Curse tablets and binding spells

What is attested

The physical object is not in doubt. Some thousands of thin metal sheets — usually lead, sometimes lead-tin alloy or pewter, occasionally other metals — survive from across the Greek and Roman world, inscribed with text intended to constrain, harm, silence or compel a named person; no defensible total exists, since the count depends on what is included and on how much unpublished material is reckoned in. Greek calls the act katadesmos or katadesis, 'binding down'; Latin uses defixio, from defigere, 'to fix, to nail down'. The vocabulary is consistent and physical, and the objects were treated accordingly: many were rolled or folded shut, some were pierced with a nail, and most were deposited somewhere that put them beyond ordinary reach — a grave, a well, a spring, a sanctuary of an underworld deity, the foundations of a building.

The earliest examples currently identified come from Sicily — Selinous above all, with material also from Akragas and Gela — and are placed at the very end of the sixth or in the first half of the fifth century BC. That dating is worth stating carefully, because it rests largely on letter-forms and dialect features rather than on stratified findspots: many of the early Sicilian tablets have no secure archaeological context, which means the claim of priority is only as firm as the method that supports it. From the fifth and fourth centuries there are substantial Attic groups, and the practice is then continuously attested into late antiquity and across the empire, from Roman Britain to Egypt and North Africa. The geographic and chronological spread is one of the few things about the corpus that can be asserted without qualification.

What the texts actually do is narrower than their reputation suggests. Many name a target and then specify what is to be bound: the tongue, the hands, the mind, the soul, the work, the lawsuit. The use of a matronymic in place of a patronymic is often described as a marker of the genre, though how common it actually is in the corpus has been questioned. Some address a deity or a daimon directly; some address the dead person in whose grave the tablet was placed; some contain no address at all and simply list names. Many later tablets carry charaktêres (invented sign-like marks resembling letters with small circles at the terminals) and voces magicae (strings of untranslatable syllables), and some carry drawings of bound figures. The formulaic overlap between tablets found far apart is most economically explained by transmitted written models, though a shared oral repertoire cannot be excluded.

The functional range that scholarship has drawn out of the texts is broadly fourfold — curses connected with lawsuits, with athletic and theatrical and later circus competition, with trade and commercial rivalry, and with erotic attraction or the frustration of a rival's attraction — plus a distinct group, best documented at Bath and Uley in Britain, in which the writer states that something was stolen and asks a deity to punish the unknown thief and recover the property. These last read as complaints laid before a god, sometimes with conditional and legalistic phrasing, and they behave differently from a straightforward binding text: the writer claims to be the wronged party, and the target is frequently unnamed because unknown.

The Bath deposit is the single best-documented body of this material. Roughly 130 inscribed tablets of lead and tin alloy were recovered from the sacred spring at Aquae Sulis in excavations of 1979–80 and published by R. S. O. Tomlin in 1988 as part of Barry Cunliffe's report on the temple of Sulis Minerva. The great majority concern theft — of clothing, bathing gear, small sums of money — and address the goddess. Because the group has a known findspot, a known deposition context and a coherent internal character, it allows conclusions about who was writing (people of modest means, including some with shaky Latin), what they wrote about (petty loss), and how the object was used (dropped into the goddess's water) that the scattered tablets from unprovenanced contexts cannot support.

How the rite worked

The procedure that can be reconstructed from the objects themselves has four observable stages: acquire and prepare a metal sheet; inscribe it, usually with a stylus or a sharp point, sometimes in mirror writing or with letters deliberately disordered; treat it physically — fold, roll, pierce; and deposit it in a place associated with the dead, with chthonic power, or with water. Every one of these stages is visible in surviving artefacts. What accompanied them in speech or gesture is not.

Formularies on papyrus from Roman Egypt — the body of texts conventionally cited as the Greek Magical Papyri, PGM — preserve recipes that specify materials, timing, wording and accompanying actions for binding operations. One of the fullest, PGM IV 296–466, prescribes the making of two figurines, the piercing of one with needles at named parts of the body, and the burial of the figurine with an inscribed lead tablet at the grave of someone who died young or violently. A clay figurine now in the Louvre, found in Egypt inside a terracotta vessel, matches these instructions closely: kneeling, arms bound behind, pierced with thirteen needles, and accompanied by a lead tablet naming a woman, Ptolemais, as target and a man, Sarapammon, as the person acting. The correspondence between recipe and object is the strongest single demonstration that written formularies were actually executed.

Cost and personnel are largely inferential. Repeated formulae, professional-looking hands, and tablets from the same deposit written by a single scribe for different clients point to paid specialists in at least some places and periods; other tablets are clumsy, misspelled, or barely literate and look self-made. Lead was cheap and widely available as scrap — water-pipe offcuts and sheet trimmings — which is part of why the practice reached people who left almost no other written trace. No ancient source gives a price for a defixio that can be confirmed.

Deposition was the operative act, and it was directional. Graves of the untimely dead (aôroi) and the violently killed (biaiothanatoi) recur as favoured sites, on the reasoning — stated in some tablets themselves — that such spirits were restless and available. Wells, springs and pools appear repeatedly, as at Bath and at the Anna Perenna fountain in Rome. Sanctuaries of Demeter and Persephone, and of underworld or vengeful deities, account for further groups. The consistent principle is transfer: the text is put somewhere it cannot be retrieved and where a power that can act is thought to be.