inferas larvạs

B Dis Manibus hos v(oveo) L(ucium) C(a)eli(um) C(aium) HAEB+++ (traces) et siquos alios hos[tes] habeo + 5 neca illa noṃ[ina]

392

 Germania Bibliography: Weckerling 1887: no. 5; Wünsch 1897: 28; Adami 1903: no. 5; DT 97; CIL XIII, 2, 1, 7555, 3; Weckerling 1904: no. 6; Jeanneret 1917: no. 97; Preisendanz 1933: 153; Kropp 2008: no. 5.1.4/4; DTM, p. 187; Urbanová 2018: no. 75.

the tablets were found together and share a similar hand, it seems quite probable that a single author wrote both texts). The reasons that led to the writing of this curse are unknown, but we do learn that the practitioner wished to consign his/her victims ad inferas larvas and to the Manes (for parallels, see 17 and 48). Despite the curt nature of the text, the use of the word hos[tes] (B, l. 3) allows us to categorize this curse as a juridical defixio (on this, cf. section I.7.5). Of all the victims, only L(ucium) C(a)eli(um) (B, l. 2, who is identified with a patronymic) is named explicitly; the rest are grouped together with an allinclusive formula (for a parallel, see 471, l. 3: et omnes…).

Image source: CIL XIII, 2, 1, 7555, 3. Commentary: this defixio was discovered with 471 in 1885 in the Roman necropolis of ancient Cruciniacum, specifically in area 54, section 158/25 (see Geib and Petry, apud Oxé 1926: 144). It was written on an irregularly shaped sheet of lead.

Translation: ‘(These) names have been handed over to the ghosts of the underworld. I consecrate to the Manes Lucius Caelius, Caius Haeb..., and any other enemies I may have. Kill these people.’

An opisthograph, the inscription contains seven lines of text that run from left to right and were written in an old Roman cursive similar to that found on 471 (given that

470. Cruciniacum, Bad Kreuznach Provenance: necropolis. Current Location: Stadt Worms Museum. Date: 1st century CE. [Inv. No.: 4679 b] Material: lead. Measurements: 7.6 × 13.4 × 0.1 cm. Reading: 5



10

Vale(n)tis Sinto sive alii inimici (traces) Valentinus Sinto inimcus sic comdi plumbum subsidet sic Sintonem et Martialem Sinṭ[onis] et Adiutorium Sintonis et quisqụis contra Rubrium fr[atre]m et me Quạrṭionem siqui contravenerit Sintonem et Adiuṭorium eius Sintonis defero ad infero sic nusquam contra nos [inve]ṇissẹ respoṇs[io-] nis cum loquantur inimi[c]i sic desumạṭ non parentem tanquam inferos

Bibliography: Weckerling 1887: no. 2; Wünsch 1897: 28; Adami 1903: no. 2; DT 98; CIL XIII, 2, 1, 7554; Weckerling 1904: no. 3; Jeanneret 1917: no. 98; Preisendanz 1933: 153; Kropp 2008: no. 5.1.4/5; DTM, p. 186; Urbanová 2018: no. 76.

text was written in old Roman cursive that can be dated to the first century CE based on palaeographic criteria. Based on its content, this text can be categorized as a juridical defixio, since it includes the word inimicus (ll. 1, 2 and 9) as well as makes clear allusion to the victims’ capacity to speak (ll. 8–9; cf. section I.7.5). We must note that within the text both the cursers (Quartius and Rubrius) and cursed (mostly referred to with their patronymics) appear. Among the victims, we find Sinto, son of Valens (l. 1, he also appears in 468 B, l. 13). We must also note the use of a similia similibus formula (ll. 2ff., where we find comdi for quomodo), by means of which the text establishes a persuasive analogy between the curse’s own deposition and the fate of the victims (for a similar case, see 157, 518 and perhaps 530). Regarding the names found in the text, the cognomina Sinto (l. 1; see OPEL IV, 84) and Rubrius (l. 5; see OPEL IV, 32) deserve special mention.

Image source: Weckerling 1904: 94, fig. 2. Commentary: this defixio was discovered in the Roman necropolis of ancient Cruciniacum in 1885, specifically in area 54, section 158/25 (see Geib and Petry, apud Oxé 1926: 144). At the moment of its discovery, the tablet was found rolled up with 468 inside a clay cinerary urn, where two coins from the reign of Vespasian were also discovered (for a parallel deposit, see 125–26). The curse was written on an irregularly shaped lead sheet and is still whole and in a good state of conservation, as I confirmed during an autopsy (March 2018).

After being inscribed, the curse was rolled.

The inscription contains 10 lines of text which run from left to right. The author left a wide right margin, which was later used to correct several omissions in the text. The

Translation (Urbanová 2018: 299, modified): ‘Sinto (son?) of Valens, or other enemies. Sinto Valentinus, the 393

Sylloge of  Defixiones from the Roman West enemy. Just as the lead sinks down, so I hand over to the gods of the Underworld Sinto and Martialis, (son or slave) of Sinto, and Adiutorius, son of Sinto, and whoever [is] against my brother Rubrius and me, Quartio. If anyone comes out against [us], Sinto and Adiutorius, son Sinto. In

this way, (they can) never contrive(?) a response against us, when the enemies speak out. In this way, may (this lead tablet) afflict? (Sinto) absent (at court?) just like (i.e.: as if he was in(?)) the Underworld.’

471. Cruciniacum, Bad Kreuznach Provenance: necropolis. Current Location: unknown. Date: 1st–2nd centuries CE. Material: lead. Measurements: 5.5 × 6.5 cm. Reading: 5

Ṃansu[e]tụ[s] Secụṇdina et oṃnes qui illi ass[u]nt +et doc[e]n[t illu]m

old Roman cursive similar to that of 469. Given that the tablets were found together and share a similar hand, it seems quite probable that a single author wrote both texts. Based on its content and the analysis of Wünsch, this curse belongs to the group of juridical curses and is directed against Mansuetus (who also appears at 468 B, l. 9), Secundina, et oṃnes qui illi ass[u]nt et doc[e]n[t illum] (ll. 3–5), that is to say all those who would go to court and testify. In this all-inclusive formula (cf. 469), note the form assunt (l. 4, for adsunt). For the well attested cognomen Mansuetus, see OPEL III, 51–52.

Bibliography: Weckerling 1887: no. 6; Adami 1903: no. 6; DT 99; CIL XIII, 2, 1, 7555, 4; Weckerling 1904: no. 7; Jeanneret 1917: no. 99; Preisendanz 1933: 153; García Ruiz 1967: no. 33; Kropp 2008: no. 5.1.4/6; DTM, p. 188; Urbanová 2018: no. 77. Image source: CIL XIII, 2, 1, 7555, 4. Commentary: this defixio was discovered alongside 469 c. 1885 in the Roman necropolis of ancient Cruciniacum, specifically in area 54, section 158/25 (see Geib and Petry, apud Oxé 1926: 144). The curse was written on an irregularly shaped sheet of lead and contains six lines of text, which run from left to right. It was written in an

Translation: ‘Mansuetus, Secundina, and all those who assist and inform him.’

472. Cruciniacum, Bad Kreuznach Provenance: necropolis. Current Location: Bonn Museum. Date: 2nd century CE. [Inv. No.: 4523] Material: lead. Measurements: 4.8 Ø × 0.1 cm. Reading: A nomina data dẹḷạ ta le[ga]ṭạ ad inferos 5 u[t] illos per vim [c]orrip[i]ạṇt

394

 Germania 5



B Silonia Surum Cae`nu(m)´ Secundum ille te (s)ponsus proca[t] il(l)um âmo

Bibliography: Klein 1891: 131–36; Wünsch 1897: 28; DT 100; CIL XIII, 2, 1, 7550; Weckerling 1904: no. 8; Diehl 1910: no. 857; Jeanneret 1917: no. 100; Preisendanz 1933: 154; Kropp 2008: no. 5.1.4/7; DTM, p. 184; Urbanová 2018: no. 78.

Based on its content, it seems that this defixio belongs to the group of erotic curses. On side A the defigens uses a formula, common among the group from Cruciniacum, to consign the victims to the infernal powers (see 468–69 and 476 for parallels) in order that per vim [c]orr[i]piạṇt (A, ll. 5–6). On side B we find a list of personal names: the first is in the nominative, while the rest are in the accusative. This can be understood as the defigens followed by his victims, although see a discussion of mixed cases in section I.7.2. Pointing to the layout and the use of smaller letters, Wünsch (apud CIL) rightly observed that it appears that the names Surum and Caenu(m) were added at a later point. The same scholar has offered the following interpretation of the final formula: ‘Ille (Secundus) te (Siloniam) sponsus procat: (ego autem) illum amo.’ Of the names in the text, we must highlight the Latin cognomina Silonia (B, l. 1; see OPEL IV, 81) and Caenus (B, l. 2; see OPEL II, 18).

Image source: Klein 1891: Tab. VI, 2 and 2a. Commentary: this defixio was discovered in 1885 during the excavation carried out in the Roman necropolis of ancient Cruciniacum, specifically in area 54, section 158/25 (see Geib and Petry, apud Oxé 1926: 144). Inside one of the graves, excavators found a broken glass urn (21 cm in height), which contained the burnt remains of a person and the defixio under consideration. The curse was written on a lead disc (for this shape, see 64, 145 and 514), which has been well preserved despite slight damage to its surface. An opisthograph, the inscription contains 12 lines of text, which run from left to right and were written in an old Roman cursive. The curse has been dated to the second century CE based on palaeographic criteria.

Translation (Urbanová 2018: 307, modified): ‘The names given, handed over, bound(?) to the Underworld, so that (the infernal gods) seize them by force. Silonia (curses) Surus, Caenus, Secundus. That (Secundus) proposes to you (Silonia). (But) I love him(, too).’

473. Cruciniacum, Bad Kreuznach Provenance: necropolis. Current Location: Bonn Museum. Date: 1st century CE. [Inv. No.: 4524] Material: lead. Measurements: 7.5 × 10.7 × 0.1 cm. Reading:



inimici et inimici Caranita[n]i Abilius Iuenis Ṣabinus ap[p]aritor Arria Gavisa (?) Optatus Silonis Privatu[s Se]veri Cossus Maesi M`a´rcus aerariụ[s] Atta Marci uxsor Camulia uxso[r] [-c.1-]Gamati Amans Valeri Atticinus [Am]monis Terentius Attiso Iulia (?) Attisonis Narcis(s)us Calipunis Caliph[o]nti[s e]ṭ Pudentis et Pudes ++++[-c.9-10 (?)-]VS B[a]lbus Vicinus [---] (traces) ++++ATE (?)



(along the left and right edges)

12 13

sic te morbo a(d)dicant dii M[anes] [---a]d(d)icere inferis [et p]unire

5



10

395

Sylloge of  Defixiones from the Roman West Bibliography: Klein 1891: 136–44; Wünsch 1897: 28 and ff.; DT 101; CIL XIII, 2, 1, 7551; Weckerling 1904: no. 2; Jeanneret 1917: no. 101; Preisendanz 1933: 154; Kropp 2008: no. 5.1.4/8; DTM, p. 184; Urbanová 2018: no. 79.

of the tablet, as was customary in letter writing (cf. Sarri 2018: 112). Even if the exact circumstances that led to its writing are unknown, this curse bears a great similarity to 468. Under the general heading of inimici et inimici (for inimicae; l. 1), the text provides a long list of Caranitanus’ adversaries. Among these individuals, Optatus Silonis (ll. 4–5), Atticinius Ammonis (l. 7) and Terentius Attiso (ll. 7–8) were all also cursed in 468 (B, ll. 5 and 4, respectively). Alongside the victims who are identified with a patronymic and by their profession, we also find two wives (uxsor, ll. 5 and 6; note the hypercorrection xs for x) who are likewise cursed. The list of names concludes with a final curse through which the defigens unleashes illness on his victims (morbo; for a parallel, see 129: accedat morbo cotidea) before sending them to the underworld. As far as names goes, note Arria (l. 3; see OPEL I, 174), Cossus (l. 4; see OPEL I, 80), Atta (l. 5; see OPEL I, 205 and following), Amans (l. 6; see OPEL I, 88) and Atticinus (l. 7; see OPEL I, 209).

Image source: Klein 1891: Tab. VI, fig. 3. Commentary: this defixio was discovered c. 1885 in the Roman necropolis of ancient Cruciniacum, specifically in area 54, section 158/25 (see Geib and Petry, apud Oxé 1926: 144). The Museum of Bonn acquired the tablet in 1886, where it is still housed today. The curse was written on a rectangular lead sheet with rounded corners. The tablet has been broken in the centre and along its lower edge, where the bottom-left corner has been lost. It contains 13 lines of text, which were written in old Roman cursive by a trained hand. The text runs from left to right, except for the final two lines: l. 11 runs down the left margin, l. 12 along the bottom edge, and l. 13 back up the right side of the tablet. There is a simple explanation for this rather unusual layout: the author ran out of space and decided to use the margins

After being inscribed, the tablet was rolled up.

474. Cruciniacum, Bad Kreuznach Provenance: necropolis. Current Location: Bonn Museum. Date: 1st century CE. [Inv. No.: 4299] Material: lead. Measurements: (6) × (9) × 0.1 cm. Reading: A [S]ẹverinus et Santius [defe]ro in[f]e[r]is nomina [mi]sere aut sic mihi non suspendat 5 [---]suspendat [Sev]eriṇ[us] B [---]IA in sunt Bibliography: Klein 1891: 144–46; Wünsch 1897: 29; DT 102; CIL XIII, 2, 1, 7552; Weckerling 1904: no. 9; Jeanneret 1917: no. 102; Kropp 2008: no. 5.1.4/9; DTM, p. 185; Urbanová 2018: no. 80.

the curse has been dated to the first century CE. Though the circumstances that led to the creation of this curse are unknown, we do learn that Severinus and Santius (A, l. 1; the name is for Sanctius) are consigned to the underworld through the formula [defe]ro in[f]e[r]is nomina (A, l. 2; see 468, 469, 472 and 476 for parallels). Blänsdorf (apud DTM) has interpreted the phrase non suspendat (A, l. 4) as non respondat, which certainly would make sense for a juridical defixio. That said, this proposal cannot be accepted on palaeographic grounds, at least not without conducting an autopsy to revise the reading. Regarding the names found in the curse, Severinus and Sanctius are two well attested Latin cognomina (see Kajanto 1982²: 257 and 252, respectively).

Image source: Klein 1891: Tab. VI, fig. 1. Commentary: this defixio, which has been broken into four fragments, was probably discovered in 1885 in the Roman necropolis of ancient Cruciniacum, specifically in area 54, section 158/25 (see Geib and Petry, apud Oxé 1926: 144). In its fragmentary state, the tablet’s right edge and lower portion have been preserved, despite two cracks running across the curse.

After being inscribed, the tablet was rolled up.

An opisthograph, the inscription contains eight lines of text, which run from left to right and were written in old Roman cursive. Based on palaeographic considerations,

Translation: ‘Severinus and Santius. I hand their names over to the underworld...’ 396

 Germania 475. Cruciniacum, Bad Kreuznach Provenance: necropolis. Current Location: Bad Kreuznach Museum. Date: 1st–2nd centuries CE. [Inv. No.: 5144] Material: lead. Measurements: 9.3 × 9.5 cm. Reading: 5



10

Potitus Fusci ADV+[-c.1-]+ Ịṿisum Valli Marullum Pusionis Maxsumus Prịụni Ṇ[e]rvinum Paterni Maturum Suavis Turicum Macri Sulpiciuṃ Secundae [-c.1-]+ Prudentẹṃ Solve(n)di Mensor Marulli Novim[arum ?] Marulli s[ervo]s Secundạ +S lingua[-c.4-]+CI datur

Bibliography: Oxé 1926: 144–45; AE 1927, 68; Finke 1929: no. 236; García Ruiz 1967: no. 34; Solin 1968: 14; Kropp 2008: no. 5.1.4/10; DTM, p. 188; Urbanová 2018: no. 81.

on its content and the use of the word lingua (‘tongue’, l. 10; instead of litis as printed in the editio princeps), this curse can be placed among juridical defixiones, even though the preserved text consists of little more than a list of personal names (cf. section I.7.5). The victims are listed with a bipartite formula made of a cognomen and a patronymic. Sulpicius (l. 6) is the one exception: as my autopsy has confirmed, we find the matronymic Secundae instead of a patronymic (the editio princeps has Secundani). Within the list of names, it is noteworthy that three are in the nominative case: Potitus, Maxsumus (with the hypercorrection of xs for x) and Mensor (ll. 1, 3 and 8); the rest of the names appear in the accusative. Normally, this is interpreted as a way of separating the cursed from cursers (see 472 and 495 for examples), though see section I.7.2 for a discussion. This list closes with a final formula meant to give (l. 10, datur) the victims over to the deities invoked. The Latin cognomina that deserve special attention are Marullus (l. 2; see OPEL III, 62), Nervinus (l. 4; see CIL XIII, 3088) and Turicus (l. 5; see OPEL IV, 134). There are no parallels for Ivisus (l. 2). After being inscribed, the tablet was folded.

Image source: Oxé 1926: 144, fig. 1. Commentary: this defixio was discovered c. 1885 in the Roman necropolis of ancient Cruciniacum, specifically in area 54, section 158/25 (see Geib and Petry, apud Oxé 1926: 144). The tablet had been deposited alongside two fibulae (one of bronze, the other of iron), a hand-shaped urn, a ceramic cup decorated with stripes and a terra sigillata dated to the first half of the first century CE (the curse has been assigned a later date that this pottery). The defixio was inscribed on what must have originally been a square sheet of lead. As my autopsy (March 2018) has confirmed, the top and right edges of the tablet have been wholly conserved, while the left side has been almost completely lost and the bottom edge has been quite damaged. The curse contains 10 lines of text, which run from left to right and were written in old Roman cursive. Based

476. Cruciniacum, Bad Kreuznach Provenance: necropolis. Current Location: Bad Kreuznach Museum. Date: 1st century CE. [Inv. No.: 5145 o] Material: lead. Measurements: 3.6 × 7.9 × 0.1 cm. Reading: A data nomina haec ad inferos

397

Sylloge of  Defixiones from the Roman West B (traces)

Bibliography: Oxé 1926: 145–46; AE 1927, 69; Finke 1929: no. 237; García Ruiz 1967: no. 35; Solin 1968: 15; Kropp 2008: no. 5.1.4/11; DTM, p. 188; Urbanová 2018: no. 82.

conservation. Side B, however, is currently illegible and is reddish and off-white due to oxidation. Because of these problems with the verso, I could not check the editor’s reading of side B. The curse contains at the very least three lines of text that run from left to right and were written in old Roman cursive. We do not know the circumstances that led to the writing of this curse. On side A, however, we can read the formula data nomina haec ad inferos, which is also found on other curses from ancient Cruciniacum (468 and 472) and was used to deliver victims to the powers invoked. The names of the victims were probably written on side B, as Oxé has suggested, though this cannot be confirmed today.

Image source: Oxé 1926: 145, fig. 2. Commentary: this defixio was discovered c. 1885 in the Roman necropolis of ancient Cruciniacum, specifically in area 54, section 158/25 (see Geib and Petry, apud Oxé 1926: 144). The defixio was written on a rectangular sheet of lead. As I was able to confirm during an autopsy (March 2018), side A has been fully preserved and is in a good state of

Translation: ‘These names have been handed over to the underworld.’

477. Frankfurt am Main Provenance: necropolis. Current Location: unknown. Date: 2nd century CE. Material: lead. Measurements: (5.5) × (7) cm. Reading: A [i]nimicos Sexti ut [s or h]ic non [p]ossint [con-] [t]ra Sext[um] fac veni+[---] nec agerọ quicq[uam] 5 possint [u or e]t sic [sint] vani et ṃuti Q[---] DI et illi qui in ITOAC

5



10

398

B Loṭum loqui Va[le-] ntinus et [Fron-] to et Ripanus et Le[---] [---] et Iuventin[-c.2-]+++[---] [---] et Luci[u]s e[t -c.4-]GAR[---] [F]rontonem [---] [---]LI adversari [---] sint vani et ṃ[uti] [qu]omodi ista GARV[---] [---]S[---]

 Germania Bibliography: Woelcke 1926: 80–84; Schillinger-Häfele 1977: no. 123; AE 1978, 546; Solin 2004: 128; Kropp 2008: no. 5.1.2/2; Blänsdorf and Scholz 2011: 66–69; ILGIL 592; Urbanová 2018: no. 71.

Based on the use of the words [i]nimicos (A, l. 1) and adversari (B, l. 7) as well as the expression sint vani et muti, which is a clear allusion to the silencing of the victims (A, ll. 5–6; B, l. 8), the curse can be grouped among the juridical defixiones (cf. I.7.5). Regarding orthography, note agero for agere (A, l. 4) and quomodi for quomodo (B, l. 7). Following Bohn (apud Woelcke 1926: 81–82), Schillinger-Häfele accepted the reconstruction of the final phrase ist garu[li] (B, l. 9; also perhaps in l. 5). Nevertheless, such a phrase would find no parallels in the corpus.

Image source: Woelcke 1926: 83, fig. 18. Commentary: this defixio was discovered in 1925 during the excavation of the necropolis of Frankfurt-Praunheim. The tablet was found inside crematory burial no. 13, which contained a rich array of grave goods dated to the second century CE. The grave goods included five terra sigillata cups (Drag. 27 and 37), the remains of sigillata from southern Gaul, two pitchers and a lamp (for the archaeological context, see Woelcke 1926: esp. 80–81). The defixio disappeared during World War II.

The victims are grouped together at the outset of the text with the all-inclusive [i]nimicos Sexti (see 137 for a parallel) and then appear listed individually on side B. All listed names are well-attested Latin cognomina. Among these, note: Lotus (B, l. 1, cf. Kajanto 1982²: 336), Ripanus (B, l. 3, cf. OPEL IV, 30), Iuventinus (B, l. 4, see OPEL II, 211).

The curse was written on what now appears to be an irregularly shaped sheet of lead, though it was perhaps originally rectangular. At the time of publication, the defixio was in a poor state of conservation: it did not have any of its original four edges and had been greatly corroded (especially side B). The text contains 18 lines, which run from left to right and were written in old Roman cursive. As far as the text’s date is concerned, Bohn (apud Woelcke 1926: 82) placed the curse within the same time frame as the rest of the findings, and therefore dated it to the second century CE.

Translation (Urbanová 2018: 298, modified): ‘[I commend] the enemies of Sextus, so that they are not able to come out or take whatever actions against Sextus... so that they are untrustworthy and mute in such a way... and those who... to speak... Valentinus and Fronto? and Ripanus and... Iuventinus and Lucius... Fronto [in the accusative]... the adversaries... may they be untrustworthy and mute just like....’

478. Frankfurt am Main Provenance: necropolis. Current Location: Frankfurt-Praunheim Museum. Date: 2nd century CE. [Inv. No.: α 11] Material: lead. Measurements: unknown. Reading: A rogo Mane[s] Inferi ut Ṃ[a-] rius Fronto a[dv-] ersarius Sex[ti] 5 sit vanus neq ue loqui pos [s]it contra [S]extum ut F[r]onto fiat 10 mutus qu [m] access [e]rit

399

Sylloge of  Defixiones from the Roman West B c ̣onsular [e]m ut sit mutus ne que possit 5 loqui ne que qui[c-] quam ag[e-] re tanqu am nullo 10 ab inf[e-] ris

Bibliography: Exner 1938; García Ruiz 1967: no. 32; Solin 1968: no. 13; Schillinger-Häfele 1977: no. 122; AE 1978, 545; Reuter and Scholz 2004: 49–50; Fasold 2004: 31; Kropp 2008: no. 5.1.2/1; Blänsdorf and Scholz 2011: 70–77; ILGIL 594; Urbanová 2018: no. 70.

victims we find a Fronto. While it is not strange to find various curses directed towards one shared goal (see, for example, 125–26, 134–36, etc.), what is surprising is that these curses were found in different archaeological contexts. Perhaps the curses were not only deposited in different tombs, but also at different times?

Image source: Archäologisches Museum Frankfurt. See also App. IV.3, SD 478.

The notable linguistic and lexical aspects of the text are qum for cum (A, ll. 10–11) and consularis (B, l. 1), which is unparalleled in the corpus. According to Scholz, the practitioner was making reference to the provincial governor (legatus augusti pro praetore), whose job it was to dole out justice either in Mogontiacum or Nida. Indeed, it appears that Sextus was so frightened of the testimony of Marius Fronto, a Roman citizen, that he wrote two curses with the stated goal of silencing the victim. Note the similarities between sit vanus neque loqui pos[s]it (A, ll. 5–7) and sit mutus neque possit loqui (B, ll. 2–5) with the formula sint vani et ṃuti from 477 (A, ll. 4–5 and B, l. 8).

Commentary: this defixio was discovered in 1902 during the excavations of a necropolis located in FrankfurtPraunheim. Although we do not know the exact details about the tablet’s archaeological context, it seems that it was deposited in the tomb of a Roman soldier, which has been dated to c. 100 CE (for this question, see Reuter and Scholz 2004: 50). The curse was written on an irregularly shaped lead sheet, which has been broken into two parts that fit together. The tablet is in relatively good condition, though it is missing two thirds of the left edge and the top-right corner. An opisthograph, the text contains 23 lines, which run from left to right and were written in old Roman cursive. Based on its content, this text not only belongs among the juridical defixiones, but also seems to be closely related to 477: in both cases, the defigens is named Sextus and among the

Translation (Urbanová 2018: 78, modified): ‘I ask [you], Manes of the underworld, may Marius Fronto, the enemy of Sextus, be untrustworthy, may he be unable to speak against Sextus, may he become mute when he approaches the legate, may he be mute and unable to speak or do anything as if he were nobody in the underworld.’

479. Waldmössingen Provenance: aquatic context (?). Current Location: MUF Freiburg. Date: 75–200 CE. [Inv. No.: 1979-24-276-1] Material: lead. Measurements: 5.3 × 10.2 × 0.7 cm. Reading: A fib(u)lam Gnatae qui involavit aut qui melior est animi conscios ut illum 5 aut illam aversum faci ant di(i) sicut hoc est 400

 Germania B aversum et qui res illaeus sus tulit

Bibliography: Nuber 1984; Kropp 2008: no. 5.1.7/1; Faraone and Kropp 2010: 389–90; ILGIL 589; Urbanová 2018: no. 230.

Based on its content, the tablet should be grouped among the defixiones in fures, since we read about the theft of a fibula. The text presents the following orthographic features: fibla for fibula (A, l. 1), conscios for conscius (A, l. 4) and illaeus for illius (B, l. 2). The verb involo is quite apt in this context and is well attested in the curses from Britannia (e.g., 236 and 249).

Image source: Nuber 1984: 379, fig. 2. Courtesy of G. Seitz. See also App. IV.3, SD 479. Commentary: this defixio was discovered in 1979 during the work carried out at the reservoir located on the northeast side of Schramberg-Waldmössingen, near Rottweil (Arae Flaviae). The curse was found at the bottom of a hill among the materials that had been dredged up at a depth of 0.9 m. In an attempt to better understand the archaeological context, Nuber analysed a series of maps (from the year 1897) and found that there had been two fountains at the bottom of the above-mentioned hill. In his opinion, these fountains could have served as the place where the curse was originally deposited. This theory seems all the more likely, since investigations have not turned up the remains of any building in which curses are normally found, such as a necropolis or sanctuary.

The following phrases merit our attention: animi conscios (A, ll. 4–5), qui res illaeus sustulit (B, ll. 1–3) and ut illum aut illam aversum faciant di(i) sicut hoc est aversum (A, ll. 4–6 and B, l. 1). The first alludes to the thief’s accomplice or a witness, who is also a victim of the curse (for a similar idea, see 302 B, ll. 3–4: qui medius fuerit). The second phrase appears to target the new owner of the fibula; this interpretation is based on other parallels (see 121, 450 and 483) where tollo is used as a synonym for involo. Accordingly, this phrase returns to the thief, who has previously been mentioned (qui involavit, A, l. 2). For a punishment, the defigens asks ut illum aut illam aversum faciant di(i) sicut hoc est aversum (A, ll. 4–6 and B, l. 1). Here the text combines an all-inclusive formula with a persuasive analogy. After attempting to identify the robber, the defigens compares the victim to the way in which the text has been written. Given this comparison, the layout of aversum (B, l. 1) is even more complex than that of the rest of the inscription since it combines right to left writing with ‘mirrored’ and upside-down letters (see the ending -ersum). For parallels for the phrase aliquem aversum facere, see 464 A, ll. 2–3: pervese agas comodo hoc perverse scriptu(m) est.

The defixio was written on a lead sheet with the shape of a tabula ansata. This shape is common in writing practices and in this case was adapted for magical purposes (the only known parallels are 337, 463, 518, and perhaps 220). The curse was found in an excellent state of preservation. While the surface is a bit uneven and corrosion has damaged the ansae (especially on the left side), the tablet is practically whole. Below the first line at the height of the word Gnata, we can clearly identify a hole. The editor has argued that this hole is the result of corrosion rather than the act of piercing the tablet with a nail on the grounds that there is no oxidation or burrs along the hole’s perimeter.

Finally, the invocation of Di(i) (A, l. 6) could be taken as calling upon the Di Inferi, Di Manes, or even a more general way to summon a divinity, whatever its name. Gnata is a Latin cognomen attested in Gallia and Germania (see OPEL II, 168).

An opisthograph, the curse contains nine lines of text, which run from right to left. Nearly all of the letters have been written in the ‘usual’ direction, although some of them (e.g., the G in Gnatae, in A, l. 1) are ‘mirrored’ (i.e., from right to left), or upside down (see, for instance, aversum, in B, l. 1). The text was written in capitals that can be dated between the years 75 and 200 CE. The letters range from 0.3 to 0.9 cm in height on side A and from 0.3 to 0.6 cm on side B.

Translation (Urbanová 2018: 50, modified): ‘Whoever has stolen the brooch of Gnata or knows something more about it, may the gods render him/her averted, just as this (text) is averted; and the one who stole her thing.’

401

Sylloge of  Defixiones from the Roman West 480. Roßdorf, Kr. Darmstadt/Dieburg (Hessen) Provenance: necropolis. Current Location: Ober-Ramstadt Museum. Date: 2nd century CE. [Inv. No.: HM OR 1969: 21/27,7] Material: lead. Measurements: 3.8–2.5 × (4.5) × 0.15 cm. Reading: A Gn(aeus) Họr[at-] inị(us) [G]ṇ(aei) ḷ(ibertus) Arar icus B Flav[i]a Finita

Bibliography: Nuber 1996; AE 1996, 1172 a–b; Kropp 2008: no. 5.1.6/1; ILGIL 591.

The defixio was written on a roughly rectangular lead sheet, which still has its top, bottom and left edges intact, though they have been fractured. This is not the case with the right edge, which is now incomplete. The surface of the tablet has an off-white patina and has been severely corroded to the point that some letters are practically illegible.

Image source: Nuber 1996: 241. Courtesy of G. Seitz. Commentary: this defixio was discovered in 1969 during the excavations of a Roman crematory necropolis. Specifically, the curse was located on the edge of the necropolis about a half metre north-east of grave 27. Here archaeologists found a rather unique deposit comprised of an upside-down cinerary urn, inside of which there were burnt organic remains, an iron nail and the tablet currently under consideration. Unfortunately, we do not know exactly what type of burnt organic remains were inside the urn and hence cannot confirm, as the editor has suggested, that they belonged to the curse’s victims and served as an ousia. Entries 85 and 526 present interesting parallels for this curse, since each of these curses was found with ceramic vessels that also contained organic remains; to this list we may need to add a find from the sanctuary of Anna Perenna, where a small clay jar covered with a lead sheet was discovered (see Polakova and Rapinesi 2002: 39 and 49 for a discussion). Finally, note the presence of an iron nail alongside the curse under consideration: this was certainly an important tool in imprecatory rituals, even if it did not pierce the tablet.

An opisthograph, the inscription contains six lines of text, which run from right to left and were written in capitals that are very different in size and some of these letters have been written ‘mirrored’ (i.e., from right to left, such as l and n in B, ll. 1–2). While the lines on side A run from top to bottom, on side B they are organized from bottom to top (for parallels for this layout, see 134–36, 149 and 303). Based on palaeographic and onomastic considerations, the text dates to the second century CE. Such a dating coincides with the date of the urn in which the curse was deposited. The circumstances that led to the writing of this curse are unknown given that the inscription only includes the names of the victims. On side A, Nuber has proposed the reconstruction Gn(aeus) Họr[at]inị(us) Araricus. Gn(aeus) would come from the first two letters (g and n, l. 1, reading from right to left), Horatinius (ll. 1–2) would be a gentilic attested in CIL XIII, 8285 and Araricus (ll. 3–4) would be an uncommon cognomen (found at CIL XIII, 5711 and 5489). Side B contains a female personal name: Flav[i]a Finita.

481. Aventicum, Avenches Provenance: necropolis. Current Location: Roman Museum of Avenches. Date: 2nd–3rd centuries CE. [Inv. No.: 89/7852-9] Material: lead. Measurements: 8 × 12.5 cm. Reading:

Marius Cinne(m) suus et eum qi exin co(n)ciliavit aequa(m) a vita

402

 Germania Bibliography: Frei-Stolba 2002; AE 2002, 1055; Kropp 2008: no. 5.1.1/1; Urbanová 2018: no. 227.

concilio (l. 3, which can mean ‘serve as an intermediary’) and the word exin (l. 2, frequent in Latin verse). As FreiStolba has noted, ‘la structure de la proposition principale n’est pas claire, car le verbe co(n)ciliavit s’insère dans la subordonnée commençant par le pronom relatif qui’ (2002: 119). Following the same author, this structure can be explained by supplying the implied verb discedo, which together with the final a vita (l. 4) would take on the meaning ‘die’. This is quite common in Late Antiquity and undoubtedly apt in this context.

Image source: Aventicum-Site et Musée romains d’Avenches. Commentary: this defixio was still folded upon its discovery in the necropolis of Avenches-En Chaplix in 1989. The curse was found in the back of the so-called ‘northern funerary moment’, which has been dated to 28 CE thanks to dendrochronological analysis (see Flutsch and Hauser 2000).

Concerning the names found in the text, Frei-Stolba has interpreted Cinne(m) (l. 1) as a Gallo-Roman personal name without any known parallels. Kropp, however, has argued for different word division and hence she reads Marius Cinnesuus (l. 1), which would be a personal name in the nominative singular instead of an accusative. Although the names Cinnenius (see CIL XIII, 5191) and Cinneus (CIL VIII, 8548) are attested, the proposed Cinnesus does not have any known parallels. For this reason, we must proceed with caution.

The curse was written on a rectangular lead sheet, which lost its two left corners and has been severely damaged on the top-right side, of which only a small section has been conserved. The inscription contains four lines written in old Roman cursive (the maximum letter height is 0.7 cm) and runs from right to left. While most of the letters run from left to right, some of them have been ‘mirrored’ (i.e., they were written from right to left), such as c, q, s and r. Noteworthy linguistic and orthographic features include the following: Cinne for Cinnem (l. 1), suus for suum (l. 1), qi for qui (l. 2), cociliavit for conciliavit (l. 3) and aequa for aequam (l. 3). Regarding diction, note the use of the verb

After being inscribed, the tablet was folded and pierced with an iron nail, traces of which have been preserved on the lower-right angle of the defixio.

482. Groß Gerau Provenance: domestic context. Current Location: private collection. Date: 1st–2nd centuries CE. Material: lead. Measurements: 3.5 × 13.5 × 0.1 cm. Reading: A deum · maxsime Atthis tyranne totumque duodeca theum comme ndo deabus iniurium fas ut me vindic [e]tis a Priscil(l)a Caranti (filia) quae nuberi era5 vit pe[r] matrem deum vestrae {ut} [v]indicate sacra pater[na or -ni (?)] P[ri]scil(l)[a] pere[at]

403

Sylloge of  Defixiones from the Roman West B per matrem deum intra dies C cito vindicate numen vestrum magnum a Priscilla quae detegit sacra Pris cillam usqum nullam numero nu[p-] 5 sit gentem tremente Priscilla quam erante

Bibliography: Scholz and Kropp 2004; AE 2004, 1006a– b; Reuter and Scholz 2004: 70–71; AE 2006, 14; Scholz, Kropp and Blumer 2006; AE 2007, 148; Kropp 2008: no. 5.1.3/1; Versnel 2010: 300–05; ILGIL 588; Melzow 2013: no. 72; Urbanová 2018: no. 228.

for maxime (A, l. 1), Atthis for Attis (A, l. 1), deum for deorum (A, l. 1), Priscila for Priscilla (A, l. 4 and perhaps 6), eravit for erravit (A, ll. 4–5) and erante for errante (B, l. 7). Also noteworthy is the influence of Greek on the diction: tyrannus (A, l. 1; also found in 486), duodeca (A, l. 2; formed from a combination of Latin duodecim and Greek δυοδεκα), theum (A, l. 2; Latin transcription of Greek θεων). Concerning diction more broadly, the editors have highlighted the use of sacra (A, l. 6) which in conjunction with detegit this would mean ‘secret’; nullus (B, l. 4) is used as a synonym for ‘dead’ (cf. 477, where it has the same meaning); gens (B, l. 5) clearly used in a pejorative sense; and tremente (B, l. 5) which has a sexual connotation.

Image source: Scholz, Kropp and Blumer 2006: 182, fig. 2 and 183, fig. 3. Courtesy of M. Scholz. See also App. IV.3, SD 482. Commentary: this defixio was discovered in 1999 during an emergency excavation carried out at the vicus of GroßGerau. During this intervention, workers uncovered a small ditch near the entrance of a building dated to the reign of Trajan. Inside of the ditch, about 80 cm below the surface, was the defixio currently under consideration. It was found alongside some 100 stucco fragments, an as from the reign of Vespasian (see RIC, 482) and fragments of a tegula with the stamp [LEG] XIII GMV. The curse had been folded over five times. It is possible, as the editors have suggested, that the ditch where the tablet was deposited was part of the victim’s house. In that case, this defixio would be grouped with the few curses that have been found in domestic settings (such as 17, 127, 524, 528, etc. cf. section I.6.5.2).

As far as genre goes, this text belongs to the group of erotic defixiones. Here, the defigens denounces Priscilla for inurium fas (A, l. 3). She is handed over to Attis, Cybeles and totumque duodeca theum. The invocation of Attis and Cybele is suitable, given the context of love betrayed, which is clear in both the Ovidian myth and the defigens’ own situation. Furthermore, their mention has led the editors to posit a connection between this defixio and the collection from the sanctuary from Mainz. Also noteworthy is the invocation to the totumque duodeca theum, who are attested here for the first time in a Latin defixio. With this phrase, which almost resembles an all-inclusive formula, the practitioner seeks to secure the intervention of the principal deities of the Roman pantheon. As the editors have pointed out, deabus (A, l. 3, in place of deis) may refer to the Erinyes, who are summoned in the defixio 522.

The defixio was written on a strip of lead, which was broken into three pieces that all fit together. X-ray analysis reveals that this sheet was made from a small but thick piece of lead that was hammered out into a sheet. An opisthograph, the inscription contains 15 lines of text, which run from left to right. The line spacing is irregular: the space is greater on side B than on side A, where the space is gradually reduced between the lines. The text was written by a trained hand in a mixture of capitals and old Roman cursive (note the cursive shape of b, and both [cursive and capital] d, etc.), which can be dated to c. 100 CE.

Priscilla’s punishment ought to be doled out intra dies C (B, l. 1), which is a rather long time frame when compared to other texts (in Britannia the time limit is generally nine days; see 267, 340, 456, etc.). Also of note is the use of the patronymic (instead of the more usual matronymic) to identify the female victim. The use of the defigens’ name (in this case Paternus) is common in the agōgai. The name Priscilla is well-attested (A, ll. 4 and 7, B, ll. 3–5; cf. OPEL III, 162), Carantus (A, l. 4) is Gallic in origin and Paternus

The text is almost fully preserved (except for several letters that have been lost due to corrosion). Noteworthy linguistic and orthographic features include the following: maxsime 404

 Germania a well attested Latin cognomen (A, l. 6). After the curse was inscribed, the tablet was folded five times.

(?). By the Great Mother, may you punish (her betrayal of) the ancestral rites (or: sacred objects)! May Priscilla die!

Translation (Gordon 2012a: 10, slightly modified): ‘Greatest of all gods, Atthis, Lord, and all the Twelve Gods! I lay the wrong done me as a claim before the goddesses, that you may punish Priscilla, the daughter of Carantus, for me, who has done wrong in getting married

By the Mother of the Gods, may your grand divine power punish Priscilla within a hundred days, quickly, who betrays the sacred rites! I value Priscilla as nothing worth. She has married a worthless fellow(?), because Priscilla is as weak as she is unsteadfast (?).’

483. Groß-Gerau Provenance: unknown. Current Location: private collection. Date: 1st century CE. Material: lead. Measurements: 8.7 × 10.3 cm. Reading: 5



10



(H)umanum qui{s} sustulit Verionis palliolum sive res illius qui illius minus fecit ut illius mentes memorias dei`ci´ctas sive mulierem sive eas cuius Verionis res minus fecit ut illius manus caput pedes vermes cancer vermitudo interet membra medullas illius interet

Bibliography: Blänsdorf 2007a; AE 2007, 1049; Kropp 2008: no. 5.1.3/2; DTM, p. 183; ILGIL 13; Urbanová 2018: no. 229.

of intrare). Concerning the tablet’s diction, the editor has drawn attention to the expression minus fecit (ll. 4–5) meaning ‘to diminish’, which is only paralleled in 120. Also, note the abovementioned interet (ll. 11–12), which is first attested here with the meaning ‘attack’ or ‘infest’. Derived from vermes, the word vermitudo (ll. 10–11) is a neologism formed by analogy with third declension nouns like valetudo. The noun refers to the infestation of worms from which the curse’s victim ought to suffer. Here, the letter-space is much wider than in the rest of the text, as if someone would be dictating the word to the writer (Blänsdorf 2007: 64) or as if the author of the text was not sure about how to write it. Finally, note the use of the verb tollo (sustulit, l. 2) to refer to theft. This usage is common in the British defixiones against thieves (cf. 450).

Image source: Blänsdorf 2007a: 61. Commentary: this defixio was discovered in 2002 during construction work that unearthed the remains of a Roman settlement dated to the first century CE. The tablet, which was rolled up at the time of discovery, was found among the debris produced during construction. As a result, we do not know the exact context in which the curse was originally deposited. The defixio was written on a rectangular sheet of lead, which was corroded due to oxidation. The text was written by a trained scribe in capitals and contains 12 lines, which run from right to left. The letters (which measure 0.5 cm in height), were mostly written ‘mirrored’ (i.e., from right to left), though there are some exceptions, such as in l. 1 where only the letter n appears ‘mirrored’.

Accepting the interpretation of (h)umanum as a vulgarization of hominem (l. 1) rather than as a personal name (see Kajanto 1982²: 62, 222), this word would have served as the basis for an all-inclusive formula. Taken with sive mulierem sive eas (ll. 6–7), this catchall phrase would have allowed the defigens to identify the thief (for a parallel, see 249 B, ll. 1–4). Following these all-inclusive formulae, we find a list of the thief’s body parts that are cursed: first the text takes aim at the mentes memorias dei`ci´ctas (ll. 5–6) and then turns to the manus, caput and pedes (ll. 9–10) before finishing with the victim’s internal organs (ll. 11–12).

Judging from its content, we can place this text among the defixiones against thieves. Noteworthy linguistic and orthographic features include the following: quis for qui (l. 1); deicetas (l. 6; later corrected to deicictas) for deiectas; eas for eam (l. 7); cuius for qui (l. 7) and interet (ll. 11–12, from the verb interare, which is a vulgar form 405

Sylloge of  Defixiones from the Roman West This anatomic list appears to reflect the process of the theft: first the thief thought up a plan (hence the mind) and then carried it out (with his hands). As in English, the mention of the head and feet is a manner of referring to the thief’s entire body, including his innards and entrails. The punishment to be meted out consists of two terrible diseases: cancer and vermes, which were two difficult disseases to cure in Antiquity (Blänsdorf 2007: 64). Both of these illnesses, however, are hardly attested in the corpus of Latin defixiones: the first is also found in 146 and 147 ([---omi]ne cancer braca[[r]]rice persequiris…), while the second comes up in 491 (B, ll. 7–8: vermes adque alia portenta…). The name Verio

(ll. 2 and 8) is a cognomen already attested (see OPEL IV, 159). After being inscribed, the tablet was rolled up and likely deposited with the iron nail that is now stuck to the back of the sheet and protrudes past the lower edge. Translation: ‘The man who stole the little cloak of Verio or his things, who diminish them (i.e., his possessions), that his/her mind and memory be destroyed, or the woman or those who diminish the things of Verio, that in his/her hands, head, feet, the worms, the cancer, the worms penetrate, (that they) penetrate in his/her limbs and marrows.’

484. Mogontiacum, Mainz Provenance: sanctuary. Current Location: IMMH; TA. Date: 1st–first third of the 2nd century CE. [Inv. No.: 182, 16] Material: lead. Measurements: 10.5 × 10.2 × 0.07–0.1 cm. Reading: A PRI B (around the edges of the tablet) Prima Aemilia Nar cissi agat quidquid co nabitur quidquid aget omnia illi inver (on top of l. 4) 5 sum sit 6



10



(at the tablet’s midpoint) amentita surgat amentita suas res agat quidquid surget omnia interversum surgat Prima Narcissi aga como haec carta nuncquam florescet (under l. 1) sic illa nuncquam quicquam florescat

Bibliography: Blänsdorf 2004a: no. 112; AE 2004, 1024; Blänsdorf 2005a: no. 2; Kropp 2008: no. 5.1.5/4; Blänsdorf 2010b: 151–52, no. 5; Faraone and Kropp 2010: 383–84; Blänsdorf 2011: 66–67; DTM 15; Blänsdorf 2012b: 21; Blänsdorf 2015c: 33–34; Urbanová 2018: no. 91.

consisted of two zones: the sacred precinct, which was walled off and consisted of a temple and garden, and the surrounding field. Three small magical figurines were found outside of the temenos, while the curse tablets were deposited in several different locations within the enclosed zone. Accordingly, 24 of the tablets were found behind the temple inside a rectangular well that was divided into different sections. Inside, defixiones were found at various strata, the latest dating to the reign of Domitian (82 CE). On top of the well, an altar was built, which was subsequently replaced by two altars of the same size. Offerings were placed directly into the fire burning on the altar: many curse tablets were deposited in this way, some of which remain intact, while others show clear signs of having been placed in the fire. Furthermore, excavators found ‘remains of lead from the same features, melted into little

Image source: U. Weichhart apud DTM 15. Landesarchäologie Mainz. See also App. IV.3, SD 484. Commentary: in 1999 the excavations of the sanctuary of Isis and Mater Magna unearthed a cache of 34 new defixiones, which were deposited between the time of the Flavians and the reign of Trajan or Hadrian. This sanctuary was constructed on the site of an old cemetery (probably dating back to the Iron Age), ‘because the place was already considered sacred’ (Witteyer 2005: 106), and 406

 Germania lumps, [which] hint at the idea that originally they were also tablets, perhaps even figurines’ (Witteyer 2005: 116). In a niche adjacent to these altars, archaeologists found two more defixiones. Three further curses came from a small deposit located outside the sanctuary’s walls. The sanctuary was abandoned and covered over with a thick layer of tiles. In this stratum, there were also some coins from the reigns of Trajan and Hadrian, which has given us a terminus post quem of 125–28 CE for all the defixiones found in that stratum. Finally, we must note that in this same layer of tiles five other tablets were discovered (for this deposit, see especially DTM, pp. 1–6, Witteyer 2005 and 2013).

(e.g., DT 218, 227, 275–82), is completely unparalleled among the tablets from Mainz. We do not know the circumstances that led to the writing of this curse, which is directed against Prima Aemilia. The inscription on side B presents the following noteworthy linguistic and orthographic features: aga for agat (l. 11), como for quomodo (l. 11), nuncquam for numquam (ll. 12 and 13), and the agreement of quidquid with the plural omnia. Further, Blänsdorf (2005a: 14) has highlighted the lexeme amentita (ll. 6–7), which he translates as ‘out of one’s mind, mad’ and which would then be paralleled by amentia found at 48 (A, l. 27: obiicias illae amentiam). Another lexeme that deserves mention is carta, which is used to refer to the defixio itself (l. 11, see 96 for a parallel). The structure of the text is based on two similia similibus formulae: the first compares the victim to the text’s layout so that everything in her life be topsy-turvy, inversum sit (l. 4–5 and also ll. 7–9; see Faraone and Kropp 2010 for other examples). The second formula establishes a persuasive analogy between the victim and the tablet so that she can never thrive (ll. 11-14; florescet; see OLD 3). For other curses that draw analogies between the tablet and the victim, see Sánchez Natalías 2018: 12ff).

This defixio was written on a quadrangular sheet of lead and contains 18 lines of text, which were written in capitals (for a palaeographically similar text, see 215 and for a discussion see Blänsdorf 2012 b). The inscription is an opisthograph, though side A only contains the letters PRI (surely for Pri(ma)). Perhaps the author of the text was simply practising before writing the full curse on side B, where we encounter the same personal name. We must note the layout of side B, which is unparalleled among the other curses from Mainz: l. 1 begins in the top-right corner of the tablet and the next four lines run around the edges of the curse anti-clockwise, the last of which is above l. 4. The remainder of the inscription (ll. 6–14) runs from left to right and from top to bottom and is found in the centre of the tablet. However, the author did not accurately judge how much space (s)he needed for writing the entire curse. For that reason, the practitioner had to write the final two lines (ll. 13–14) right underneath l. 1, leaving a space of about 2 cm between ll. 14 and 6. This layout, which is slightly reminiscent of that of some North African curses

After being inscribed, the tablet was folded. Translation (Blänsdorf 2010b: no. 5): ‘(Whatever) Aemilia Prima, (the lover?) of Narcissus may do, whatever she attempts, whatever she does, let it all go wrong. May she get up (out of bed) out of her senses/mind. Whatever she strives after, may her striving in all things be reversed. May this befall Prima, the lover of Narcissus: just as this letter never shall bloom, so she shall never bloom in any way.’

485. Mogontiacum, Mainz Provenance: sanctuary. Current Location: IMMH; TA. Date: 1st century–first third of the 2nd century CE. [Inv. No.: 411, 4] Material: lead. Measurements: 6.5 × 6.5 × 0.08–0.12 cm. Reading: 5

Avita(m) noversa(m) dono tibi et Gratum [do]no tibi [---]+EMESMANT+[---]

407

Sylloge of  Defixiones from the Roman West Bibliography: Blänsdorf 2004a: no. 113; AE 2004, 1025; Blänsdorf 2005a: no. 4; Kropp 2008: no. 5.1.5/6; Blänsdorf 2008: no. 6; Blänsdorf 2010b: 150–51, no. 4; Blänsdorf 2011: 67-68; DTM 8; Blänsdorf 2015c: 30–31; Urbanová 2018: no. 88.

the mention of a noversa (l. 1, for noverca), it appears that there had been a sort of family feud. The editor has suggested the possibility that the defigens was a son from a previous marriage and became involved in a dispute with his stepmother. Accordingly, he employs the well attested formula dono tibi to consign Avita and  Gratus  (perhaps the stepmother’s son?) to Mater Magna (for the formula, see 451). The curse may contain a third victim in the final line, but the remaining text makes no sense. Noteworthy linguistic and orthographic features include noversa for noverca (l. 1, probably a result of writing from right to left), and the loss of final m in Avita(m) and noversa(m) (l. 1). Avita and Gratus are well Latin attested cognomina.

Image source: U. Weichhart apud DTM 8. Landesarchäologie Mainz. See also App. IV.3, SD 485. Commentary: this defixio, which was found in the sanctuary of Isis and Mater Magna (for the archaeological context, see 484), was written on a roughly quadrangular sheet of lead, which is now missing the lower-right corner as well as part of the bottom edge. The curse contains five lines of text, which run from right to left and were written in ‘mirrored’ capitals (i.e., from right to left: note the reversed b, r, etc.).

After the tablet was inscribed, it was folded. Translation (Blänsdorf 2010b: 169): ‘I give you Avita the stepmother; I also give you Gratus...’

Although we do not know the circumstances that led to the writing of this curse, based on its content and

486. Mogontiacum, Mainz Provenance: sanctuary. Current Location: IMMH; TA. Date: end of the 1st century CE. [Inv. No.: 201 B 36] Material: lead. Measurements: 11.5 × 10 cm. Reading: B neque se possit redimere nulla pe{r}cunia nullaque re neq(ue) abs te neque ab ullo deo nisi ut exitum malum 5 hoc praesta rogo te per ma iestatem tuam

A bone sancte Atthis tyran ne adsi(s) advenias Libera li iratus per omnia te rogo domine per tuum Castorem 5 Pollucem per cistas penetra les des ei malam mentem malum exitum quandius vita vixerit ut omni cor pore videat se emori prae10 ter oculos

408

 Germania Bibliography: Blänsdorf 2004b; AE 2004, 1026; Blänsdorf 2005a: no. 6; Blänsdorf 2005b: 681; Blänsdorf 2005c: 107–08; Kropp 2008: no. 5.1.5/2; Blänsdorf 2010b: 147– 50; no. 2; Versnel 2010: 305–09; Blänsdorf 2011: 71–72; DTM 5; Blänsdorf 2013: no. 77; Urbanová 2018: no. 85.

cistae penetrales refers to ‘the sacred boxes kept in the sanctuary that contained what were supposed to be Attis’ severed genitals’ (Blänsdorf 2010b: 148). The defigens asks that Liberalis be punished with a malam mentem, malum exitum quandius vita vixerit ut omni corpore videat se emori praeter oculos (A, ll. 6–10). That’s to say, he should suffer from madness and a slow and terrible death that he himself witnesses throughout his whole life (quandius vita vixerit [A, ll. 7–8]; cf. 250, B, l. 4: quoad vixerit and 350, ll. 3–4: usque die`m´ qụọ moriatur). The victim could not atone for his deeds and escape the curse through the help of his money or other valuables; rather, he simply must suffer an awful death (nisi ut exitum malum, B, ll. 1–4). The idea that one could only repay for their deeds through death is well attested (e.g., 491, 249, 271 and 304). To conclude the curse, the practitioner reiterates his petition (this is common in British curses); in this case, (s)he calls upon the deity’s majesty (see 128 and 365).

Image source: U. Weichhart apud DTM 5. Landesarchäologie Mainz. See also App. IV.3, SD 486. Commentary: this defixio, which was found in the sanctuary of Isis and Mater Magna (for the archaeological context, see 484), was written on a quadrangular sheet of lead. An opisthograph, the inscription contains 16 lines of text, which run from left to right and were written in an elegant old Roman cursive. As the editor (2010b: 149) has pointed out, the text employs a ‘rhetorical structure, consisting of a series of tricola and dicola’ and was likely written by a practitioner who came from high society. The circumstances that led to the writing of this curse against Liberalis are unknown. In the text, we note the following noteworthy features: the absence of final s in adsi(s) (A, l. 2), the loss of final m in vita(m) (A, l. 8), the hypercorrection of Atthis and the word tyranne (A, ll. 1–2, see 482 for a parallel). We must highlight the way that the defigens invokes the deity: bone sancte Atthis Tyranne (A, ll. 1–2; for a parallel, see 189, l. 1: bona san(c)ta nomen piạ…) and dominus (A, l. 4; cf. 121). The practitioner also insists that per omnia te rogo, per tuum Castorem, Pollucem, per cistas penetrales (A, ll. 3–6). Here, we find a solemn invocation of the deity and the mention of the

After being inscribed, the tablet was folded. Translation (Blänsdorf 2010b: 167): ‘Good, holy Att(h) is, Lord, help (me), come to Liberalis in anger. I ask you by everything, Lord, by your Castor (and) Pollux, by the cistae in your sanctuary, give him a bad mind, a bad death, as long as he lives, so that he may see himself dying all over his body —except his eyes. And may he not be able to redeem himself by (paying) money or anything else, either from you or from any other god except (by dying) a bad death. Grant this, I ask you by your majesty.’

487. Mogontiacum, Mainz Provenance: sanctuary. Current Location: IMMH; TA. Date: 1st century–first third of the 2nd century CE. [Inv. No.: 103, 2] Material: lead. Measurements: 4 × 11.3 × 0.04–0.07 cm. Reading:

Trutmo Florus Clitmonis filius

Bibliography: Blänsdorf 2005a: no. 1; AE 2005, 1128; Blänsdorf 2008: no. 3; Kropp 2008: no. 5.1.5/3; Blänsdorf 2010b: 156, no. 10; Blänsdorf 2011: 66; DTM 21; Urbanová 2018: no. 86.

figurine made of clay (for the archaeological context, see 484). The figurine has an erect penis and was pierced some eight times. This figurine was intentionally broken at the waist and each half was carefully deposited next to the defixio so that one half was lying prone and the other supine (see Witteyer 2013: 328–31). This was, perhaps, a further attempt to destroy the victim (Blänsdorf 2010: 156). The curse, which was written on an irregularly shaped sheet of lead, contains three lines of text, which run from left to right and were written in capitals by a

Image source: U. Weichhart apud DTM 21. Landesarchäologie Mainz. See also App. IV.3, SD 487. Commentary: this defixio was found in the Isis and Mater Magna sanctuary together with a small magical

409

Sylloge of  Defixiones from the Roman West trained hand. The circumstances that led to the writing of this curse are unknown since the text only contains the victim’s name, which is qualified by a patronymic (for other examples, 215, 235, 362, etc.). Though the text does not proffer much information, Blänsdorf has suggested, ‘[t]he erect phallus may indicate that the defixio was an erotic one’ (2010b: 156). The personal

names found in the text are undoubtedly Germanic, although Trutmo also had a well-known Latin cognomen (Florus). After being inscribed the tablet was rolled up. Translation: ‘Trutmo Florus, son of Clitmo.’

488. Mogontiacum, Mainz Provenance: sanctuary. Current Location: DL; GKE. Date: 1st century–first third of the 2nd century CE. [Inv. No.: 182, 14B] Material: lead. Measurements: 4 × 7.5 × 0.05 cm. Reading: A mando et rogo religione ut man data exagatis Publium Cutium 5 et Piperionem et

B Placida et Sacra filia eius sic illorum membra liquescan(t) quatmodum hoc plum5 bum liquescet ut eo ru exsitum sit

Bibliography: Blänsdorf 2005a: no. 3; Blänsdorf 2005b: 672; Blänsdorf 2005c: 108–09; AE 2005, 1127; Blänsdorf 2008: no. 11; Kropp 2008: no. 5.1.5/5; Blänsdorf 2010b: 156–57, no. 12; Blänsdorf 2011: 70; DTM 11; Blänsdorf 2012b: 21–22; Blänsdorf 2013: no. 79; Blänsdorf 2015c: 35; Urbanová 2018: no. 488.

and Piperion’ (2010b: 157). As far as diction is concerned, note the phrase mando et rogo religione (A, ll. 1–2), which is unparalleled. According to the editor (2010b: 157), the use of religio here is meant to underscore that the defigens’ request is ‘in accordance with the rules of a religious prayer’ (on this, cf. also Veale 2017: 286–87). The curse concludes with a similia similibus formula that draws a persuasive analogy between the death of the victims and the way in which the tablet was deposited: the victims’ limbs are to liquify in the same way that lead melts (cf. the curse from Montfo, which was found inside a well; see 157 A, ll. 1–4: qomodo hoc plumbu(m) non paret et decadet, sic decadat aetas, membra, vita; for other such analogies, see Sánchez Natalías 2018: 12ff.). Whereas the personal names Cutius, Sacra and Placida are well attested Latin names, Piperon is not. After being inscribed, the tablet was rolled up and thrown into the temple’s sacred fire along with 498 (whose text ought to be taken as a continuation of this curse).

Image source: U. Weichhart apud DTM 11. Landesarchäologie Mainz. See also App. IV.3, SD 488. Commentary: this defixio, which was found in the sanctuary of Isis and Mater Magna (for the archaeological context, see 484), was written on a roughly rectangular sheet of lead, whose lower edge has been lost. An opisthograph, the curse contains 11 lines of text, which run from left to right and were written in old Roman cursive by a trained hand (see Blänsdorf 2012 b for a palaeographic discussion). Noteworthy orthographic features include the following: loss of the final -m in Placida(m), Sacra(m), filia(m) (B, ll. 1–2) and eoru(m) (B, l. 6), the omission of the final t in liquescan(t) (B, l. 3) and the hypercorrection in exsitum (B, l. 6; for exitum).

Translation (Blänsdorf 2010b: 178): ‘I hand over (to you), and, observing all ritual form, ask that you require from Publius Cutius and Piperion the return of the goods entrusted to them. Also Placida and Sacra, her daughter: may their limbs melt, just as this lead shall melt, so that it shall be their death.’

Judging from the content, Blänsdorf has argued that a theft motivated the writing of this curse and hence calls for ‘the return of the goods that had been entrusted to P. Cutius 410

 Germania 489. Mogontiacum, Mainz Provenance: sanctuary. Current Location: IMMH; TA. Date: 1st century–first third of the 2nd century CE. [Inv. No.: 109, 5] Material: lead. Measurements: 13.5 × 6 cm. Reading: A Fo[r]tunam dolus q[u]otti[die]I sed or q[u]ot ti[bi] sed vir pa(tri?)+++deo meo I meo +V [sp]oliav[it] 5 IVNCNAO (vacat) REIANTI

B mentem memori am cor cogita tum il[le q]uisquis patrem meum con[s]p[exit?] 5 illi et eius (or sui te illi)

Bibliography: Blänsdorf 2005a: no. 5; Blänsdorf 2008: no. 13; AE 2008, 978; Kropp 2008: no. 5.1.5/7; Blänsdorf 2010b: 156, no. 11; Blänsdorf 2011: 66–67; DTM 16; Urbanová 2018: no. 92.

the inscription contains 10 lines of text, which mainly run from right to left (except for the first two lines of side A) and was written in capitals. In addition, some of the letters are ‘mirrored’ (i.e., written from right to left, such as e in B, l. 1, the final s in B, l. 3, the final s in B, l. 4). We do not know the circumstances that led to the writing of this curse, in which the victim’s mental abilities were systematically targeted (mentem memoriam cor cogitatum, B, ll. 1–3; for this sequence, see 339 and 349).

Image source: U. Weichhart Landesarchäologie Mainz.

apud

DTM

16.

Commentary: this defixio, which was found in the sanctuary of Isis and Mater Magna (for the archaeological context, see 484), was written on a roughly rectangular sheet of lead, which has lost its upper-right corner and part of its lower-left edge. The central section of the inscription has been severely damaged by corrosion resulting from a thick iron nail that pierced the defixio. An opisthograph,

After being inscribed, the tablet was pierced by the abovementioned iron nail. Translation (side B only): ‘Mind, memory, heart, thought. Whoever fixed his sights (?) on my father...’

490. Mogontiacum, Mainz Provenance: sanctuary. Current Location: IMMH; TA. Date: end of the 1st century CE. [Inv. No.: 111, 53] Material: lead. Measurements: 5.5 × 12 × 0.05 cm. Reading: A Tiberius Claudius Adiutor in megaro eum rogo te Ma t[e]r Magna megaro tuo re cipias et Attis domine te 5 precor ut hu(n)c (h)ostiam accep tum (h)abiatis et quit aget agi nat sal et aqua illi fiat ita tu facias domna it quid cor eoconora c(a)edat (?) 411

Sylloge of  Defixiones from the Roman West B devotum defictum illum menbra medullas {AA} nullum aliud sit 5 Attis Mater Magn

Bibliography: Blänsdorf 2005a: no. 7; AE 2005, 1125; Blänsdorf 2005b: 683–85; Blänsdorf 2005c: 103–04; Kropp 2008: no. 5.1.5/8; Blänsdorf 2010b: 173–75, no. 8; Blänsdorf 2011: 68–69; DTM 4; Blänsdorf 2013: no. 76; Blänsdorf 2015c: 32–33; Urbanová 2018: no. 87.

un affranchi impérial employé dans l’administration’. In favour of this hypothesis concerning who is the curser and who is the cursed, we can adduce parallels such as 215 (ll. 1–5), in which we find a similar opening phrase: Docilianus Bruceri deae sanctissim(a)e Suli devoveo eum [q]ui… Accordingly, the proposal of the AE editors appears probable.

Image source: U. Weichhart apud DTM 4. Landesarchäologie Mainz. See also App. IV.3, SD 490.

The practitioner asks Mater Magna to accept the victim in her temple (megaro tuo recipias, A, ll. 3–4). As the editor observes, ‘this act of “receiving” (…) connotes the surrender of the victim to the god, and is in fact the curse that aims to produce the target’s death’ (Blänsdorf 2010b: 153–54). Next, the defigens employs a similia similibus formula to establish a persuasive analogy between the victim’s future deeds and salt: he asks that they fail as surely as salt must dissolve in water. A second hand wrote side B, where the curse takes aim at different parts of the victim’s body (menbra and medullas, B, ll. 2–3) and then evokes Attis and Mater Magna (B, l. 6).

Commentary: this defixio, which was found in the sanctuary of Isis and Mater Magna (for the archaeological context, see 484), was written on an irregularly shaped sheet of lead. An opisthograph, the text contains 14 lines, which run from left to right and were written in capitals by two poorly trained hands, who wrote lines on side A to keep their text straight (see 503 for a parallel). The circumstances that led to the writing of this curse are unknown. Noteworthy linguistic and orthographic features include the following: huc for hunc (A, l. 5), ostiam for hostiam (A, l. 5), abiatis for habeatis (A, l. 6), domna for domina (A, l. 8) and menbra for membra (B, l. 2). As far as diction is concerned, we must highlight the use of the Greek lexeme megaron (A, ll. 2–3) to refer to the temple. Also note the lexeme eoconora (A, l. 8), the vulgar form of iecinora, as well as the phrase acceptum habere (A, ll. 5–6; cf. 1), which is frequent in commercial contexts. According to the editors of AE (contra Blänsdorf 2005a: 18, 2005b: 684, 2005c: 104, 2010a: 153 and DTM 4), Tiberius Claudius Adiutor is the defigens, rather than the victim. Furthermore, they suggest that he was ‘peu-têtre

After the text was inscribed, the sheet was folded. Translation (Blänsdorf 2010b: no. 7, modified): ‘I, Tiberius Claudius Adiutor, ask you, Mater Magna, to receive him in your temple. And Lord Attis, I ask you that you may credit him as a sacrifice to your account (i. e. enter him in your accounts under “Offerings”); and whatever he does or busies himself with, may it become salt and water for him. Him cursed and “caught” —in his limbs, strength – let there be nothing else— Attis, Mater Magn(a).’

491. Mogontiacum, Mainz Provenance: sanctuary. Current Location: IMMH; TA. Date: 1st century–first third of the 2nd century CE. [Inv. No.: 72, 3] Material: lead. Measurements: 26.7 × 12.3 × 0.16 cm. Reading:

412

 Germania A Mater Magna te rogo p[e]r tua sacra et numen tuum Gemella fiblas meas qualis sustulit sic et illam REQVIS (rogo (?)) 5 adsecet ut nusquam sana si[t] quomodo galli se secarunt sic ea(m (?)) velis nec se secet sic uti planctum ha[be]at quomodo et sacrorum deposierunt 10 in sancto sic et tuam vitam valetudinem Gemella neque hostis neque au ro neque argento red imere possis a matre 15 deum nisi ut exitum tuum populus spectet. Verecundam et Patern nam sic illam tibi com mendo Mater Deum 20 Magna rem illorum in AECRVMO DEO VIS qua le rogo co(n)summent[u]r quomodo et res meas vire sque fraudarunt nec se 25 possint redimere nec hosteis lanatis

B nec plum{i}bis nec auro nec ar gento redimere a numine tuo 5 nisi ut illas vorent canes vermes adque alia portenta exitum quarum 10 populus spectet tamquam quae {C} FORRO or MO L+auderes comme[ndo] duas TAMAQVANIVCAVER{S}SO 15 scriptis istas AERIS+ADRICIS+S+LON A+ illas si illas cistas caecas aureas FECRA E[-c.2-]I[-c.1-]LO[-c.2-]AS 20 OV[-c.1-]EIS + mancas A

413

Sylloge of  Defixiones from the Roman West Bibliography: Blänsdorf 2005a: no. 8; AE 2005, 1124; Blänsdorf 2005b: 677–80; Blänsdorf 2005c: 104–07; Kropp 2008: no. 5.1.5/9; Blänsdorf 2010b: 159–61, no. 17; Blänsdorf 2011: 78–81; DTM 1; Burrer 2013: no. 73; Blänsdorf 2015c: 38–39; Urbanová 2018: no. 231.

of vires, see OLD 26). As the curse goes on to explain, the victims are to suffer a violent death in which they are eaten by canes, vermes adque alia portenta (B, ll. 6–8; for the worms, see 483). Blänsdorf has pointed out that this death ‘can be explained by some parallels from Herodotus, the Bible and the history of pagan persecutors of Christian martyrs: wicked men are killed by worms eating their entrails’ (2010b: 161).

Image source: U. Weichhart apud DTM 1. Landesarchäologie Mainz. See also App. IV.3, SD 491. Commentary: this defixio, which was found in the sanctuary of Isis and Mater Magna (for the archaeological context, see 484), was written on a rectangular sheet of lead, which is almost entirely preserved and in good condition (as I verified during my autopsy in March 2018). An opisthograph, the inscription contains 46 lines (26 on side A and 20 on side B), which run from left to right. According to Blänsdorf, two different hands wrote the text in capitals (but note the old Roman cursive shape of b, in A, ll. 3 and 18; and d in B, 13). Following the editor, their hands can be distinguished by different calligraphic styles as well as by the size of their letters (on side A the letters measure about 0.5 cm, whereas on side B they are 1 cm in height). In my opinion, it is also possible that there was just a single one author, who got too tired of writing after inscribing side A (producing bigger letters on B, with a wider style). It is worth noting that both sides present an uncommon indentation.

Both of these curses were written with the expressed desire that the victims could not avoid their due punishments by making an offering to the goddess: neque hostis, neque auro neque argento redimere possis (contra Gemella, A, ll. 12–14) and nec se possint redimere, nec hosteis lanatis, nec plum{i}bis, nec auro nec argento (contra Verecunda and Paterna, A, ll. 24–26 and B, ll. 1–3). Furthermore, the three women are to be publicly executed: exitum tuum populus spectet (contra Gemella, A, ll. 15–16) and exitum quarum populus spectet (contra Verecunda and Paterna, B, ll. 9–10). This public punishment, which resembles those given of condemned criminals, finds some parallels in 493 and 128. Finally, we must highlight the use of the lexeme exitus to refer to death (A, l. 15 and B, l. 9), which we find instead of the more common mors (cf. 486, 488 and 493); also note the cistas caecas, aureas, sacras (B, ll. 17–18), which may be equivalent to the cistas penetrales from 486.

Based on content, the curse belongs to the group of defixiones against thieves. The text actually contains two different curses that are directed at three different victims. In both cases, the author follows the same outline, in which (s)he first evokes Mater Magna, then describes the theft and finally asks for the wrongdoers to be punished without any possibility of escaping punishment.

After being inscribed, the curse was rolled up. Translation (Blänsdorf 2010b: 184–85, modified): ‘Mater Magna, I ask you by your sanctuary and your divine force. Gemella who stole my brooches (I ask you) may she (…) so that no part of her be healthy. Just as the galli have cut themselves, so (may) she want to do. And may she not cut herself so, that she may lament herself. As they have deposited the holy things in the sanctuary, so also your life and health, Gemella. Neither by offerings nor by gold nor by silver may you be able to redeem yourself from the Mater Deum, except that the people may watch your death. — Verecunda and Paterna: for thus I give her to you, Mater Deum Magna, their property (…) I ask they may be destroyed just as they have defrauded me of my property and resources, nor may they be able to buy themselves free either by offering sheep or by lead (tablets); neither by gold nor silver may they buy themselves free from your divine power, until dogs devour them, worms and other horrible things; may the people watch their death just as… two (…) with writings (…) them, if [someone takes?] those hidden golden holy boxes… holy ones.’

The first curse targets Gemella (A, l. 3), who is accused of stealing a fibula (fiblam, A, l. 3; for a parallel, see 479). Her punishment is laid out in two similia similibus formulae: the victim should suffer from an open wound so that she bleeds out just as the ‘the self-castrated galli, the frenzied followers of Mater Magna’ suffer from their injuries (Blänsdorf 2010 b: 159). Continuing with the same imagery, the text next claims that just as the galli deposit sacred objects in the temple, so too must the victim deposit her vitam and valetudinem (for this type of offering, see 451). The second curse is directed against Verecunda and Paterna (A, ll. 17–18), who are brought before the goddess and accused of ripping off the defigens, who claims that res meas viresque fraudarent (A, ll. 23–24; for this meaning

414

 Germania 492. Mogontiacum, Mainz Provenance: sanctuary. Current Location: DL; GKE. Date: 1st century–first third 2nd century CE. [Inv. No.: 1, 29] Material: lead. Measurements: 9 × 16 × 0.2 cm. Reading: A rogo te domina Mater Magna ut tu me vindices de bonis Flori coniugis mei qui me fraudavit Ulattius 5 Severus quemadmo[dum] hoc ego averse scribo sic illi

B omnia quidquid agit quidquid aginat omnia illi aversa fiant ut sal et aqua illi eveniat quidquid mi abstulit de bonis 5 Flori coniugis mei rogo te domina Mater Magna ut tu de eo me vindices

Bibliography: Blänsdorf 2005a: no. 9; AE 2005, 1122; Blänsdorf 2005b: 672–73; Blänsdorf 2005c: 102–03; Kropp 2008: no. 5.1.5/10; Blänsdorf 2010b: 152–53, no. 7; Faraone and Kropp 2010: 382–83; Blänsdorf 2011: 70– 71; DTM 3; Blänsdorf 2013: no. 75; Blänsdorf 2015c: 40; Urbanová 2018: no. 233.

is common: cf. 365 and 486). The curse is built around two similia similibus formulae: the first compares the way that the curse is written to the victim, his life and deeds (A, l. 6, B, ll. 1–2). Here we find the phrase averse scribo, which does not refer to writing from right to left, but rather to ‘the hostility with which she writes (…), which is in itself to affect Severus and everything he does or tries to do’ (Blänsdorf 2010b: 153; for parallels, see 464 and perhaps 532 among others). The second analogy, ut sal et aqua illi eveniat, compares the dissolution of salt in water to the futility of the victim’s future deeds (B, l. 3, see 490 and perhaps 493). The cognomen Florus is well attested, while Ulattius is a personal name of Celtic origin (derived from the root