Greek Curse Tablets: What People Wrote to the Gods

Greek curse tablets were usually thin sheets of lead inscribed with the names of people someone wanted to silence, restrain, defeat or separate from a partner. The Greek word katadesmos refers to binding; scholars also use the Latin term defixio, especially when discussing the wider Greek and Roman evidence. The earliest surviving Greek tablets date to the late sixth and fifth centuries BC, although different forms of the practice continued into Roman antiquity. Their writers generally did not intend other living people to read them.

The tablets frequently ended up in graves, wells and other places thought to give access to the world below. They concerned lawsuits, marriage, trade and theft. One writer might inscribe only a set of names; another could address spirits or gods and specify how an opponent’s hands, voice, judgement or personal relationships should be affected. Such private requests reveal conflicts and worries that seldom find their way into official inscriptions or the narratives of ancient historians.

The words record what a writer wanted, but seldom reveal what subsequently happened to the target. Even when the intended victim is named, the person who ordered or wrote the spell often remains unidentified. The differences between short classical name lists and complex Roman-period spells also make a single account of ancient ‘magic’ inadequate. Wording, archaeological setting and the treatment of the lead offer firmer evidence about individual acts.

Where Greek Curse Tablets Came From

Some of the earliest surviving Greek binding tablets come from western Sicily, notably Selinous and Himera, where Greek settlements existed alongside other communities with their own civic and religious lives. The inscriptions often give little beyond personal names. Jessica Lamont’s study of the Sicilian evidence points out that early examples sometimes have none of the explicit commands, divine appeals or verbs meaning ‘bind’ found on later tablets. Writing down the names may itself have been the essential operation.

By the fourth century BC, tablets from Athens and Macedonia display more varied language and methods of deposition. Litigation and trade produced disputes, while obligations between workshops and households could generate conflicts of their own. A speaker in a lawsuit might fear an eloquent opponent; a trader could worry about customers; a woman could see her hopes for marriage threatened. Binding a rival offered a private response. The finds do not tell us that curses were equally widespread in every community, or that everyone who used one hired a specialist.

The English term ‘curse tablets’ covers several ancient practices. Fritz Graf’s discussion of Greek cursing distinguishes an ara, a spoken curse or imprecation, from a katadesmos, an act of binding. Some inscriptions ask for an enemy to suffer, others try to stop a person acting, and others place a complaint before a god. The grievance of someone asking a deity to punish a thief differed from that of someone trying to ruin an opponent’s court case, even if both inscribed their requests on lead.

Why Lead, and What Happened to the Tablet?

Lead was practical for the task. A thin sheet could be beaten out of scrap and scratched with a pointed instrument. Once buried or submerged, it also had a better chance of surviving than papyrus or other perishable writing materials. Its durability partly explains why archaeologists encounter so many lead tablets. It cannot show that every magical text was written on the metal, still less that the heaviness of lead meant the same thing in every ritual. The local context and the actual words must guide any symbolic interpretation.

After inscribing the lead, its maker might roll or fold it, pierce it with a nail or deposit it beside other objects. Where the inscription asks for someone to be fixed or bound, such treatment may have reinforced the words. A nail hole alone, though, does not tell us what its maker intended. Some tablets were left flat; others had been folded so tightly that conservators had to open them to expose the writing. Creases and the direction of the letters can occasionally help establish whether the metal was inscribed before it was folded and deposited.

The place where a tablet was left could also matter. A grave might bring an appeal to the dead, whether directly or through the dead as intermediaries. A well provided a way below the surface; a sanctuary could associate the request with a particular sacred place. These explanations come from find patterns and surviving invocations, since most writers did not explain why they chose a location. Nor is evidence from a disturbed layer as secure as a find recorded in an intact tomb.

Figurines, Nails and the Body of the Target

Athenian binding practices could involve small figures as well as inscribed sheets. From the Kerameikos comes a male lead figurine enclosed in a miniature lead coffin, accompanied by a judicial curse against named opponents. It dates to about 400 BC. The University of Hamburg’s study of the object explains why its history is difficult to reconstruct: parts of the little coffin were recovered separately, and scholars dispute the sequence in which they were deposited or disturbed. The finds cannot sustain a precise account of one uninterrupted ceremony.

A restrained figurine could represent a person whom the writer wanted to restrain, particularly when an accompanying inscription names legal opponents. But that parallel does not make nails, figures and tablets interchangeable ritual objects. Archaeologists have found these things in the wider body of evidence for ancient rites. They cannot be supplied to an individual curse unless the excavation or inscription actually records them. Both the objects and the inscriptions need to be interpreted in their own archaeological contexts.

Graves, Wells and the People Below

A well excavated by the German Archaeological Institute in the Kerameikos cemetery at Athens contained numerous lead tablets. Jutta Stroszeck’s 2019 report on Well B 34 identifies a change in the pattern of deposits. In the classical period tablets often went into graves; wells became prominent among later Athenian finds. Stroszeck suggests that late-fourth-century BC changes to cemetery rules may have helped redirect these deposits. Her proposal explains a shift in the archaeological evidence rather than establishing a universal legal rule about cursing.

Excavations in the Athenian Agora have produced several groups of tablets. Jaime Curbera and Jessica Lamont published 25 fourth-century BC lead tablets from the area beneath the Tholos, the southern industrial quarter and a well near the Crossroads Enclosure. Twenty-four were curses. The Tholos had an administrative role in Athenian government, but finding curses nearby does not identify officials as their patrons. It does show that acts generally considered private left traces close to the places where Athens conducted its civic and commercial business.

Roman-period finds from the Agora also record disputes involving money and personal status. In David Jordan’s publication of 30 Roman-period curses, 22 tablets came from a well near the civic offices. The surviving complaints and instructions mention enslaved people, foreigners and unpaid loans, among other concerns. If a curse refers to an enslaved person running away, it gives us the claimant’s wish to exert control. It is not independent confirmation of what the enslaved person did. The same problem arises whenever a curse-maker’s accusation is taken as an objective account.

A Woman's Marriage Curse at Pella

An inscription discovered at Pella in 1986 gives this material an unusually personal voice. Pella was a major centre of the Macedonian kingdom. The writer wants to stop Dionysophon marrying Thetima and have him instead for herself. She calls upon Makron, a dead man, and associated spirits. The marriage is to remain blocked unless she herself recovers and unfolds the lead. Her instructions refer to a particular contest over marriage and attachment, rather than merely naming people to be bound.

She also asks to grow old beside Dionysophon. Thetima is the immediate obstacle, although the curse extends to any other woman he might marry. Damage to the writer’s name has prompted uncertain restorations. Nothing surviving establishes whether she was a wife, a lover or a potential bride. Emmanuel Voutiras, who studied the inscription, warned that her relationship with Dionysophon cannot be assigned a definite legal or social status. Her words establish what she wanted, but do not prove that Dionysophon had abandoned her or reveal the rest of her life.

The Pella inscription matters for its Greek as well. Its dialectal features are associated with Macedonia and are often called Doric or Northwest Greek. They have been brought into debates over what the ancient Macedonians spoke, although a single tablet cannot resolve the linguistic history of a kingdom. Unlike a state decree, it records language used in an intimate dispute. That gives its spellings and expressions particular value alongside the literary Attic Greek more familiar from surviving texts.

The Nemea Tablets and the Limits of Interpretation

The sanctuary of Opheltes at Nemea produced several folded, pierced sheets of lead, including four with legible erotic curses. Opheltes was a child hero whose death figured in the traditional account of the Nemean Games’ beginnings. A shrine dedicated to someone said to have died young would have particular associations for appeals involving the dead. The sheets were deposited at a hero’s sanctuary, rather than in an ordinary grave.

Interpreting the words is difficult. Jorge J. Bravo III argues that the Nemea spells are more ambiguous than a straightforward wish to separate lovers. Their lists of body parts and the verb apostrephō, meaning ‘turn away’, allow different grammatical readings. Depending on how the uncertain subjects and commands are read, the speaker may be asking to redirect affection, turn away a body or disrupt an existing relationship. Repetition in a spell does not remove these problems of interpretation.

Folding and piercing were clearly part of the history of these tablets, although no accompanying instructions tell us in which order every action took place. Excavation establishes that the sheets were left at Opheltes’ shrine. Connecting that choice to beliefs about a child hero and contact with the untimely dead is a reasoned interpretation based on his cult and other deposits. The general aims of the erotic spells are clearer than the full set of religious assumptions held by their authors.

What People Wrote About Court Cases

In Athenian lawsuits, the spoken argument carried considerable weight. Citizens usually presented their own cases to jurors, supporting them with witnesses and documentary testimony. An opponent able to persuade a jury posed a direct threat to the outcome of a case. Judicial curses seek to suppress tongues, memories or the capacity to present an argument; some also name witnesses and allies alongside the principal adversary. Their authors were trying to influence the relationships and testimony on which a case might turn.

An Athenian tablet, dated by the National Archaeological Museum to approximately 345–335 BC, contains 111 names identifying at least 96 people. Its binding formulas target words and actions. Citizens of standing appear alongside women, resident foreigners and people in trades. The museum raises legal, political, commercial and associational explanations, without choosing a proven motive. Such a varied list cannot securely be assigned to a routine court dispute.

The list records a range of people who seldom appear together in elite literary accounts. A citizen may be marked by a demotic, the name of the deme, or local civic district, to which he belonged. Traders and craftsmen can appear where elite authors offer little information about their work. But a shared place on a curse does not establish friendship, an alliance or conspiracy among its targets. The writer may have assembled a group for the purpose of binding people whose actual relationships were quite different.

Trade, Rivalry and the Problem of Motive

References to occupations, workshops and profits on other tablets may reflect economic disputes. A merchant might want customers to desert a competitor, or a craft worker might hope to disrupt an order. These are possible motives, seldom complete accounts supplied by the inscriptions themselves. Philip Venticinque’s study of commercial curses collects roughly sixty texts concerning trades, shops, workshops or profits, mainly dating from the fifth through third centuries BC. Venticinque cautions that a reference to trade does not necessarily establish a direct contest for customers.

Members of a household and workers in different trades could appear together in one curse. Their connections might have involved debts, witnesses, marriage ties or personal reputation, as well as buying and selling. Credit and guarantees often depended on other people, so a disagreement in a workshop could reach a court or household. The text records what its author wanted stopped, but usually says little about the quarrel’s beginnings. Where only names survive, a precise story about a violated contract or damaged business cannot be recovered.

Theft, Injury and Appeals for Justice

Some lead appeals arose from a stated injury, particularly the theft of property. Scholars use the expression prayers for justice for petitions in which someone asks a deity to expose or punish the wrongdoer. They can resemble binding spells in material and location, yet the petitioner presents the request as redress for an offence. Such texts preserve the petitioner’s accusation and request for divine intervention without independently proving the alleged offence.

A great deal of the familiar evidence for theft petitions belongs to Roman Britain rather than classical Greece. One British Museum lead sheet from Gloucestershire asks Mercury to avenge stolen gloves. Its date and language place it in a different historical setting. It would be misleading to use it as proof that classical Athenian curses all appealed to justice. The line between an attack on another person and a request for rightful punishment could be contested in antiquity and remains difficult for historians to draw.

Love, Separation and Coercion

Love spells also pursued different kinds of control. Some sought to end a couple’s relationship, some to prevent a wedding, and others to bring a desired person to the petitioner. The convenient modern labels ‘separation’ and ‘attraction’ do not always fit an individual text. At Pella, the writer wishes to separate Dionysophon from Thetima precisely so that he can be hers. At Nemea, the direction of certain instructions is disputed. These texts record particular attempts to interfere in others’ choices, rather than one settled ancient doctrine of romantic magic.

Some of the demands were explicitly coercive: writers wanted a target deprived of sleep, taken away from an existing partner or compelled to submit to their wishes. The people named rarely leave accounts of what they felt or chose, and their consent cannot be assumed. Yet the spells also register the fears of petitioners who faced uncertain marriage arrangements, social exclusion or financial insecurity. We learn a good deal about the writer’s predicament while hearing little from the person on the receiving end.

Roman-era Greek and Egyptian magical texts sometimes give elaborate directions for compelling desire, including the use of figurines, divine names and prescribed actions. They preserve traditions of ritual instruction from a later period. They do not justify supplying a detailed Roman Egyptian procedure to every fourth-century BC lead sheet. A late-antique papyrus and a classical curse can be related without sharing their theology, their instructions or the people for whom they were written.

The Gods, the Dead and Other Powers

Invocations could call on Hermes in his underworld aspect, Persephone, Hecate and other powers associated with death or transition. Some tablets call upon the dead, daimones or agents that cannot easily be identified. The Greek term daimōn had meanings that varied across contexts and is not equivalent to a demon in Christianity. The Greek understanding of the afterlife also changed between authors and communities. Other spells name no divine power at all, which prevents us from treating every tablet as a letter addressed to a particular god.

At graves, the dead might be called upon as intermediaries. Some traditions gave special potential to those who had died too soon, but this was not necessarily the explanation for every grave deposit. At Nemea, the tablets lay at the sanctuary of Opheltes, a child hero whose death formed part of the local cult. A graveyard could also provide a secluded place for leaving an object. Excavation shows a deposit’s location; determining the petitioner’s beliefs requires the inscription and careful comparison with other contexts.

Greeks approached divine powers in public sacrifice and prayer, in oracular consultation and in private ritual. Communities did not necessarily agree about which approaches were acceptable. In Plato’s Republic 364b–365a, Adeimantus speaks of travelling religious practitioners who claim that binding spells and incantations can injure enemies. Plato’s larger argument concerns the moral implications of such claims. The passage tells us that his fourth-century readers could recognize these practices and specialists, not that the advertised spells succeeded.

The Writing Itself: Names, Dialects and Signs

A few inscribed names could be enough for a binding act, as the Sicilian material suggests. Other texts repeat commands, enumerate bodily capacities or include unusual letterforms. Their scripts and spelling reveal differences in how those formulas were written. Some cramped or irregular lettering reflects the difficulty of scratching lines into metal; erasures need not indicate a coded alphabet. Later tablets sometimes carry charaktēres, signs evidently intended to function as ritual symbols rather than ordinary written words.

Dialect and spelling can suggest a text’s regional and social setting, though an inscription cannot always tell us where its writer was born. Pella provides a particularly valuable set of dialectal forms. Identifying names and family connections can also illuminate the movement of people and the presence of women, workers and foreigners in local communities. None of this establishes whether an individual author wrote the words personally or paid another person. Skillful lettering might come from a specialist; awkward strokes might result from an unfamiliar surface or haste.

The Greek Magical Papyri preserve comparatively elaborate recipes and invocations from Roman Egypt, including strings of sounds described as voces magicae. They supply valuable parallels, but cannot decode all earlier lead sheets. Any comparison has to respect the text’s date and origin, and the difference between a ritual prescription in a handbook and an object actually found in a tomb or well. Similarities may indicate shared practices without licensing the reconstruction of words that the earlier tablet never contained.

Why Some Lead Inscriptions Are Not Curses

Lead was used for other kinds of inscriptions too. At Dodona in Epirus, visitors to Zeus’s oracle recorded questions on small sheets, often asking about marriage, property or travel. A 2026 study by Hugh Bowden considers how thousands of those questions were written, deposited, erased and reused. Their writers sought advice, rather than the restraint of an opponent. The questions may resemble curses in size and material, yet their function differs. The inscription and archaeological context, rather than the lead alone, establish the purpose of an individual sheet.

A museum photograph without archaeological context can make such distinctions hard to see. Folded lead and Greek letters may suggest a binding spell, but only the inscription itself can establish its nature. Even a known curse fragment might leave its target unknown. The British Museum’s fragment from Athens, for example, is identified as part of a lead curse text but has a broad Hellenistic-to-Roman date. Script and material do not always allow a narrower attribution.

From Classical Greece to the Roman World

Greek-language curses continued under Roman rule, alongside a large body of Latin examples. In Britain, surviving petitions frequently concern stolen goods; other Roman-period writers attempted to hinder sportsmen, horses or chariot teams. Their rituals have affinities with classical Greek binding, although formulae and social circumstances differ. Imperial and late-antique sporting curses cannot simply be used to describe the normal interests of fifth- or fourth-century Athenian petitioners.

The Oxford Handbook chapter on curse tablets and sport places much of the surviving evidence for these spells in the imperial period or later. A study of a fifth- or sixth-century AD Aramaic circus curse from Antioch adds another language and city to the history of these practices. These examples document similar practices in other languages and cities centuries after the first surviving Greek tablets. They do not prove that an unchanged set of instructions persisted from archaic Sicily to late-antique Syria.

How Archaeologists Read a Folded Curse

Excavators may initially encounter a tablet as a corroded lump whose writing is invisible. Conservators record its condition and may use radiography or other imaging to establish how it is folded before deciding whether to open it. Although lead is malleable when worked, ancient folds can crack, taking inscribed strokes with them. Careless cleaning or unfolding may destroy the evidence. Once the object is stable, changing light angles, microscopy and digital images can help epigraphers separate original incisions from scratches and later damage.

The published transcription marks preserved letters and identifies gaps or doubtful readings. Comparisons with names, alphabets and other inscriptions can suggest restorations, but those suggestions must remain distinguishable from what survives. The edition of Curbera and Lamont’s Athenian Agora tablets also demonstrates the value of recording the excavation alongside the writing. A tablet’s date, the objects around it and the way it was left can affect interpretation. A remarkable text without reliable find records is consequently less informative about the rite in which it was used.

Repeated formulas or similar writing hands might suggest common conventions, copying or the work of someone experienced in preparing curses. That does not necessarily mean a professional wrote each tablet. Few inscriptions identify the person wielding the stylus, and the transaction between a paying client and a magician is generally invisible. Comparison of scripts and formulas can suggest shared conventions, but neat lettering alone cannot identify a paid specialist.

Were Curse Tablets Illegal or Effective?

Not all Greek writers judged harmful rites alike. Plato’s Laws 933a–e discusses wax figures, curses and fear of sorcery in the course of proposing penalties for an ideal state. It is a philosophical proposal, not proof that every Greek city had a statute criminalizing possession of a lead curse. Buried deposits often suggest a wish to keep a rite hidden. Secrecy could, however, be part of the ritual itself as well as a response to social or legal dangers.

For the person who buried a tablet, success meant the hoped-for change in another person’s affairs. A historian cannot trace any such outcome to supernatural causation, even when a later event happens to fit the words. The surviving object does establish that someone sought intervention through a written rite. The act may have given some petitioners a sense of control or deepened an antagonism already present. Neither response should be attributed to every writer, and no complete record survives of the results.

The surviving texts from Athens, Pella and Nemea differ sharply in their aims and wording. A fourth-century Athenian list names scores of potential adversaries; the Pella author names a prospective marriage and calls on Makron; the Nemean erotic texts still resist a single reading of their instructions. They belong to a broad tradition of binding, while preserving distinct grievances and uncertain interpretations. Their writers left the sheets where other people were unlikely to see them. They now preserve particular efforts to influence uncertain lives, with the words of the petitioner more readily accessible than the experience of the target.