Hippocratic Diagnosis, Solomonic Therapy, Roman Amulets: Epilepsy, Exorcism, and the Diffusion of a Jewish Tradition in the Roman World

Two contrasting portraits of exorcism in the Roman period for patients with symptoms consistent with epilepsy, drawn by Josephus ( A.J. 8.45–47) and Lucian ( Philops. §16), illustrate a substantial albeit contested diffusion of that ancient technique from the Jewish tradition to a wider Mediterranean public. The process is reflected in a similarly complex traditional background and textual composition of a group of inscribed Greek amulets for epilepsy. A sidelight on attitudes towards the practice of exorcism, on its way to wider popularity, and the conception of epilepsy is cast by these amulets, which have not yet been studied as a group. Their texts witness the application of precise Greek medical terminology, yet to an end, and in a compositional company, that authors in the Hippocratic tradition would have rejected. More generally, the artifacts offer a cross-section of amuletic practice and its diversity in the Roman and late ancient periods.

learned the art (τέχνη) of demonology for the benefit of humanity, including incantations (ἐπῳδαί) and modes of exorcism (τρόποι ἐξορκώσεων) for healing and driving out demons. This knowledge was still in contemporary use, as illustrated by the cure (θεραπεία) performed by Eleazar, a countryman of Josephus, in the presence of Vespasian. The cure as practiced there consists of passing beneath the nose a signet ring with a root under its bezel, as recommended by Solomon, then adjuring the demon in Solomon's name not to return, applying also some of the incantations that he composed.
God also granted (Solomon) knowledge of the art concerning demons (τὴν κατὰ τῶν δαιμόνων τέχνην) for the benefit and healing of mankind: he composed incantations (ἐπῳδάς) for the alleviation of illness and left behind modes of exorcism (τρόπους ἐξορκώσεων) for those in need, to chase out demons such that they would never return. This healing prevails widely among us to the present day: for I have witnessed one of my countrymen, Eleazar, in the presence of Vespasian … releasing from demons those possessed (λαμβανομένους) by them. The mode of healing was as follows: he would bring his ring up to the nostrils of the demoniac, which had under its seal one of the roots recommended by Solomon, then draw the demon through the nostrils of the patient as he smelled. The person fell at once, and (Eleazar) would adjure (ὥρκου) (the demon) never to return to him, mentioning Solomon and uttering the incantations that he composed. Josephus, Based on the conjunction of the vocabulary of seizure (λαμβανόμενοι) and falling (πεσών), the episode has long been supposed to concern epilepsy, but we have only the account of Josephus, not the perceptions of Eleazar or his patient on diagnosis and causation. It is uncertain whether any of them would have recognized epilepsy in those terms. At least the historian, under the influence of Greek paideia, does apply medical terminology (θεραπεία, νοσήματα) as opposed to a traditional discourse on evil spirits, or sin and redemption.
To pass from praise to parody, Lucian, native of neighboring Syria, could speak in the second century of a more general tradition of Palestinian professionals of exorcism absent any explicit confessional affiliation. In and of itself, that is evidence for a diffusion of the technique. But more interesting ("premiers exorcismes") identified the exorcist (on the siege episode see Mason,History,. On the exorcism see further Duling, "Eleazar Miracle"; Deines, "Josephus"; Bohak,Ancient Jewish Magic,[100][101][102][103][104][105] is the way in which Lucian proceeds to ridicule it. In the Lovers of Lies, the sophist describes a fictional gathering of philosophers whose discourse on the paradoxical, and their uncritical attitude towards it, mark them as unworthy of their profession: an account of these "lies" (ψεύσματα) is the occasion for the framing story of the more skeptical Tychiades, who bears witness, and his interlocutor Philocles. Natural remedies involving sympathy, healing incantations, and erotic magic all come in for mockery by Tychiades before one of the philosophers, Ion, who represents the Platonic school, offers an example that he expects will be credible due to how well-known it is. The discussion amounts in the end to an unsuccessful attempt to convince Tychiades of the existence of daimones, phantoms, and ghosts, among other supernatural manifestations and magical operations. A professional (σοφιστής) specializes in the treatment of epilepsy, explicitly associated with the influence of the moon-reflecting a facet of popular belief to be discussed further on-which is diagnosed as a form of demonic possession. From this he makes a large profit. The exorcist engages the demon in dialogue, forcing it to reveal its origin and means of entry in its native tongue, and then compels it to withdraw, applying adjurations (ὅρκους ἐπάγων) and making threats (ἀπειλῶν). Witnesses are able to see the demon itself as it departs.
Everyone knows the Syrian from Palestine (τὸν Σύρον τὸν ἐκ τῆς Παλαιστίνης), the expert in this field (τὸν ἐπὶ τούτῳ σοφιστήν), how many patients he takes on, who fall under the influence of the moon (καταπίπτοντας πρὸς τὴν Σελήνην) and roll their eyes with their mouths full of foam, and stands them up and sends them off with sound mind, relieving them of their terrible suffering for a high price (ἐπὶ μισθῷ μεγάλῳ). When he stands over the fallen and asks from where (the demons) entered the body, the patient himself is silent, but the demon answers, speaking Greek or a foreign language (ἑλληνίζων ἢ βαρβαρίζων) depending on its origin, how and from where it entered the person. He drives the demon out by applying adjurations (ὅρκους), and even, if it does not obey, by making threats (ἀπειλῶν). I myself have seen one coming out, with a complexion like dark smoke. Lucian,Philops. §16 If Lucian meant to imply a Christian affiliation of this exorcist, as already claimed by a scholiast, he certainly could have done so more explicitly.3 Tychiades remains unconvinced, and his rejoinder adds on a jibe at the Platonic "forms" (ἰδέαι) as no more substantial than the demons, which the Platonists nevertheless claim to be able to see clearly.
What specific features of exorcism oppose it to the "truth" and "correct reasoning" that Tychiades holds up in the end as the "antidote" (ἀλεξιφάρμακον, §40) to lies? Besides the apparent contradiction in the diagnosis as the result of both lunar influence and demonic possession, there may well have been an assumption of familiarity with and acceptance of the tenets of Hippocratic medicine, which, as discussed below, explicitly denies the causation of epilepsy by either the moon or demons. While the notion of demonic forces as entities that one could address, and even engage in dialogue, would not have been ridiculous in and of itself (recall the guiding daimon of Socrates), to do so in the face of sound medical advice, to converse with the demons rather than debate with the doctors, surely would have moved an educated Hellene to derision. Lucian and his audience, with a shared foundation of Greek paideia, would have recognized the Hippocratic concept of epilepsy, not a demon, and shared a laugh at the idea of communicating, let alone remonstrating, with the disease.
In the New Testament writings shortly preceding Lucian, there is abundant evidence that early Christians were heirs to the same ancient tradition of exorcism, making the scholiast's remark at least chronologically possible. This much-commented topic will not be considered in any detail here, but some aspects will be significant for the context of the developments witnessed by the Greek amulets. Demonic possession was taken as a given in the Judean-Palestinian background of the gospels and frequently provided a setting for miracle-accounts. Jesus himself uses only direct commands to the demons, such as ἔξελθε "come out," never the technical terms for adjuration ὁρκίζω and ἐξορκίζω.4 The emphasis is on the sanctity and charisma of the exorcist as operative force, a point stressed in later Christian discourse, rather than on received knowledge of traditional ritual. In parallel, the Christian tradition developed more formal institutions and texts around the technology of exorcism. The chief witnesses to such texts are preserved through their later reflexes in the Byzantine church.5 The recommended Christian response to demons, already in the gospels, centered on prayer (Mark 9:29). The call for recently Karavas,"Luciano,, but details, such as the "high price" associated with the exorcisms and the "Greek or foreign" speech of the demon, do not correspond to the New Testament picture of exorcism, and Lucian freely applies the term Χριστιανός elsewhere. 4 On exorcism in relation to the practice of Jesus: Kee, "Terminology"; Twelftree, Jesus the Exorcist; Grappe, "Jésus exorciste"; Twelftree, "Jesus the Exorcist"; Nickel, "Jesus." 5 See the text and commentary of Delatte, office byzantin. prayer over the sick in the epistle of James (5:14-15) was also later interpreted as an endorsement of exorcism specifically.6 Later authors added the name and sign of Christ, as well as the recital of narratives about Jesus, broadly parallel to the role of Solomon in the Jewish tradition, and even non-textual phylacteries for the body.7 Great emphasis is laid by way of apology on simplicity: either the name of Christ suffices alone, or the plainest of adjurations.8 Christian apologists, such as Origen, were on guard against accusations of sorcery and superstition. Above all, a distinction was pressed between superstitious reliance upon demons in the form of invocations of them, and Christian practice of commanding demons in the name of a higher power. 9 Such discourse included in turn attacks on Jewish exorcists, which are revealing, if to be taken with much caution, for the background of Jewish predecessors or even contemporary competitors in the provision of exorcism. Particular censure is reserved for the misplaced faith of Jewish exorcists in adjurations by human names-one thinks again of Eleazar's reference to Solomon-and in additional apparatus such as fumigation and "bindings" (κατάδεσμοι), possibly including textual amulets.10 While maintaining personal charisma as central to the performance of exorcisms-the healing and apotropaic exorcism that concerns us now developing separately from the baptismal one11-the Christian church increasingly relied on specialists to fill this role, which mirrors in turn the inclusion of exorcism in the arsenal of ritual specialists who, capitalizing on the asset of literacy, offered the exorcistic amulets discussed in what follows. By late antiquity the exorcism had became the responsibility of holders of a regular ecclesiastical office. In the mid-third century, an exorcista is mentioned by Cyprian of Carthage among the clergy of that city.12 At the 6 Cyril of Alexandria, Worship and Cult,ed. PG 68:472. 7 Narratives about Jesus: Origen, Cels. 1.6, τῷ ὀνόματι Ἰησοῦ μετὰ τῆς ἀπαγγελίας τῶν περὶ αὐτὸν ἱστοριῶν. Phylacteries: a particle of miraculously, accidentally transubstantiated eucharist bread worn around the neck: Anastasius of Sinai, Spiritually Beneficial Tales 43 (ed. Nau, "texte grec," 63, with Wortley, "Repertoire," 116 no. 111). 8 Origen, Cels. 7.4: Christians drive out demons "with prayer alone and simple adjurations and the methods that a simple person would be able to apply" (μόνῃ εὐχῇ καὶ ὁρκώσεσιν ἁπλουστέραις καὶ ὅσα ἂν δύναιτο προσάγειν ἁπλούστερος ἄνθρωπος). Gregory of Nazianzen, Prayer for Baptism 17 (PG 36:381A): Christians have no need of "amulets and incantations" (περιαμμάτων καὶ ἐπᾳσμάτων), since the invocation of the Trinity provides "the great and noble phylactery" (τὸ μέγα καὶ καλὸν φυλακτήριον). 9 Origen, Cels. 1.6. 10 So the Christian in Justin, Dial. 85.3, who censures the pagan practices of Jewish exorcists (τῇ τέχνῃ ὥσπερ καὶ τὰ ἔθνη χρώμενοι), including "fumigants and bindings" (θυμιάμασι καὶ καταδέσμοις). 11 On the latter see Leeper, "Alexandria to Rome." 12 Cyprian, Epistles 23.
beginning of the fourth century, the future St. Romanus served as both deacon and exorcist (διάκονός τε καὶ ἐξορκιστής) at Caesarea Maritima.13 At the synod of Antioch in 341, the exorcist is recognized as a clerical grade alongside lectors and subdeacons.14 2

Exorcistic Amulets
So far only a limited role for writing can be seen in the ritual of exorcism itself, aside perhaps from the "bindings" (κάταδεσμοι) alleged to feature in the practice of pagan and Jewish exorcists in anti-Jewish polemic, mentioned above. In a development of the Roman period, however, an artifact-type that can be termed exorcistic amulets, following the classification of Roy Kotansky,15 burst onto the scene. Here, elements of the liturgical exorcism are adapted for protection in a more durable form. The texts are attested in Hebrew and Aramaic, as would be expected from their ultimate origins,16 and occasionally in bilingual forms,17 but the focus of the present discussion will be on the Greek texts, as Greek is the primary instrument of the diffusion of the technique beyond the eastern cradle of the exorcism and the language in which the intersection with Greek medical science, of central interest here, can best be observed. A central witness is the corpus of the amulets themselves, on metal and papyrus, from the Greco-Roman Mediterranean, which will be discussed in their turn. As these artifacts often lack archaeological context, not to mention information from contemporary informants on their perceived mechanisms and cultic context, it is fitting to begin with a complement from a literary text, which also suggests a wide diffusion of the underlying concept, even if the particular setting is fictional. In his novelistic account of the life of Apollonius of Tyana, Philostratus mentions an episode in his subject's visit to the Brahmans of India that attracted the attention of the Christian author Eusebius.18 There is an allusion to a background of just such a type of ritual object, with which third-century readers could have been expected to be familiar. of a boy possessed by a demon is unable to bring her son before the Indian sage Iarchas for healing, so a letter is produced as a substitute, full of threats (ξὺν ἀπειλῇ καὶ ἐκπλήξει) that will produce the desired result once the demon reads it (ἀναγνοὺς ταῦτα). Such exorcistic letters, substituting for the in-person exorcism, would go on to enjoy a career as a motif among healing miracles in Christian hagiography.19 In addition to the amulets on metal at the center of this study, textual amulets belonging to the Christian tradition are also found among the papyri, which take up in turn the earlier Jewish tradition. One Byzantine text cites the demons' previous oath to Solomon and lists at length the places where they are forbidden from hiding.20 This same Solomonic precedent is cited, with the addition of the archangel Michael, in an apotropaic amulet on a late ancient gold lamella found at Rome, where the exchange between Solomon and the demons is elaborated with reported speech from the latter.
I adjure (ὁρκίζω) all the spirits (πνεύματα) …, every evil spirit, as they recall the covenant (μνησθέντα τῆς διαθήκης) into which they entered in fear of Solomon and the angel Michael, when they swore the great and holy oath in the holy name (ὤμοσαν τὸν μέγαν καὶ ἅγιον ὅρκον ἐπὶ τοῦ ὀνόματος ἁγίου) and said, "We will flee and not lie."21 This tradition resulted in a dedicated, pseudo-autobiographical treatise of Solomon on his mastery of demons, including their names, functions, and methods of adjuration, his well-known Testament, which circulated in a broad manuscript tradition including late ancient papyri.22 Designs for a "seal" (σφραγίς) attributed to Solomon have an independent career as an apotropaic motif against demonic assault.23 The rise of the written in this context suggests a progression from oral performance by charismatic individuals to textual objects with graphic permanence and portability, perhaps produced and bestowed on the patient alongside oral components that do not survive. This development may connect with another trend of the Imperial period identified among amulets, in which natural objects originally credited with amuletic efficacy by inherent virtue were later 19 E.g., the sixth-century Life of Eugendus of Jura, 143-46, ed. Martine, Vie des pères. 20 P.Vind. G 337 (PGM P 10); see most recently Maltomini, "Note," 121. 21 SEG LIII 1110.5-15. 22 Edited by McCown, Testament of Solomon; see in general Johnston, "Testament of Solomon," and on papyrological fragments of the Testament, Daniel, "Testament of Solomon." 23 See recently Trnka-Amrhein, "Seal." elaborated, as part of the well-known epigraphic habit, with inscribed versions of texts first developed as oral performances.24 In Egypt there is another contemporary analogy in the adaptation of institutional liturgy and cult on a private scale, such as in the miniaturization of the "opening of the mouth" ritual for conferring life-force on effigies, transferred from cult statues in temples to the creation of magical figurines via Greek formularies on papyrus.25 Exorcisms in the magical papyri from late ancient Egypt offer a glimpse of the aftermath of their textual spread in the Imperial period. The technique had apparently been integrated into local ritual practice, such that by late antiquity its form, informed by both Christian and Jewish traditions, stood apart from the rapidly growing Egyptian Christianity.26 The bilingual Greek-Old Coptic papyrus codex PGM IV offers two relevant witnesses. The first is a recipe for the treatment of demoniacs attributed to a sage with an Egyptian name, Pibechis.27 Along with a complex ritual procedure and the inscription of a short amulet (φυλακτήριον) on a tin sheet containing only magical words and the short command "protect so-and-so" (φύλαξον τὸν δεῖνα), a lengthy spoken exorcism includes the phrase "I adjure you by the God of the Hebrews, Jesus" (ὁρκίζω σε κατὰ τοῦ θεοῦ τῶν Ἑβραίων Ἰησοῦ), and continues through a detailed aretalogy of the fearsome deeds of the Jewish god, including the events around the biblical exodus.28 There is also mention of Solomonic precedent, in the form of a powerful seal (σφραγίς) placed on the tongue of the prophet Jeremiah,29 and the invocation is explicitly tied to Judaism in a closing claim that it is a "Hebraic spell" (λόγος Ἑβραϊκός). The second witness has the broader title "Excellent procedure to cast out demons" and prescribes a shorter spoken exorcism, which is in fact bilingual.30 Beginning in Old Coptic with an invocation of the Christian god as "God of Abraham, Isaac, and Jacob," alongside Jesus Christ and the Holy Spirit (ⲭⲁⲓⲣⲉ ⲫⲛⲟⲩⲑⲓ ⲛ̅ⲁⲃⲣⲁⲁⲙ ⲭⲁⲓⲣⲉ ⲡⲛⲟⲩⲧⲉ ⲛ̅ⲥⲁⲕ [ⲭⲁ]ⲓⲣⲉ ⲡⲛⲟⲩⲧⲉ ⲛ̅ⲁⲕⲱⲃ ⲏⲥⲟⲩⲥ ⲡⲓⲭⲣⲏⲥⲧⲟⲥ ⲡⲓⲁⲅⲟⲥ ⲛ̅ⲡⲛⲉⲩⲙⲁ), it continues in Greek to address the demon and adjure it to come out, while the first-person speaking voice claims to bind and consign it to an infernal pit. 24 Faraone, Vanishing Acts. 25 Dieleman, Priests, Tongues, Rites, 170-82; Zellmann-Rohrer, "Overlooked Magical Formulary." 26 On exorcistic terminology in the magical papyri see recently Love, PGM IV. 3007-86. The name probably transliterates Pꜣ-bỉk "The one of the Falcon": see Thissen, "Ägyptologische Beiträge," 295. 28 See Knox, "Jewish Liturgical Exorcism"; Bonner, "Technique of Exorcism," 41-43; Horst,"Exorcistic Formula, This episode is not otherwise attested in scripture: see Perdrizet, "ΣΦΡΑΓΙΣ," 58; Knox,"Jewish Liturgical Exorcism, PGM IV. 1227-64.
A more prominent role for writing in the deployment of the exorcism itself comes across in one further magical papyrus. The document is a magical formulary on a book-roll, which offers a shorter exorcism not against a demon, but, in broadened scope, a perhaps demonically deranged member of the human body. The target is the wandering female womb, an affliction recognized by classical authors and which continued to be diagnosed into late antiquity and beyond.31 The womb is adjured to return to its natural seat in the following terms: I adjure (ἐξορκίζω) you, womb (μήτραν), by him who was set over the abyss before the existence of heaven or earth or sea or light or darkness, who established (κτίσαντα) angels, first among whom is Amichamchou kaichouchao cheroei oueiacho odou proseionges, and who sits over the Cherubim, carrying his own throne (...)32 The deity in whose name the adjuration is made, seated on an appropriately glorious seat to serve as analogue, is unmistakably foreign to Egypt and recalls the Genesis creation narrative of the Hebrew Bible, while the occult name for the "first" angel belongs to the angelology of the intertestamental period. This text is not to be uttered, but rather inscribed on a tin sheet, preserving its words as a more permanent amulet and also participating in the epigraphic fashion of its day.

Epilepsy: Diagnosis and Ritual Remedies
It is now time to turn to the affliction probably at issue in the passages of Josephus and Lucian, and certainly in the Greek amulets of interest here. According to modern medical conceptions, epilepsy is a chronic neurological disorder causing seizures, including convulsion, loss of consciousness, and abnormal behavior, arising from excessive or improperly regulated neural signaling.33 The characteristic seizures are preceded by intense physical and emotional distress and hallucinations, followed by convulsions and falling. Treatment is now available via medications that reduce neural activity, or surgery to remove affected areas of the brain. 31 Faraone, "Magical and Medical." 32 PGM VII. 260-71 with Faraone,21. 33 For the evolution of medical conceptions of the disease see Temkin, Falling Sickness.
In the pre-modern world, without these pharmacological methods of treatment and neurological modes of analysis, diagnoses of epilepsy, that is, sudden and otherwise unexplained seizures and the attendant psychological malaise, as supernatural possession were current, as were treatments by ritual means. Epilepsy is associated with divine or demonic affliction in ancient Babylonia, the cradle of the healing exorcism, and from its earliest attestations in Greek as the "sacred disease" (ἱερὰ νόσος).34 The more clinical "seizure" (ἐπιληψία, ἐπίληψις) also appears from an early date in the Hippocratic corpus (Morb. sacr. §10) and is the term preferred by Galen.35 The magical character of ancient cures, again prefigured in Babylonia, is also censured early on in the Greek tradition. According to the Hippocratic treatise On the Sacred Disease, contemporaries wrongly blame its infliction on the goddess Hekate, foreshadowing a popular association with the moon, and attempt to heal it with ritual purifications (καθαρμοί) and incantations (ἐπαοιδαί).36 Galen for his part describes the apparent success of an herbal periapt against epilepsy but explains its mechanism as pharmacological (De simplicium medicamentorum temperamentis ac facultatibus 11:859-860), and hence implicitly not demonological. Among the Hellenistic accounts of dreams and healing at the incubation sanctuary of Asclepius at Epidaurus, a fragmentary record of the cure of an epileptic (ἐπίλαμπτος, IG IV.12 123.115-17) recounts a dream in which the god appeared to apply a finger-ring to the patient (τῶι δακτυλίωι πιέξαι), an instrument that also figured in the exorcism ritual for demonic attack by Eleazar, as Josephus describes it. The connection between epilepsy and exorcism suggested by Lucian is not borne out explicitly in the New Testament, but the symptoms of one demoniac so healed, marked as a particularly difficult case, do recall its characteristics: convulsion, foaming at the mouth, and grinding of the teeth.37 The association with the moon, censured in Hippocratic medicine and parodied by Lucian, is duly reflected in Matt 17:15, of a patient as "moon-struck" (σεληνιάζεται). The ritual treatment of epilepsy, in this Roman environment following on significant engagement of Judeans with Greco-Roman culture and vice versa, is a microcosm of the conjunction of Greek traditions in the medical definition of the disorder, folklore about its means of infliction and cure, and the ancient technique of exorcism, particularly embraced and disseminated in Judaism. The small corpus of textual amulets that concern epilepsy, and combine exorcism with various other means, all depending on Greek for the inscriptions themselves and the formularies behind their copying, provide a more particular illustration. The amulets appear to take up the Greek diagnosis and classification of epilepsy and related concepts, but address it through techniques foreign to Greek medical, and even Greek ritual tradition, around the time of Lucian's parody.
Before the amulets are introduced, the background of Jewish attitudes and responses to conditions that Greeks would have identified as epilepsy should be examined. The Hebrew Bible describes symptoms in two places that have later been identified as epilepsy: the "falling" experienced by both Balaam (Num 24:4) and Saul (1 Sam 19:24), the latter associated also with the attack of an "evil spirit." The biblical text, however, uses no more specific description than literally "falling" (the root npl).38 Epileptic symptoms are rare in the Talmud. They can most often be recognized under the term nikpeh, "one who writhes" or perhaps is "bent" by a demon,39 and interested the compilers for the legal implications of disease, not medical analysis.40 Such were the opportunity to invalidate a marriage with a woman whose epilepsy has been concealed from her husband, a slave sale where the slave's epilepsy has not been declared, and the unfitness of epileptic priests and cantors for service. Occasional etiologies are given, reflecting a complex mixture of popular beliefs beyond medicine, including illicit sex or proximity to same; catching the light of a candle (b. Pesaḥ. 112b) or the moon (Der. Er. Rab. 11) while nude; and demonic attack including that of anonymous demons associated with the lavatory, with bloodletting, and with mills. The suspicion cast upon the mill (b. Ketub. 60b) has not yet been explained,41 but a 37 Mark 9:14-29 with Printz, "Quelques réflexions"; for the demonology of the New Testament see recently Martin, "Angels." 38 See in general Preuss,Biblical and Talmudic,[299][300][301][302] Preuss, 300. 40 Kottek, "Epilepsy," 349-59. 41 Preuss, Biblical and Talmudic, 300 n. 24. relation of sympathy between the grinding of teeth in a seizure and the action of the millstones may underlie the belief. Never is exorcism mentioned, let alone recommended. Demonic encounters are conceived as fleeting attacks, not lasting possession, a concept which will surface in the textual amulets, and hence may have rendered the performance of a formal exorcism impractical. Pharmacological and other medical therapies, however, are similarly omitted, and the legal rather than medical interest of the sources must be recalled once again. A definition of the "certified amulet" (qameʿa mumḥe) is nevertheless of special interest (t. Šabb. 4.4-9), as it includes textual-even if unspecified-and herbal components and seems to place epilepsy at its center. Certified amulets may "serve not only an epileptic but even to prevent epilepsy" (ibid.), and epilepsy is paradigmatic for the application of amulets. A probably metaphorical story of two epileptic Israelites, a priest and a layman, given amulets by a physician (Lev. Rab. 26.5), brings demonic attack as cause into focus once again: the layman, but not the priest, is furnished with detailed instructions for use, because the priest was already forbidden from walking in cemeteries, the latter apparently a place of demonic encounter and hence particular peril.
Philo, who combined a Jewish background and education with Greek paideia, shows particular interest in and sympathy with medical science among authors of the Second Temple period. Holding physicians in high regard-the skilled doctor, for example, serves as a metaphor in philosophical argumentshe also cites Hippocrates directly and shows a more general familiarity with and interest in medical observation of the human body, medical treatment, and drugs.42 Philo does not mention epilepsy in his surviving works, but his view of angels and demons as manifestations of the same divine concept, and rejection of fear of demons as superstition,43 suggest that he would not have placed much stock in the use of exorcism, let alone in amuletic form, to cure it, but rather cast his lot with the Hippocratic tradition.
Josephus, a key witness for the application of exorcism to an epilepsy-like affliction, showed both a shallower acquaintance with Hippocratic medical terminology and a less positive regard for physicians.44 He highlights, for example, their role in administering poison (B.J. 1.598) and failures to treat illness successfully (A.J. 15.245-246). More diverse medicinal traditions he cites with approval, such as the herbal lore of the demonifuge baaras plant, where demonic agency is acknowledged (B.J. 7.181-185), even if the role of God as 42 Hogan,Healing,[191][192][193][194][195][196][197][198][199][200][201][202][203][204][205][206] Philo, Gig. 16; Hogan,189. 44 Hogan, ultimate bringer and healer of illness is maintained. Josephus speaks of epileptic symptoms in the first instance as "sickness" and "seizure," and their cure as "therapy," bringing them already into the orbit of medical discourse, while holding up the Solomonic tradition of exorcisms with pride, as worthy of note and patronage by his Roman audience. Thus, while neither scripture nor its exegete Philo link epileptic symptoms with cure via exorcism, the connection of the pathology with the demonic is already suggested by the former and taken up more firmly by Josephus at least in part through the lens of Greek medical terminology. The oral exorcism described by Josephus in turn is absent from the Talmud, where it may already have been converted, so to speak, into a textual form in the otherwise unspecified inscriptions of the "certified amulet" for epilepsy. The background of causation suggested by the Talmud, including lunar influence and demons inhabiting specific sites (lavatories and cemeteries), implies a broader circulation of the epilepsy-exorcism-amulet nexus in popular belief, incorporating elements, especially the moon, from external traditions.

Exorcistic and Other Greek Amulets for Epilepsy
The ritual approaches to the treatment of epilepsy, outlined above from the Greek tradition, extended to textual amulets on metal that make specific reference to the pathology in medical terminology, which in turn incorporated textual exorcisms as a reflection of the popularity of that technique. Adapting the Jewish traditions of ascribing epileptic symptoms to demonic attack and dealing with the latter via exorcism, they now circulated in a language (Greek) that facilitated access by a more international public. The sort of handbook that must have underpinned that circulation can be glimpsed in the late ancient magical formulary from Egypt discussed above, which recommends, alongside a longer, spoken exorcism, a short sequence of magical words to be inscribed on a tin sheet.45 More direct evidence comes in a group of inscribed metal sheets (lamellae) in which the technique is applied in a permanent medium, and which in their diverse composition provide a suitable crosssection of amuletic practice in the later Roman and late ancient periods with a wide geographic diffusion, encompassing the Near East, Asia Minor, and Italy. These artifacts, of which seven may be counted (Table 1 below) that feature epilepsy as a central if not exclusive concern, can be analyzed into three textual categories.

5.
Syria (purchased at Damascus). Gold lamella, fourth/fifth century ce. Kotansky, Greek Magical Amulets, no. 57. An invocation (ἐπικαλοῦμαι) of angelic and other divine names and magical signs, after which the "lord archangels, gods, and divine signs" (κύριοι ἀρχάνγελοι, θεοὶ καὶ θῖοι χαρακτῆρες) are asked to avert all ills and in particular epilepsy (ἐπίλημψις) and headache from a bearer identified by an egyptianizing metronymic formula.
6. Unknown provenance (now in Baltimore). Silver lamella, late Imperial period. Unpublished; cf. SEG XLI 1745. The first eight lines, adjuring (ἐξορκίζω) a spirit blamed for the infliction of epilepsy in the name of the "sole god" (εἰς τὸ ὄνομα τοῦ μόνου θεο ̣ [ῦ]), which in part transliterates Egyptian divine epithets, have been transcribed from a photograph provided by the Walters Art Museum (acc. 57.1961). The rest (twenty-four lines) is a sequence of magical words and signs ending with a command to protect (φύλαξε, 33) a named bearer.
The first group presents lamellae inscribed with non-phonetic signs and Greek letters consisting purely of divine names or magical words distinct from syntactical Greek text. 49 Two are known thus far (Table 1, nos. 1-2). Both are thin silver sheets, originally rolled up in tubular containers. One was found at Sagalassos 48 L. 3 apparently gives the noun "ram" (sr: a form of the god Chnum) followed by the names of the gods Amun and Ra (Re), the word "gods" over the line-break (nṯr.w), and in L. 4 the phrase "the name (pꜣ-rn) of Kmeph" (for the latter deity see Thissen, "Κμηφ"); for ιω εσηφικα in L. 2 as containing an Egyptian acclamation (h(ꜣ)y) of a divine name cf. ιω ερβηθ in no. 4, while εσηφικα, so far unparalleled, perhaps contains reflexes of Egyptian ḥsy "glorified" and ḥyq "magic." 49 For the magical words see in general Tardieu, Kerchove, and Zago, Noms barbares, and for the signs, Dzwiza, Schriftverwendung; "Magical Signs"; and Gordon, "Charaktêres." in Asia Minor (no. 2), the other at Rome (no. 1). In both cases, these letters and signs are preceded by the direct copying of a rubric from what must have been the amulet-maker's written formulary.50 In the text from Sagalassos, further directions have also been copied, notably a prescription that gold, not the eventual silver, be used as the substrate. This mistaken copying of the heading from a formulary is the only sign that these finished products concern epilepsy at all: hence it is likely that more amulets targeting epilepsy lie unidentified among amuletic lamellae bearing solely these non-syntactic elements.51 The second group reaches a more discursive level, with longer invocations in syntactic Greek. The Jewish "God of Abraham, Isaac, and Jacob," angels, and divinities with gnostic roots are asked to rescue named bearers from epilepsy, demonic possession of other kinds, and illness more generally. The three examples (nos. 3-5) date between the second and fifth centuries, one on silver and two on gold lamellae. In the two cases where provenance is known, a wide range is covered, from Syria (no. 5) to northern Italy (no. 3). The deities whose help is enlisted against epilepsy indicate a diverse cultic background. In one case (no. 4) it is simply the Jewish God of the patriarchs-though further divine names, some egyptianizing, are appealed to further on. In others (nos. 5, 7) it is a plurality of gods, alongside archangels and the same sort of nonphonetic signs (no. 5) as applied in the first group, now qualified as "divine" (θῖοι) in their own right. In yet another (no. 3), it is the "holy first-begetter" (ἅγειε πρωτογενέτωρ) followed by a long sequence unintelligible in Greek, eialdalxaō blathammachōrixaēkaiaōsych pheōth perphthō iachthōgnedaseōps chthemmech, in which, at the very least, the name of the gnostic deity Ialdabaoth52 should be recognized as the referent of that epithet, alongside some vocables that recall Egyptian (perphthō).
The final group attests the specific technique of exorcism. Of the two amulets, one, now in London, has only recently been identified (no. 7), and the other, a silver lamella of unknown provenance now in Baltimore (no. 6), was known in print thus far only through a brief description. It can be read as an exorcism of the "spirit of the sacred disease of epilepsy, of the moon's order" (τάξις), which is adjured by "the sole god" and "the holy god" to depart. A sequence of occult epithets given to this deity proves to contain transliterations of traditional Egyptian divine epithets, while the continuation of the text in twenty-four lines of magical words and non-phonetic signs is consistent 50 For comparable cases see Faraone,"Scribal Mistakes." 51 For the evidentiary problems of anepigraphic amulets cf. Faraone,Transformation of Greek Amulets, For his appearance in Greek magical texts cf. Brashear,"Greek Magical Papyri," 3587. with the egyptianizing tradition of the later Roman Imperial period and late antiquity. The London amulet (no. 7) is of particular importance as an early witness to this group and to the complexity, already at that time, of conceptions of pathology and means of cure. The exorcism in turn shares space with an element from the second textual category, the invocation. The text, dated to the late first or early second century ce, begins with an invocation of the object itself to serve as amulet (φυλακτηρίασαι, 1) for a named bearer. In a second invocation multiple gods are referenced, but only a few are selected for address, hailed as the "greatest" (ο ̣ [ἱ] μέγιστοι θεῶν, 2-3), reflecting the attitude of megatheism as defined by Angelos Chaniotis.53 Their anonymity is supplemented by the epithet "those who have been passed down as" or "afforded divine names" (οἱ τὰ θεῖα ὀνόματα παραδοθέντες, 3-4), with the lack of specifics setting those names in the realm of unwritten, secret knowledge. What these deities are invoked (ἐπικαλοῦμαι, 5) to do is also omitted, but the same φυλακτηρίασαι may be applied to them, as their numinous presence acts as guarantor of the exorcism to which the text turns next (7 and following).
Before the characteristic adjuration is reached (ὁρκίζω σαι, 15), the list of demonic forces who should heed it is drawn up. The demonic is first couched in general as "spirit" (πνεῦμα, as also nos. 4, 6), consistent with the Judeo-Christian as opposed to Greek tradition (δαίμων and derivatives), the two strands later being combined (no. 3). It is then linked to nine potential causes or sources, unparalleled among ancient amulets but anticipating Byzantine liturgical exorcisms.54 Epilepsy heads the list, followed by mental illness that is probably symptomatic; then ill speech, either accompanying that disturbed state, or the blasphemy or cursing on which the disease might be blamed; then two cases of causation by humans: poisoning and a mysterious ἐπικλασμός. As demonic possession is presented as the result of an ἐπικλασμός made "for good or ill" (ἤ τίς ποτε ἐποίησεν ἢ ἐπὶ κακοῦ ἢ καλοῦ ἐπικλ ̣ ασμόν, 10-12), a trope from Christian hagiography might be compared in which Christians resort to magic to cure illness, bringing about, or worsening, demonic assault.55 The word could then stand for the more common ἐπίκλησις "invocation," if it is not a ritual action related to the etymological sense of ἐπικλάω "break." The next pair of causes refer to midday and evening as times of high demonic activity, both paralleled 53 Chaniotis, "Megatheism." 54 For discussion of these relations see the commentary in the edition of Zellmann-Rohrer, "Greek Exorcistic Amulet." 55 E.g., Gregory the Great, Dialogues 1.10.1-5 (ed. de Vogüé, Grégoire le Grand) with Wortley,"Repertoire, in liturgical exorcisms, and in the former case recalling the infamous "demon of midday" from the popular apotropaic Ps 90 (91).56 Epilepsy is specified twice more by name, reflecting two new facets of the pathology: it is equated with a demon, and that demon with sites regarded as haunted in popular belief, reflected also in the Talmud as discussed above, tombs and watercourses.57 The powers specified so far are finally recapitulated with the general "spirit" once again, now said to be entering the patient's body, and formally adjured (ὁρκίζω, 15). The first-person voice of that verb "I adjure" is distinct from that of the patient, who remains in the third person, perhaps a remnant of a separation of roles between exorcist and patient in a more formal, oral performance. The inscription finishes with another first-person claim to be destroying the demons directly, now considered as a "race" (δι ̣ αφθ ̣ ε ̣ ί ̣ ρ ̣ ω ̣ τὴ ̣ ν φυλὴν ὑ ̣ μ ̣ ῶ ̣ ν ̣ , 18-19). Some absences from the London amulet are also suggestive of a relatively early position, closer to a direct inheritance of the apotropaic exorcism from the Jewish tradition. Magical words, signs, and figures, characteristic of the often egyptianizing magical tradition of the later Roman and late ancient periods and reflected in the majority of the other epilepsy amulets, are lacking. So too are common signs of Egyptian influence in later Greek texts, the identification by mother's name (metronym) with the formula "whom (so-and-so) bore" as in the amulets from Arco and Syria (nos. 3, 5)58 and the transcription of Egyptian divine names (nos. 4, 6). Two probable Semiticisms in the text of the London amulet, the phrases ἡ σήμερον ἡμέρα "the present day" and ἡ ἄρτι ὥρα "the current hour,"59 bear out a role for translation of exemplars from the Jewish tradition in the composition of the London text.
The sheer multiplicity in the conception of and approaches to epilepsy in these amulets is a fitting illustration of the broad diffusion and recombination of traditions in the Roman empire, mediated through writing especially in Greek, which penetrated into the Roman heartland (nos. 1, 3). Epileptic symptoms are described with medical terminology (ἱερὰ νόσος, nos. 3, 6; ἐπιλημψία, nos. 4, 6, 7; ἐπίλημψις, no. 5) borrowed from the same Hippocratic tradition that would have condemned these objects and other ritual approaches, and the designation by lunar influence rejected there also appears (nos. 1, 2, 6). Causation is multiply determined: symptoms are listed (fainting and falling, 56 For the use of this Psalm see Kraus, "Archäologische Artefakte." 57 Eliav, "Roman Bath," 429-30. 58 See in general Curbera, "Maternal Lineage"; Dieleman, "What's in a Sign." 59 The nexus ἡ σήμερον ἡμέρα and (ἡ) ἄρτι ὥρα (as also in no. 3) do not appear in Greek before the Septuagint (e.g., Gen 19:37) in the former case and the Pauline epistles (1 Cor 4:11) and Josephus (A.J. 19.171) in the latter; for further parallels in amulets see Zellmann-Rohrer, "Greek Exorcistic Amulet." nos. 2, 4; mental derangement, no. 7), distinct maladies are added on (all disease, demonic attack, and fright, no. 3; headache and all other ills, no. 5), and besides the effects of the moon and of demons (the latter finely gradated in no. 7), origins including the anger of gods and the Fates (no. 3) and poisoning and other human actions (no. 7) are catalogued. A fittingly comprehensive definition, uniting Jewish (and later Christian) attribution of disease to evil spirits with Hellenic lore about the moon upon the fulcrum of the medical term "epilepsy" itself, is formulated in one of the specifically exorcistic amulets: "spirit of the sacred disease of epilepsy, of the moon's order" (no. 6). Alongside exorcism (nos. 6, 7), there are appeals to divine protectors from the Jewish, gnostic, and ancient Egyptian traditions (nos. 3-5, 7), and direct application of ritually efficacious text (nos. 1, 2; suggested also in the invocation of the amulet itself in no. 7). Further contrasts are implied in the engagement of the people behind the texts with the disease so conceived. The exorcism entails a discourse between human and demon or personified illness, which Lucian holds up for ridicule, while other techniques, more palatable to educated Greeks as at least harmless, direct the speech to beneficent deities who are to perform the healing on the patient's behalf. The combination of the two strands in the London amulet (no. 7) ultimately show the irrelevance of that intellectual distinction at a more popular level. 60

Conclusion
The ritual therapy of epilepsy and related conditions in the Roman Mediterranean witnessed a conjunction of Greek traditions in the medical definition of the disorder, folklore about its means of infliction and cure, and the Near Eastern technique of exorcism. As Josephus's account of one such exorcism performed in the presence of Vespasian suggests, Jews could take pride in the royal pedigree of healing exorcisms as practiced by Solomon and hope that Imperial patrons would also take an interest. A group of metal amulets with inscribed exorcisms manifest that interest in a durable form, further reflecting 60 An eighth, still unpublished amulet probably from the Near East and dated to the fourth-sixth centuries ce came to the author's attention too late for full consideration but does not substantially modify the conclusions reached here: it invokes angels and other supernatural beings to protect a women identified by the egyptianizing metronymic formula from epilepsy and other ills, makes reference to the moon, and includes ritual signs (charaktēres); a description courtesy of R.W. Daniel and A. Hollmann is given in Zellmann-Rohrer, "Greek Exorcistic Amulet." contemporary beliefs and anxieties about an increasingly hostile demonic as opposed to daemonic world.61 The texts are an eclectic mixture of motifs characteristic of the magic of the Roman Imperial period, but an early stratum in particular witnesses the fixation of the Jewish oral exorcism as written, textual amulet, in which confessional specificity has been effaced-recalling the free agency of the anonymous "Syrian from Palestine" described by Lucian. Specific parallels both in diction and in conceptualization of demonic forces between the London amulet and later Byzantine exorcisms suggest another inheritance taken up from the Romans by their self-proclaimed successors the Byzantines, the self-styled Ῥωμαῖοι.62