Roman Curse Tablets from the Baths: Manuscripts for Magic, Ritual, and Religion

—The Roman Baths in the City of Bath World Heritage Site in the United Kingdom is the origin of the well-preserved remains of one of the outstanding religious spas of antiquity and the greatest healing shrine of Roman Britain. The Roman curse tablets represent personal and private prayers inscribed on sheets of lead or lead-alloy and cast into the sacred spring of Aquae Sulis at the Roman baths complex in the city of Bath, United Kingdom. The tablets range in date from the 2 nd to the late 4 th century AD and in script from Roman Capital Letters, Old Roman Cursive, and New Roman Cursive. The Roman Curse Tablets were discovered in the late 1970s. This study seeks to briefly review the history, content, material and technique, and application of some of these tablets with a glance over the mythology.


I. INTRODUCTION
The Roman curse tablets, otherwise known as Defixiones, represent personal and private prayers inscribed on sheets of lead or lead-alloy and cast into the sacred spring of Aquae Sulis at the Roman baths complex in the city of Bath, United Kingdom. Tomlin suggests that the tablets range in date from the 2 nd to the late 4 th century AD. The Roman Curse Tablets were discovered in the late 1970s [1]. This paper is focused on a review of the history, content, and application of these tablets with a glance over the deity to which the prayers were addressed as well as the mythology.
The Roman Baths, which were in use for 400 years, comprises bathing spaces with different settings and applications. The city's one-of-a-kind thermal springs rise in the site and provide the Baths with natural hot water as of today. The Great Bath is an enormous pool lined with 45 sheets of lead and filled with hot spa water. It was built surrounded by a vast barrel-vaulted hall that rose to a height of 20 meters.
Tomlin implies that in the 1860s and following demolishing of an old inn on the other side of the road, the temple podium was found. In the late 1870s, the wall enclosing the hot spring was explored through excavations, followed by the draining of the King's Bath above the spring, which resulted in exposing the Roman reservoir underneath [2].
Reference [3] indicates that the East wing of the Roman bath house contained a sizable hypothermal bath fed by water that flowed through a pipe from the Great Bath. Over time, several heated rooms built on site, extended in dimensions and number until the 4th century A.D., when the site reached its maximum extent. Located at the heart of the site, the Sacred Spring is where hot water of 46° C has been rising at the rate of 1,170,000 litres per day since before the Romans developed the Baths. This phenomenon was attributed to the gods of antiquity. The Romans built a temple in the site and dedicated it to the chthonic goddess Sulis Minerva, who was believed to have healing powers (Fig. 1). Reference [3] also suggests that the Romans lobbed lead curses to the local Celtic chthonic goddess of the bath. Her name was Sulis and she was identified by Romans with their own Minerva, goddess of Wisdom.  The spring rises withing the courtyard of the Temple and its water feeds the Baths [3]. According to references [2] and [3], finds of an earthen bank adjacent to the Spring suggest it was a worship point prior to the establishment of the Roman temple and baths. Initially an open pool in a corner of the Temple courtyard, the Sacred Spring was enclosed within a barrel-vaulted, semi-dark building with columns and statues, all contributing to the mysterious atmosphere of the Baths. Offerings and votive objects were thrown into the Sacred Spring by the worshippers of Sulis throughout the Roman era ( Fig. 1.) [2].
The vaulted building collapsed into the Spring in the 6 th to the 7 th century. During the archaeological investigations on the Sacred Spring in 1979, the brick rib-vaulting of the roof was exposed lying as it had fallen (Fig. 2) [2].

A. Deity
When the Romans built the bath-and-temple complex in the city of Bath in the late first century A.D., they called the place Aquae Sulis [4], meaning the waters of Sulis, a British deity who was equated with the Roman goddess Minerva [2]. All classes of people came to Aquae Sulis, to visit the temple of Sulis Minerva, the hot springs, and the Great Baths [5].
Based on the etymology of the goddess's name, and her several other specifics, such as the association with sight, civic law, and epithets relating to light, at least in pre-Roman periods, Sulis has been recognized as a solar deity [6].
In the Encyclopaedia of Celtic Mythology and Folklore, Sulis's role is mentioned as the de facto Celtic solar deity [7], while the associated Sulevia and like names are the goddess's attestations elsewhere [8].

B. Mythology
To depict a situation, in which a Roman would resort a curse, Jordan imagines a Roman guy, Gaius, at the Baths [9]: When Gaius stepped naked from the bath at Aquae Sulis, he found out his cloths had been stolen. He was left naked, wet, and desperate. The thief had disappeared and_ according to the writers of the Roman law_ no policeman would ever find the thief. At the very best, they could determine a penalty, as harsh as labor in the mines of the bathhouse, in case the thief were a civilian. For a Roman soldier, the penalty would be immediate, dishonorable discharge from service. For Greeks, too, burglary counted as a civil offense. However, stealing someone's clothes from the bath would be punished by death [2] IF you got caught. Jordan mentions that almost no one was [9], while Tomlin does not completely rule it out [2].
As Gaius waited for his friend or his slave to bring him a new suit of clothes, suffering the injustice, he decided to enact a legal fiction and donate his stolen clothes to the goddess of the waters Sulis Minerva. Gaius urged Sulis to find the thief and to torture him until he would return her own clothes to her [9].  McKie has identified a petitioner named Solinus [10], to whom Tomlin attributes a similar story [2].
According to references [11] and [12], Sulis had received such gifts before and on similar occasions. Ref. [12] explains that finds from Sulis spring comprise over one hundred quasi-legal, quasi-magical documents of the late 2nd and the 3rd century A.D.. Ref. [2] does not completely rule out the probability of more objects still buried in the spring.
The gifts included not only clothes, but coins, jewelry, iron pans, curse casts, and even some thieves [9] (Fig. 3, 4). Among the finds from the site was the gilded bronze head of Sulis Minerva herself [2] (Fig. 5).
People also wrote to the Roman goddess Sulis Minerva asking for wrongs to be put right or for revenge; Tomlin calls it "a nervous respect the uncanny waters which could heal_ and harm". In short, the Romans did not have police, but they did have something else that worked as a deterrent: the fear of the gods.
The Roman vestiges of the Bath have remained at the heart of the city's development since their inception and are among the most famous and important Roman remains in the northern Alps. The Roman Curse Tablets are the earliest known surviving prayers to a deity in Britain and apparently the richest known archive of such manuscripts so far [2].
As Watson cites in her book, according to Jordan, "Defixiones, more commonly known as curse tablets, are inscribed pieces of usually lead, in form of small, thin sheets, intended to influence, by supernatural means, the actions or welfare of persons or animals against their will" [13]. The terminology of the practice of magic in antiquity is subtle, varied, and specific at the same time. The most frequent term in this terminology is defixio_ with the plural form of defixiones or defixionis_ which means "enchantment" or "a binding", from the Latin verb defigo, meaning "to fasten" [14], [15]. That explains why many curses request that someone be bound [16], [17], [18].

III. MATERIAL AND TECHNIQUE: WHY LEAD?
Most of the Roman curse tablets from the votive deposit in the Sacred Spring of Sulis Minerva at Bath consist of thin pieces of lead or lead alloy inscribed with a metal nib, likely a stilus [2], [19]. Extensive research about the use of lead in antiquity justifies the reasons why lead was the choice for creating curse tablets. Lead is fairly cheap and abundant, easily pliable, and simply readable when inscribed [2], [20].
Since lead is not a precious metal, the location of lead mines is not usually specified in the textbooks about the Roman era but the position of silver mines is often marked because of the importance of the metal in both Attica and the Roman Empire. Usually, silver mines are also a lead mine because lead and silver occur together. However, the ratio of lead to silver varies in each mine: British mines had the highest ratio of lead to silver [21].
Reference [2] reports that the curse tablets retrieved from the Roman Baths belong to a wider and longer-lived tradition than a single handbook or professional scribe. According to him, the texts are written in Roman Capital letters [2], Old Roman Cursive, and New Roman Cursive (Fig. 6, 7, 8). There is one tablet that has been written in the two latter scripts [2].  There is another piece, whose script remained unknown for more than twenty years until the editors of Roman Inscriptions of Britain decided the tablet was not inscribed at all but that the marks were intended to look like writing [2]. Section V of this paper will focus on this piece.
The same recipes and writing ideas can be found in more than one tablet, though not to the level of duplication. There are also formulas, which can be traced from the early second to the fourth century and which are also found at other cult centers than Bath [22].
In his 1992 paper, Tomlin suggest that out of the 89 tablets he had been able to tabulate, only two were from the same hand, which were in fact the two halves of the same document. Given that and considering even the tablets that look alike were "too incompetent to be professional work", he concludes there were no such people as tablet writers. Instead, Tomlin suggests that petitioners probably did seek advice from the staff of the temple on what to say or write, though they "were expected to write it for themselves" [2].
IV. VINISIUS TO NIGRA: A LETTER READ TWICE Sayce, an Assyriologist at Oxford specializing at cuneiform tablets was the first one to get curious about a piece of sheet-lead roughly measuring 76 by 44 mm (Fig. 9, 10) and found in 1880 in the hot spring of the Sacred Bath. Both sides of the tablet indicated seemingly handwritten scratches. In 1889, when Sayce visited Bath, Haverfield, an Oxford Ancient History professor suggested the scratches were not letters and that there was no inscription on the tablet. However, in 1900, Sayce suspected that the sheet actually bore a Celtic inscription, so he referred it to Nicholson, an Oxford philologist and the Bodleian Librarian for several years [22]. Nicholson recognized the script to be a combination of 'majuscules' and 'minuscules' with ligatures and abbreviations, such as Xρm standing for Christum (Fig. 11). He wrote brief palaeographic notes elaborating on the transcript and on letter-forms, though he failed to provide any evidence that the script was related to any other examples of Roman cursive, including to Thompson's [22]. Nicholson's deduction was that in this tablet "The same letter of the alphabet often has various shapes … and the tablet sometimes shows marks to be accidental which in a photograph might be mistaken for intentional cuts.". Sayce, on the other hand, suspected the script to be Celtic. Nonetheless, it turned out to be Latin and, according to Sayce, "a complete 4th cent. Latin letter written by a Christian man to a Christian woman in Britain" [22].
Regarding how this letter ended up in the hot spring, Nicholson suggested it may have fallen out of Nigra's dress, though it was probably lobbed as an offering to the sacred spring, as were the pins and coins found in the Bath. Ludwig Traube, a professor of medieval literature found Nicholson's decipherment obscure after checking it against the original text [22].
Tomlin reports that there is discordance between Nicholson's text and the new text "with its bushels of cloud and smoke". Regardless, Tomlin infers, Traube's guess that the object is a defixio sounds reasonable, for it contains recipes found in other Bath 'curse tablets' [22].
Tomlin also mentions that it is a fourth-century text, as Nicholson suggested, and that the author seems to have used 'Christian' terminology. However, the tablet is evidence of the existence of a pagan cult in the area surrounding the spring of Sulis, rather than Christianity. Nicholson's decipherment too, Tomlin suggests, cleared the doubt about the fact that "Christianity in Roman Britain is still short of written texts." [22].
According to Tomlin, Vinisius has remained questionable because the tablet has not been seen since 1939 at the beginning of WWII, when it was taken off view for safety and preservation considerations. Nevertheless, in 1979, the original find spotthe Sacred Bathwas investigated anew and more than a hundred inscribed tablets were explored [22].
Tomlin reports that unfortunately for Nicholson, the papyrologists, whom he would consult with copies of Vinisius to Nigra, were in Egypt. On the other hand, Tomlin assumes, Nicholson considered it beyond his dignity to consult his former rival, Thompson [22].
Notwithstanding, Thompson considers it nonchalantly easy to criticise Nicholson with our current knowledge of the New Roman Cursive replacing Old Roman Cursive by the end of the third century AD, given that Thompson's own Handbook includes no example of handwritten text from the third and fourth centuries [22].
Looking at Tomlin's tabulation of the letters inscribed on Roman Bath curse tablets in his book Tabellae Sulis presents a clearer view of the text (Fig. 12). Even though Nicholson's photographs do not preserve all the text, they indicate clearly that he was right, and that Haverfield was wrong; the tablet was, in fact, inscribed. Though, from palaeographical point of view, Nicholson made a blunder: he read the text upside down [22].
Based on Nicholson's photographs indicating that the tablet, like others from Bath inscribed in New Roman Cursive, was made of a thick sheet of lead or lead alloy, Tomlin makes the following observations: The parallel striations visible upon both surfaces are probably due to the use of an edged metal tool to smooth the writing-surface. The surface is rough in places, as a result of being cast, and perhaps also from corrosion after the tablet was inscribed. … The original would probably be entirely legible, but even from the photographs alone it is possible to recover much of the text." [22]. The tablet reads: "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." [22].

V. MAGIC, RITUAL, RELIGION
Tomlin refers to the Bath visitors as pilgrims and reports that a study of the 150 names that the tablets bear suggests that they came from lower economic and social strata. He also mentions that 21 of the names belong to petitioners and the rest to thieves or possibly to enemies. More than half of the names are Celtic and the rest are ordinary Roman names similar to modern Christian ones [2].
"The knife-marks are visible on three edges; the fourth edge is uncut, and has the undulating outline and rounded profile, which belonged to the edge of the original cast sheet. Therefore, the tablet was complete when found.
In specific, pointing out to the word Rostrum in a tablet, which could mean a bird's beak, Tomlin interprets it as a reference to St. Columbanus, who once lost one of his gloves and, being a saint, he realised what had happened and cursed the raven that he believed had stolen his glove [2]. Jordan also suggests that the tablets imply religious beliefs [9].
Some of these curses are ruthless.
In some examples the person is seeking revenge for theft of a bronze vessel and asks that it be filled with the blood of the thief [2] (Fig. 6, 7). Tomlin suggests that "not many of them curse thieves" [2].
According to Kiernan, Tomlin, and McKie, some curse texts included magical words and symbols or were inscribed back to front as formulaic appeal to the gods to increase the curse's strength, power, and potency [2], [19], [10]. Tomlin suggests that some writers enciphered the text by writing it backwards or in mirror-image letters, though most of them wrote as usual and then rolled up the tablet (Fig. 13) and threw it into the bath water, where only the chthonic goddess could read it [2]. According to Dungworth, there were two main ways to activate a curse: "by placing in a chthonic or haunted location or by nailing" (Fig. 14). He also suggests that "the nailing of the curse tablet may also have acted as a means of displaying the tablet (perhaps in a temple precinct) where it could be seen by all (unlike those which were rolled up and then hidden in wells, cemeteries, etc.).
Tomlin advises that_ not all, though_ most of those tablets which have been discovered in Britain belong to the former category [23]. Dungworth also quotes from Henig [24] that the term defixion suggests the nailing of a curse tablet played a significant role in the ritual "fixing" of the target [23].
On the symbolic meaning of nailing, Dungworth suggests that being found in many European cultures, nailing "may derive from the symbolism of the Crucifixion (although the survival of folk traditions connected with defixiones should not be discounted out of hand). Ferguson (1954: 178) identifies nails as a potent Christian symbol because of their use in the crucifixion of Christ. The crucifixion acted to reestablish Christ's divinity and ensured human salvation" [23].  While the practice of placing a curse on someone might be dismissed as superstition in the modern age, the Romans considered the efficacy of magic as science [16]. This is indicated not only by the continuity of the use of defixiones throughout the Roman Empire, but by investigation of magical texts and recipes used for self-protection against magic as well because some curses actually did result in people's illness or death [12].
As mentioned before, writing formulas and ideas do not appear to have been unique but can be found in different tablets, though not to the level of duplication. There are also formulas, which can be traced from the early second to the fourth century and which are also found at other cult centers than Bath [22].
Also, out of the 89 tablets Tomlin had been able to tabulate, only two resembled in terms of the hand, which turned out to be the two halves of the same manuscript. On the other hand, even the similar tablets seemed to him "too incompetent to be professional work". Therefore, it seems that petitioners were not tablet writers, although they probably did seek advice on what to say or write from the staff of the temple [2] According to Kiernan and Tomlin, the argument of the efficacy of the curses is based on "the ancient belief in the power of magic and the principle of psychosomatic illness. The thief's belief in the power of the tablets, perhaps combined with a guilty conscience and the awareness that he or she may have been cursed, would have weighed heavily in his or her mind and eventually caused him or her to become sick" [19].
The Roman curse tablets are also indicative of the level of Graeco-Roman bilingualism in the British population under Rome [25], [26] (Fig. 6). UNESCO World Heritage documents indicate that unlike many classical sources, the tablets do not tell us of the lives of great men and women; nor are they great works of literature or philosophy [27], [28].
Typically, what can be found from most Romano-British Curse Tablets is a complaint of theft and the name of a stolen property, along with several punishments requested for the thief from Sulis Minerva, including ill health [2]. In some cases, a reward is also offered to the deity for the retrieval of the stolen piece [2]. After all, defixiones were prayers for justice [9], [16], [19], [29].