2

THE ANCIENT CONTEXT

IT WAS OBVIOUS to many early modern Europeans that their ideas and images of witchcraft were at least partly inherited from antiquity. The text that was most fundamental to their culture, the Bible, was itself ancient, and the authors of the demonological texts which supported witch prosecution quoted lavishly from it, and also from the Church Fathers. They also, however, included passages from pagan Greek and Roman authors: one of the most famous of such witch-hunters’ guides, the Malleus maleficarum, cited five of those; Henri Boguet’s Discours des sorciersalso had references to five; and Martin del Rio’s Disquisitiones magicae drew on a grand total of twenty-nine.¹ Creative writers were just as disposed to use such sources. Sometimes this process was implicit: the most famous witches in the whole of early modern literature, those who deal with Macbeth in William Shakespeare’s play, were derived originally from the ancient mistresses of prediction, the Fates or Norns, and in parts the chant they use seems similar to one composed by the pagan Roman poet Horace.² At other times it is explicit, so that when Shakespeare’s only slightly less famous contemporary, Ben Jonson, added an antimasque featuring witches to his Masque of Queenes, he stuffed the footnotes to his published text with references to Greek and Roman authors.³ Recent historians of early modern attitudes to witchcraft have, understandably, been generally disinclined to follow up these links: after all, their concern is with the later period. Those who have done so have tended to be authors of general surveys of the subject and to devote a few pages to suggesting that the ancient Mediterranean world had either a similar fear of witches to the early modern European one or a different one from it.⁴ Parallel to the tremendous expansion of research into early modern beliefs concerning witchcraft and magic in the past few decades, there has been an equivalent development in the study of the same subject in ancient times; by 1995 one of the most distinguished of those involved in it, Fritz Graf, could already speak of a ‘boom’ in the field, and this has intensified further since.⁵ These two developments have occurred almost without dialogue between them, and historians of ancient civilizations have tended to confine themselves to the particular one that is their individual specialism, without cross-comparison even between them. In recent years this pattern has begun to provoke some concern among them, with calls for a greater recognition of the differences between concepts of magic in ancient cultures, and an end to ‘universalizing generalizations and reductionist approaches’.⁶ This chapter is a response to these calls, and will attempt a broad comparative survey of attitudes to witchcraft and other forms of magic across the ancient Mediterranean and Middle Eastern worlds, drawing on the mass of recent research and some of the primary sources upon which it has been based. It will build on the approach employed to anthropological data in the first chapter, by emphasizing the distinctive nature of the attitudes adopted by different cultures, and attempting to determine what is constant in them and what is not. One of the lessons provided by a worldwide survey of witchcraft beliefs has been the critical importance of local variation, and it is equally important now to see if these ancient peoples exhibited the same phenomenon.

Egypt

The ancient Egyptian attitude to the supernatural and the divine was centred on the concept of what was called heka, signifying the animating and controlling force of the universe.⁷ It was employed by divine beings to maintain the natural order of things. They had no monopoly on it, however, because individual deities could teach it to humans, who could then deploy it not just against their own kind, but against other deities, to achieve their own desires and increase their might. This was part of a world picture in which the boundary between human and divine was porous, so that goddesses and gods often needed the aid of human beings and the greatest of the latter could become deified when they died and sometimes even before then. It was accordingly entirely permissible, and even admirable, for people to try to coerce deities, and texts in royal tombs which addressed the latter mingled praise with threats, and prayers with demands. Heka was especially expressed in words, spoken or written, but also by ritual, often linked to particular stones, plants and incenses. It could also be triggered by the making and treatment of statues and figurines, so that from the dawn of Egyptian history, at the opening of the third millennium BC, kings were portrayed as striking bound effigies of enemy prisoners to favour their fortunes in war. By the middle of the second millennium, models of people were being placed in tombs to work for the deceased in the next life. This was a system of thought which completely collapsed into a single whole the categories of religion and magic, as defined at the opening of this book, and with them those of priest and magician, and prayer and spell. The same spell, indeed, could implore, cajole, flatter, threaten and lie, in its attempt to gain the compliance of a deity or spirit.

Ancient Egyptians had no concept of witchcraft, and therefore represented another of the examples of peoples scattered across the globe, as described earlier, who lacked one. One reason for this may have been that they were also one of the societies which believed in the ‘evil eye’ and feared foreigners as hostile magicians; traits associated worldwide with a reduced or absent fear of witches. Only once is the use of magic mentioned in a criminal trial, and that was of a group of conspirators at the royal court who had tried to kill King Rameses III in about 1200 BC. One of them had attempted to do so by using magic, making wax images and potions with allegedly deadly spells learned from a book in the royal library; but this was treated as the employment of just another weapon, and the culprit was convicted of treason and not of witchcraft.⁸ Heka was itself entirely morally neutral and could be legitimately employed against both public and private foes, as long as the quarrel was generally considered to be a just one. Ordinary people who wished to exercise it often obtained the services of a special kind of temple priest, the ‘lector’, who was expert in its use and drew partly for such knowledge on the books in temple libraries. Lectors functioned as service magicians, most commonly by conferring protection against misfortune or attack, or treating medical problems. Tomb inscriptions, however, suggest that they could also use lethal curses, and state officials and private persons alike were expected formally to pronounce those upon the kingdom’s foreign foes. Literature and art both support the idea that the laity, including commoners, also had specialized magical knowledge deployed for specific purposes. Some of this is summed up in references to specialists in certain kinds of spell, such as ‘scorpion-charmer’ and ‘amulet-maker’; so well integrated was magic into the whole social and religious system that there was no general word for a magician in the Egyptian language.

Egyptians believed in frightening and dangerous spiritual entities, some inherent in the cosmos and others the ghosts of dead humans or the agents of angry deities.⁹ They were especially associated with the realms beyond the normal haunts of humanity: night, the desert and the underworld. Magical protection was invoked against them, but they were not considered to be intrinsically evil, but as having a mixture of positive and negative qualities, the former or latter predominating according to context. If they could be turned against enemies, then they became powerful helpers, and texts existed to do just that. Likewise, at least from the early first millennium BC onwards, it was considered possible, and even admirable, for a proficient magician to make a supernatural being into a personal servant. Especial significance was attached to knowing the true names of such beings, which could confer power over them upon those who possessed such knowledge. This belief system underwent no substantial alteration during the three thousand years between the appearance of the Egyptian kingdom and its conquest by the Romans, only an increase in associated objects and actions. Oddly enough, for a culture as enduring, formalized and apparently as static as the Egyptian, it quite early displayed a capacity for absorbing ideas from other cultures, especially those to the east: from the mid second millennium BC, spirits with Semitic names start to abound in Egyptian texts. Conversely, by the early part of the last millennium BC at the latest, Egyptians had acquired a reputation among neighbouring peoples for excellence in knowledge of most kinds, which included that of magic. In the oldest surviving European literature, the poetry of Homer (probably from the eighth century BC), Helen of Troy, restored to her native Greece, puts a herbal potion into the drinks of her husband and his guests which has the power to remove all painful memories and banish all grief for a day. It has been given to her by an Egyptian, and Homer comments that this race is the most skilled of all in the use of herbs, as in all kinds of medicine. Greeks such as Homer did not distinguish between the chemical and arcane properties of herbs (and indeed often could not), and the properties of this drug clearly surpass those of straightforward chemistry.¹⁰ Eight hundred years later, the Jewish historian Josephus could declare that everybody regarded the Egyptians as representing the summit of all knowledge, including the arts of incantation and exorcism.¹¹ It is probable that this reputation was simply a natural consequence of the age, stability, wealth and sophistication of Egypt’s civilization, but Sir Wallis Budge, writing in the nineteenth century at the start of the systematic study of Egyptian magic, commented that the ancient reputation of the latter also derived from its remarkable degree of acceptance by, and integration into, its native society. He may well have been right.¹²

Mesopotamia

The civilizations of Sumer, Babylonia and Assyria had much in common with that of Egypt. They were also based in a great valley, in their case that of the two rivers Tigris and Euphrates which formed the plain that the Greeks called Mesopotamia and which is now in Iraq. They were likewise based on cities, with large temples staffed by a powerful priestly class, and centralized kingdoms, led by monarchs who were assumed to have a special relationship with the regional deities. They also showed a remarkable continuity over three millennia, despite cycles of stability and disorder in which particular dynasties rose and fell, sometimes precipitated or accompanied by foreign invaders. They had, too, a literate elite, who used texts composed in a standard script as a crucial component of government and religion. As part of this package of similarities, they also displayed a considerable interest in magic, partly as an aspect of official religion. This interest took forms that remained much the same throughout the whole historical period of ancient Mesopotamian culture, though evidence for it is most abundant in the first half of the final millennium BC. Attitudes to magic in Mesopotamia, however, also displayed striking differences from those in Egypt, so giving them their own strongly marked regional character.¹³

One of these differences was that Mesopotamians were more afraid of, and respectful to, their deities than Egyptians, and do not seem to have thought it possible to coerce or deceive them. Humans were not even believed to be capable of commanding spirits directly, being reliant on the help of deities to control lesser supernatural beings.¹⁴ Another difference was that they displayed a much keener interest in the influence of heavenly bodies on human affairs, thus becoming the originators of the Western tradition of astrology. By the third millennium BC they already believed that the stars and planets were associated with major deities and should decide the best time for important actions. By the first millennium astrological omens were used to predict the fates of kings and court astrologers reported regularly to rulers.¹⁵

A third major difference was that the peoples of Mesopotamia attached great importance to demons, in the sense of spirits inherent in the cosmos that were hostile to humans and a permanent menace to them, and so essentially evil. These were thought constantly to attack people, especially in their homes, and to be immune to physical barriers. Virtually all human misfortune, and especially disease, was credited to them and ritual action, both regular and ad hoc, was regarded as necessary to repel and expel them. Dealing with demons was the job of a priestly functionary called the āshipu, who mostly worked for private clients, with a mixture of incantations and actions addressed to deities, natural forces and the demons themselves. The rites included, as in Egypt, the use of figurines of wood and clay, often buried below buildings to protect them and their inhabitants, or destroyed to represent the beings that were the causes of affliction, or used as repositories for evil spirits exorcized from patients. Another similarity between the two was a willingness to import foreign ideas, in the Mesopotamian case by using spells in foreign languages. Like the Egyptians also, the inhabitants of Mesopotamia believed that to know the name of a supernatural entity was to acquire power over it; but unlike the Egyptians, they took pleasure in making long lists of demons with their characteristics. Virtually all the evidence that we possess for the practice of Mesopotamian magic consists of the records amassed by and for the āshipu: occasionally in those records mention is made of lower-grade kinds of magician, who operated among the common people – the ‘owl-man’, the ‘snake-charmer’ and ‘the woman who works magic in the street’ – but of these nothing else is known.¹⁶

The peoples of Sumeria, Babylonia and Assyria also believed in witches, in the classic sense of human beings, concealed inside their own society, who worked magic to harm others because they were inherently evil and associated with the demons which were the object of so much fear. The repertoire of the āshipu included many rites for undoing the harm that such people had wrought, while law codes prescribed death for those convicted of working such harm. The concern of the rites to avert witchcraft, however, was always to remove the affliction and not to detect the witch: indeed, the rituals themselves were supposed to bring about the death of the witch at whom they were aimed. In any case, most misfortune was blamed on angry deities, ghosts or (of course) demons. Mesopotamians also believed in the evil eye (and the evil mouth, tongue and sperm), and thought it to be destructive of both people and their livestock; it is not clear from the texts if this was thought to be voluntarily or involuntarily activated, but it was carefully distinguished from witchcraft. Actual witch trials seem to have been very rare, there are no recorded mass hunts, and the charge of witchcraft does not seem to have been a factor in political struggles. Witches were supposed to harm individual people, and not whole communities. The famous law code of the Babylonian king Hammurabi, from the early second millennium, allowed somebody accused of witchcraft to undergo the ordeal of jumping into a sacred river. If that person drowned, the charge was regarded as proven and the accuser inherited her or his estate; but if she or he survived then the accuser’s estate was handed over instead.¹⁷ It is possible, overall, that ancient Mesopotamian societies were among those in the world in which counter-magic used against presumed witchcraft was thought usually to be effective enough to remove the need to proceed against witches themselves.

The stereotypical witch mentioned in the sources is assumed to be female, which seems to match the generally low status of women in Mesopotamian society and make witchcraft an assumed weapon of the weak and marginalized. This suggestion is borne out by the other kinds of people associated with the practice of it: foreigners, actors, pedlars and low-grade magicians. In the few cases of actual prosecutions for witchcraft, which span the whole period of the various Babylonian and Assyrian monarchies, the accused were all women.¹⁸ Witchcraft was thought to be worked by the bewitchment of food or drink consumed by the victims, or of personal possessions or bodily waste taken from them (as across the world), or of images of them, or by the ritual making of real or symbolic knots. As in Egypt, destructive magic was regarded as a legitimate weapon if the cause were just, and kings formally cursed the enemies of the state. The secretive and malicious use of such magic was, none the less, clearly feared and hated in a way that was not apparent in Egypt. The texts used by the āshipu made the witch into a public enemy, capable of introducing chaos into the social order and even doing harm to deities. She was one of the menacing forces of the universe, along with foreign foes, wild animals, deities of other lands and wild lands, and (of course) demons. Fear of her seems to have increased gradually throughout ancient Mesopotamian history, so that by the mid-first millennium, she was ceasing at times to be regarded as a human being and becoming elided with a malevolent spirit of the night.¹⁹ Like the Egyptians, the peoples of Mesopotamia seemed to make no distinction between religion and magic (as defined earlier in this book), though with the difference that the rituals they performed to obtain their wishes were allowed and empowered directly by the deities whose aid they solicited. They did, however, make a major distinction between good and bad rites, good and bad practitioners of rites and good and bad superhuman beings.

Mesopotamian attitudes to magic seem to have been typical of an area far larger than Mesopotamia itself, extending from Asia Minor and Palestine in the west to the Indus Valley in the east. Three peoples found on the borders of this region, while reproducing most of those attitudes, also developed variations that were to be significant in the history of European witchcraft. The first of these were the Persians or Iranians, who occupied the region between Mesopotamia and India, and between the sixth century BC and the seventh century AD often controlled huge empires, which included Mesopotamia itself, and sometimes the whole remainder of Asia westward to the Mediterranean. By the late first millennium BC they had adopted the religion of Zoroaster, which depended on the concept of a cosmos divided between two mighty warring entities representing, respectively, essential good and essential evil. Lesser deities had become servants of these great beings, according to their dispositions, turning into the equivalents of angels and demons. Virtuous humans were likewise expected to choose the good supreme being and wicked humans the evil one. Among those who were regarded as automatic followers of evil were people who were believed to worship the demons who obeyed the Evil One and be rewarded by them with the ability to work destructive magic on others. Their rites were thought to be carried out at night, while naked. Putting a chronology on the development, or even the expression, of these ideas is very difficult, because the earliest records of them are in manuscripts dating from the thirteenth or fourteenth centuries AD, but containing texts written in the sixth and seventh centuries and based on originals which, from their language, were composed at various points dating back to the thirteenth century BC. Moreover, they were the work of the priesthood of the official religion, and give little indication of how ordinary people regarded the same issues, and of the manner in which suspected witches were treated in reality. It can be safely concluded, however, that the belief system expressed in them was fully formed by the late antique period, equivalent to that of the Roman Empire. By then, if not long before, witches were regarded by the Persians as the most evil of humans, who had to be fought (by protective and retaliatory priestly rites of the Mesopotamian sort) and punished in order to keep the land healthy.²⁰

At the opposite end of the Mesopotamian basin were the Hittites, who during the later second millennium BC developed a powerful and aggressive monarchy of their own, based on Asia Minor. Their culture also seems to have reproduced the attitudes to witches found in Mesopotamia, but with one significant difference: that the charge of being a witch was an important element in central politics at recurrent moments throughout Hittite history. This was one reflection of the Hittite tendency to try to concentrate magical power in the hands of the government, so that not only was witchcraft illegal but anybody thought to have knowledge of magic was to be brought to the royal palace for interrogation. The physical remains of purification rites carried out by priests or priestesses had to be burned at official places. Before 1500 BC, King Khattushili I banned his queen from keeping the company of certain priestesses who specialized in exorcism, and a couple of centuries later, a monarch accused Princess Ziplantawi of bewitching him and his family. In the late fourteenth century Khattushili III tried a governor for employing witches against him, and in the late thirteenth, Murshili II made the same charge against the current dowager queen.²¹

The final variation on the norm was found on the south-eastern fringe of the Mesopotamian world, among the Hebrews, who developed in the course of the first millennium BC an exceptional emphasis on one of their own gods, Yahweh, as the single deity whom they were henceforth permitted to honour. Spiritual power was therefore concentrated in the hands of priests and other holy men associated with Yahweh’s cult, and this had an impact on attitudes to magic. The Hebrew Bible applauds wonder-working prophets who serve Yahweh, above all Elijah and Elisha, even when they deploy their powers as expressions of personal vindictiveness. It invests the objects of Yahweh’s cult, especially his altar and the Ark of the Covenant, with intrinsic power, sometimes lethal. Joshua’s army stages an elaborate rite to draw on Yahweh’s power to bring down the walls of Jericho, and the god himself tells Moses to make a bronze serpent to protect his people from snakebite. The Mosaic Law includes a ceremony to determine the guilt of a woman accused of adultery by making her drink water mixed with sacred texts and dust from the Tabernacle floor (Numbers 5:11–31). All these could be termed ceremonies of a kind usually associated with magic, co-opted into the service of the official cult.

Unsurprisingly, in view of this, the Hebrew Bible also forbids recourse to magic and magicians outside that cult. It lists practitioners of magical services among the pagan Canaanites, calls them abominable and bids Hebrews turn to a prophet of Yahweh instead (Deuteronomy 18:9–22; cf. Leviticus 19:31, 20:6). Moses puts a Hebrew to death for cursing in Yahweh’s name (Leviticus 24:10–15), but the same curse is regarded as wholly acceptable when employed by a special instrument of the god such as Elisha (2 Kings 2:24). Saul is shown as behaving correctly when offering to pay a recognized Hebrew holy man to tell him the whereabouts of some lost donkeys (a classic service provided by magicians through the ages); and trying to learn Yahweh’s will through dreams and the god’s prophets (1 Samuel 9:1–10, and 28:15). However, when he hires a ba’a lot’ov, a female magician from outside the official cult (known in early modern English translation as the Witch of Endor), he does evil (1 Samuel 28:4–25). At one point the Mosaic Law ordered that a mekhashepa should not be permitted to survive, a passage officially translated in Jacobean England as ‘thou shalt not suffer a witch to live’ (Exodus 22:18). A proper understanding of this text would be possible if we knew exactly what a mekhashepa was, and was supposed to do: all we do know is that she was a specifically female practitioner of some kind of magic (and it is equally unclear from the language if she is expected to be killed or simply to cease living in the community, in other words, to be exiled). All in all, the Hebrew Bible does not spend much time on magic, as opposed to execrating the worship of other deities, and seems to make a classic incorporation of some forms of it into religion, by declaring that the same kinds of action were sanctioned for Hebrews if performed by accredited, and so divinely empowered, representatives of the one true god and execrated if offered by others.²²

A similar relative lack of interest in magic seems to obtain in the Second Temple period, between the late sixth century BC and the late first century AD, when Hebrews returned from exile in Babylonia, eventually to establish a new monarchical state in their Palestinian homeland with a single cult of Yahweh. Different strands of their literature continued to express a general animosity towards magic, as used by people not recognized as sanctified servants of the true god, but it was hardly debated. In the third and second centuries BC, the First Book of Enoch had fallen (and so corrupted) angels teach human women magic, especially by using plants (Book of the Watchers 1–36), but the Book of Jubilees (10:10–14) asserted that Yahweh sent angels to bind demons, which plagued humanity, and teach people the arts of healing, also especially by using plants. The Dead Sea Scrolls class apostasy from the true faith with magic, and order both to be punished. Exorcism, of demons from persons and places, by rites of the ancient Mesopotamian sort, remains the most commonly attested activity that may be categorized as magical.²³

More material survives from the succeeding period, following the Roman destruction of the Temple at Jerusalem and the dispersal of the Hebrew people, to complete their evolution into the Jews even as the cult of Yahweh completed its development into the religion of Judaism. The crucial collection of source material is the rabbinical literature composed between the second and seventh centuries, above all the Talmud, the collection of pronouncements and anecdotes to expound the faith, originally compiled in two separate documents in Babylonia and Palestine. The attitudes to magic expressed in it are not altogether coherent, but tend as before to credit outstanding holy men of the religion, now members of the official priesthood called rabbis, with the ability to work apparent miracles. These acts are always applauded, presumably as permitted and empowered by the true God, though this sanction for them is only sometimes made explicit. By contrast, anonymous women or heretics are treated as the natural practitioners of witchcraft, keshaphim, and are usually portrayed as being defeated by rabbis. Witches are not portrayed as having special looks or belonging to a special breed: they are just ordinary Jews, usually female, who have chosen to work harmful magic. At times they seem to operate in groups, with leaders. The Mishnah law code of c. AD 200 prescribed the death penalty for anybody who was judged to have bewitched somebody else with apparent genuine effect. How far it was actually enforced is hard to tell, as there are no references to witch trials in the same literature, or to rabbis having suspects put to death, save for the story of how rabbi Simeon ben Shetah and his followers killed eighty witches at Askelon in Palestine. That, however, has strong folkloric elements, which tell against it as a historical event: for example, it is described as having been necessary to lift all the women off the ground simultaneously to deprive them of their magical powers. The episode was supposed to have happened seven hundred years before the story is recorded. There are no accounts in the whole body of literature of financial compensation for witchcraft or references to it in divorce cases, so it is hard to see whether Talmudic stories describing the defeat of female magicians reflect social reality. One possible insight into that reality is provided by the metal bowls found buried in houses and cemeteries in Mesopotamia and western Iran, apparently dating from the fifth to the eighth centuries AD and made mostly, but not exclusively, for Jews. They are inscribed with spells to protect the owner against witchcraft and demons, and women rather than men, especially working in groups, are identified as the witches. However, women are also found as commissioning or making the bowls, and 90 per cent of the spells are directed against evil spirits alone, and not evil humans. In general, the Jewish literature of late antiquity rarely used magic as a polemical label for the religious practices of opponents, and attributed misfortune far more to the anger of the deity or the malice of demons than to witchcraft. None the less, it retained the traditional Mesopotamian belief in the existence of witches, who were generally presumed to be women.²⁴

Greece

The oldest European society from which evidence exists for attitudes to magic, including witchcraft, is the ancient Greek, which still has a far shorter recorded history than those of Egypt and Mesopotamia, extending back to the seventh or eighth centuries BC. By the fourth century at the latest the Greeks had developed their own distinctive set of beliefs, different again from any held in the great Near Eastern civilizations. One aspect of it was a distinction between religion and magic, often to the detriment of the latter, which was fundamentally that articulated at the opening of the present book, and indeed subsequently held by most Europeans until recent times.²⁵ It first appears in a medical tract concerned with epilepsy, On the Sacred Disease, which has been dated to the years around 400 BC and so may well push back the distinction concerned into the fifth century. This opposes the disreputable use of spells and medicines which seek to compel divine beings, ‘as if the power of the divine is defeated and enslaved by human cleverness’, to the legitimate actions of people who only supplicate for divine aid.²⁶ Shortly afterwards, the great Athenian philosopher Plato repeated it, attacking those who promised ‘to persuade the deities by bewitching them, as it were, with sacrifices, prayers and incantations’.²⁷ It seems, therefore, that by the central part of the classical age of Greek civilization, intellectuals, at least, were confidently articulating a matched pair of definitions that would become an enduring part of European culture. The opposition between them made by the Greeks was different from that which was applied in modern times – theirs was one between magic and normative religious practice, rather than between magic and religion as such – but it is still striking.²⁸ It was accompanied by hostility towards most categories of magician. The list of these remained more or less standard at Athens during the fifth and fourth centuries, in both stage plays and works of philosophy, and also in occasional other kinds of text: the agurtēs, a kind of wandering beggar-priest; the goēs, who is now generally assumed, from the linguistic relationships of the word, to have specialized in dealings with ghosts, either exorcizing them or setting them on people, and perhaps with other forms of spirit; the epoidos, or singer of incantations; the mantis, an expert in the revelation of hidden things, especially the future; and, most significant for future developments, the magos, who seems to have offered a range of services which incorporated most of those just described, and whose craft, mageia, became the root of the word ‘magic’. In addition to these was a string of lesser practitioners, usually noted (and condemned) in the plural: ‘oracle-mongers’ or ‘oracle-interpreters’; specialists who interpreted signs and portents; and those who ‘performed wonders’. They also included pharmakeis (masculine) or pharmakides (feminine), who seem to have specialized above all in potions; and rhizotomoi, ‘root-cutters’, who appear from their name to have worked a magic, and in modern terms also a medicine, based primarily on herbs. Greek writers did not use these terms with consistency, and the categories must have been very porous, each practitioner offering a personal portfolio of services that would often have overlapped them. Not all of the references were pejorative, the mantis, in particular, sometimes being lauded in inscriptions for helpful predictions or advice, but diviners and portent interpreters who received official approval tended also to have an official status. The Greeks lacked the powerful temple system, with its specialist priesthood, of Egypt and Mesopotamia, and their sources provide instead an insight into a flourishing world of popular magic that is barely visible in the Near Eastern texts. The very disapproval expressed by the surviving sources of that world testifies to the hold it exercised over the imaginations of many Greeks. On the other hand, it would be unwise to dismiss those expressions of disapproval as the grumbling of elitist intellectuals, attempting to reform popular beliefs. The playwrights, who also condemned magicians, had, after all, to please crowds, as the theatre at Athens was an art form with mass audiences.²⁹

It has become something approaching a consensus among experts that this hostility to magic appeared in Greece in the fifth century BC, as one response to a number of developments.³⁰ One of these was war with the Persians, which caused Greeks to define themselves more clearly against foreigners, and eastern foreigners in particular. Certainly the term magos, which gave rise to ‘magic’, was in origin the name for one of the official Persian priesthood, serving the Zoroastrian religion. Greek city states were also engaged in a greater definition of their own identities, with a new concept of citizenship, and (for some) new enterprises in imperialism. It is certainly true, in addition, that during the period of the late sixth and the fifth centuries, the Greek imagination became more interested in underworld deities and in the spirits of the dead, as entities with whom living humans might work for their own profit. It is also likely that some of the forms of magical practitioner condemned in the sources only appeared at this time. The work of the magos or goēs in exorcizing unwanted spirits does make these figures look remarkably like the Mesopotamian exorcist-priests, transplanted to Greece and turned into wandering and freelance operators.³¹ Furthermore, it is true that there is no securely dated and unequivocal condemnation of magic in Greek that can be placed before 450 BC.³² A faint note of caution may be entered, however, against this apparent absence of evidence, as sources for the subject are much scarcer before the fifth century, and those most relevant in the earlier period, such as plays and works of philosophy, are missing. The argument for a new attitude from around 450 BC has sometimes drawn attention to the apparent absence of any condemnation of magic in Homer’s poetry, and this may indeed be very significant. On the other hand, too much should not perhaps be inferred from one poet, and the kinds of magic apparently approved by him (scattered throughout the Odyssey) are those used by figures – a goddess or an official seer – or for purposes – healing or soothing – which later Greeks would also have been likely to find acceptable. It is possible that the classical Greek hostility to magic, when defined as an attempt to gain control over deities, has deeper roots than the fifth century.

What seems to be missing from this composite picture is witchcraft. There is no sense in any archaic or classical Greek text of hidden enemies within society who work destructive magic under the inspiration of evil. Plato called for the death penalty for any kind of magician who offered to harm people in exchange for financial reward, while those who tried to coerce deities, for any reason, should be gaoled. His targets, however, were service magicians offering morally and religiously dubious services in addition to the usual, theoretically benevolent, kind.³³ Furthermore, the fact that he needed to make this prescription perhaps indicates that no such laws already existed in his home city of Athens. There is no clear record of any trial of a person for working destructive magic in the whole of ancient Athenian history. There were a few in the fourth century of tragic women who gave men lethal poisons under the impression that those were love philtres. The same century also produced the case of a foreign woman who had settled in Athens, Theoris of Lemnos, whom the texts call a pharmakis, mantis, or a hiereia (priestess) and who was put to death, with her whole family, for asebeia, impiety. Unfortunately, the same texts do not allow firm conclusions to be drawn regarding the nature of her offence. One said that it was to provide ‘potions and incantations’ and another that she was an impious mantis, which together would make her a convincing victim of the Greek animosity towards many forms of magic. Another, however, accuses her of teaching slaves to deceive their owners.³⁴

The wider picture is equally enigmatic. Matthew Dickie has gathered hints that magicians were arrested and punished in Greek cities from the history of Herodotus, the drama of Euripides and a dialogue of Plato. He also points out, however, that such people were almost never practitioners of magic pure and simple, but doubled in other roles, such as priests, oracles or healers, so that their offences would be hard to match conclusively to magic.³⁵ One of the fables credited to Aesop tells approvingly how a female magician (gune magos) was sentenced to death for selling spells, which, she claimed, averted the anger of the deities, and so interfered with their wishes. This would be a perfect illustration of the Greeks’ horror of trying to coerce divine beings; but we do not know if it ever matched reality.³⁶ The city state of Teos passed a law that decreed capital punishment for any persons who made destructive pharmaka against its citizens, collectively or individually. It may be presumed that this term covered magic as well as chemical poisons, but it is not clear if the measure was enforced, and if it was intended against fellow citizens as well as outsiders. Rules for a private cult at the Greek city of Philadelphia in Asia Minor made members swear not to commit a list of antisocial acts, which included a list of magical practices; but vengeance was left to the gods.³⁷

Especially interesting and puzzling in this context are curse tablets, sheets of lead which are inscribed with spells or invocations to subject other humans to the author’s will, usually by binding, punishing or obstructing them. Their targets span all ages, and social roles and levels. Many call on the power of underworld or nocturnal deities or spirits, or the human dead, and they are found in pits, graves or tombs. They appear from the fifth century onwards, especially around Athens, and the set formulae used in many of them suggest either a widespread convention that most people understood or the use of professional or semi-professional magicians to make them. The very need for literacy indicates that specialists were involved. No known law forbids them, though Plato railed against them (and explicitly attributed them to hired magicians),³⁸ and scholars are divided over whether they were regarded as socially acceptable or would have been covered by the penalties against murder, assault and impiety. None of the authors seems concerned about the censure of society, though they do worry about the reactions of the spirit forms invoked. On the other hand, they certainly violated cultural norms, not merely by appealing to dark deities and ghosts but by using exotic names, retrograde writing and the reckoning of descent through female lines. It is hard to believe that they were ever a respectable means of mobilizing spiritual power, but how disreputable, illicit or illegal the resort to them was is completely unclear. In a different society, they would have been stereotypical acts of witchcraft, but they do not seem to have been classed as those in Greece.³⁹

It is of a piece with these patterns that there seem to be no clear representations of witches in archaic or classical Greek literature. Two characters in mythology bear some resemblance to them, as powerful female figures who work destructive magic: Circe and Medea. Circe uses a combination of a potion and a wand to turn men into animals, and Medea uses pharmaka for various magical ends, including murder. Neither, however, is human, Circe being explicitly a goddess, daughter of the sun and a sea nymph, while Medea is her niece, product of a union between Circe’s brother and either another ocean nymph or the goddess of magic, Hecate herself. Nor are they unequivocally evil, Circe becoming the lover and helper of the hero Odysseus once he overcomes her with the aid of the god Hermes, and Medea assisting and marrying the hero Jason. Medea certainly murders to help her beloved, and then again in an orgy of vengeance when he casts her off; but the attitudes of the Greek texts towards her remain ambivalent, and (like Circe) she escapes retribution for her actions. Both were to be immensely influential figures in later European literature, as ultimate ancestresses of many of its magic-wielding females; but it is difficult to see either, in their original context, as a witch as defined in this book.⁴⁰

Certainly magicians were stereotypically feminine in ancient Greek literature, and also assumed to come from the lower social ranks: it was believed that respectable women did not have the necessary knowledge. Sources that refer to the different kinds of service magician operating in real life, however, generally treat them as male: women were regarded as working mostly for their own personal benefit. Undoubtedly, the great majority of all attributed curse tablets were composed by men.⁴¹ There is a possibility that women were more vulnerable to actual prosecution for working or trying to work magic, to judge from the Athenian evidence, but it seems dangerous to generalize from such a small number of trials. When literary sources portray women as using magic, it is generally not of an aggressive kind, motivated by pure wickedness and aimed at subverting society in general, but a defensive variety, intended to win or retain a man’s affection or punish him for withdrawing it.⁴² There was a tradition, established by the fifth century BC and enduring till the end of antiquity, that Thessaly, the north-eastern part of Greece, was especially noted for pharmakidespowerful enough to drag the moon down from the sky at their command.⁴³ Why Thessalian women should have acquired this fearsome reputation has never been explained, though it signalled that the region was not quite part of Greece proper. Not much seems to have changed as the classical age of ancient Greek civilization, extending across the fifth and fourth centuries, gave way to the Hellenistic one after the conquests of Alexander the Great extended Greek culture around the whole of the eastern Mediterranean. The Sicilian poet Theocritus, who probably worked in Egypt, produced an enduringly famous work about an Alexandrian woman enacting a rite to retrieve or punish a faithless boyfriend with the help of her maid. It was entitled the Pharmakeutria, but shows how much the concept of pharmaka had moved beyond drugs, because her methods consist entirely of a mixture of incantations, material substances of different kinds and special tools.

The one distinctly new development of the age was the educated collector of magical lore, publishing books on the arcane properties of animal, vegetable and mineral substances. Like Hellenistic culture itself, this took in lands far beyond Greece, and indeed drew heavily on the accumulated traditions of Egypt, Mesopotamia and Syria: the first and most famous author in the genre, Bolus of Mendes, came from Egypt.⁴⁴ Accordingly, save that they are written in the Greek language, there is some question over how far such texts can be termed Greek. Otherwise, in the older Greek world itself, things look like classical business as usual, and it is equally hard to find the witch figure in it. An oracle from Claros in Asia Minor responded to a city in the region which blamed an outbreak of plague on an evil magos and wanted a remedy: the answer was to destroy the wax figurines used by the magician by invoking the power of the goddess Artemis. There is no indication of whether any action was also intended against an actual person identified as the culprit. Likewise, a lead tablet offered to the oracle at Dodona to the north-west of Greece posed the question, ‘Did Timo harm Aristobola magically?’ but we cannot tell who Timo was, and what would have been done if the reply had been affirmative.⁴⁵

Rome

The pagan Romans, both in their republican and imperial periods, were heavily influenced by Greek culture, and it is not surprising to find them embracing the same distinction between religion and magic; though it may equally be argued that the distinction concerned must have appealed to their own attitudes for it to have taken root. In the first century of the imperial period, and Christian era, both the playwright and philosopher Seneca and the scholar Pliny condemned magic as a wish to give orders to deities.⁴⁶ In the third century, the biographer of the holy man Apollonius of Tyana portrayed his hero as securing his acquittal from a charge of being a magician by claiming that he merely prayed to the god Heracles, who answered his plea.⁴⁷ Apuleius of Madaura, tried on a similar accusation, defended himself by contrasting somebody like him, who obeyed the deities, with a genuine magician who was popularly believed to ‘have the power to do everything that he wanted by the mysterious force of certain incantations’.⁴⁸ In the same century the Neoplatonist philosopher Plotinus accused some rivals of making incantations designed to draw down the higher divine powers to serve them.⁴⁹

There seems to be a consensus among historians that Roman attitudes to magic crystallized between the last century BC and the first one AD, and had hardened into legal and social convention by AD 250.⁵⁰ Some have attributed this development to a new desire to categorize varieties of religious experience, and to exclude activities which appeared to conflict with normative expressions of it. This has been related in turn to a greater definition of outsiders and threats to society produced by the incorporation of more and more people, and classes of people, within Roman citizenship. Others have laid more emphasis on the appearance of an individualized elite soaked in Greek ideas. The definition of the magician as an outsider and menace is generally reckoned to have appeared between the time of writers such as the poet Catullus and the politician and scholar Cicero in the mid-first century BC and that of Pliny in the late first century AD. The former still used the term magus, taken from the Greek, in its original sense of a Persian priest, but none the less Cicero could speak of the invocation of underworld spirits as a new and perverse religious practice. Pliny lambasted ‘magia’, the practices of the magi, as ‘the most fraudulent of crafts’, designed to ‘give commands to the gods’ by discovering and wielding the occult powers within the natural world, and traced its progress from Persia and the Hebrews through Greece to the Roman world, so emphasizing its foreign origin as well as its pernicious character.⁵¹

At first sight, the development of Roman law seems to follow the same trajectory.⁵² The code of the Twelve Tables, from the early republic – as far as it can be reconstructed from later evidence – forbade the specific act of luring away the profit of crops from somebody else’s land to one’s own, as an infringement of property rights; but did not specify that the means was by magic. It also outlawed ‘an evil song’, which could signify a magical incantation or merely an insult. Likewise, the Lex Cornelia of 81 BC forbade a number of means of killing by stealth, one of which was veneficium, a term which to modern eyes has the same frustrating imprecision as the Greek pharmaka, as it could mean both poison and magic: once more, ancient people would in practice often find it impossible to distinguish between the two. Things become clearer only in the second century AD, when the work of the magus in general became equated with veneficium and with maleficium, meaning the intentional causing of harm to others. By the third century, Roman law codes were adapting to this change, extending the Lex Cornelia to cover the making of love potions, the enactment of rites to enchant, bind or restrain, the possession of books containing magical recipes, and the ‘arts of magic’ in general. To own such a book now meant death for the poor and exile for the rich (with loss of property), while to practise magical rites incurred the death penalty, with those who offered them for money being burned alive. As the possession of books and the provision of commercial services were activities that could readily be proved objectively, these were relatively easy laws to enforce.

Two major problems attend any attempt to understand the actual status of magic in the Roman world. One is that very little information survives on how these laws were actually enforced; the other, that it was perfectly possible to conduct witch-hunts without having any law against magic itself, if the victims were accused simply of committing murder by magical means. It was recorded, centuries later, that in 331 BC an epidemic hit Rome, with high mortality, and over 170 female citizens, two of them noblewomen, were put to death for causing it with veneficium. This may have meant straightforward potions, as the first suspects, having claimed to be healers instead, were made to drink their alleged medicine and died, triggering the mass arrests. The years 184 to 180 BC were also a time of epidemic disease in Italy, and much bigger trials were held in provincial towns, claiming over two thousand victims in the first wave and over three thousand in the second. Again the charge was veneficium, and it is impossible to tell whether this meant poisoning in the straightforward sense, or killing by magical rites, or a mixture.⁵³ If the second or third sense of the word was what counted, and the reports are accurate, then the republican Romans hunted witches on a scale unknown anywhere else in the ancient world, and at any other time in European history, as the body counts recorded – however imprecise – surpass anything in a single wave of early modern trials.⁵⁴ Nothing like this is known under the pagan Roman Empire, but individuals were certainly prosecuted then for working magic, whether or not it resulted in physical harm to anybody else. The (alleged) case of Apollonius and the (historical) one of Apuleius have already been cited: the former was accused of using divinatory rites to predict a plague, and the latter of securing a woman’s love by casting a spell. Hadrian of Tyre, a second-century legal expert, was quoted as pronouncing that those who offered pharmaka for hire (he wrote in Greek) should be punished ‘simply because we hate their power, because each of them has a natural poison’, and they offer a ‘craft of injury’.⁵⁵ The Chronicle of the Year 354 claimed that Tiberius, the second emperor of Rome in the early first century AD, executed forty-five male and eighty-five female veneficiarii and maleficiin the course of his reign.⁵⁶ Again the terminology is cloudy, and the terms could just mean respectively poison-sellers and criminals in general; but if so, the totals seem far too low, and it is more likely that magicians are intended. Furthermore, whether or not the Romans were hunting witches by the opening of the imperial period, they were certainly imagining them in a way that the Egyptians and Greeks did not, but the Mesopotamians, Persians, Hittites and Hebrews did. Indeed, the literary images that they produced were the main ancient source cited by early modern authors to prove the long existence of the menace from witchcraft. Some of them echo, and indeed copy, the Greek and Hellenistic model of the lovelorn woman seeking to use magic to secure or retrieve a partner.⁵⁷

In addition, however, there are characters that have no parallel in Greek literature: women who habitually work a powerful and evil magic, using disgusting materials and rites and invoking underworld and nocturnal deities and spirits, and human ghosts. They appear in the later first century BC and continue into the later centuries of the empire. Such is Horace’s Canidia, a hag who poisons food with her own breath and viper’s blood, has ‘books of incantations’, and enacts rites with her accomplices to manufacture love potions or blight those who have offended her. They burn materials such as twigs grown from tombs, owl feathers and eggs, toad’s blood and venomous herbs, and tear a black lamb apart, as offerings to the powers of night. They also make images of people, and murder a child to use his body parts in their concoctions. It is predicted that they will be stoned to death by a mob, and their bodies left for animals to eat.⁵⁸ Another is Lucan’s Erictho, another repulsive old woman who understands ‘the mysteries of the magicians which the gods abominate’, because they can ‘bind the reluctant deity’. Even ordinary practitioners, Lucan assures his audience, can induce helpless love, stop the sun in its course, bring rain, halt tides and rivers, tame beasts of prey and pull down the moon. Erictho possesses in addition the ability to learn the future by reanimating corpses with a potion of dog’s froth, lynx entrails, hyena hump, stag marrow, chunk of sea monster, dragons’ eyes, stones from eagles’ nests, serpents and deadly herbs: like Canidia’s, her mixtures are the apotheosis of veneficium. Like her, too, she practises human sacrifice, but on a grander scale, even cutting children from wombs to offer up burnt on altars.⁵⁹ The rituals of both women invert all the norms of conventional religious practice. Similar figures appear in the work of other poets, though drawn in less detail. Virgil wrote of a foreign priestess with the power to inflict joy or agony on other humans by her spells, reverse the movements of rivers and stars, make trees march and the earth bellow, and summon the spirits of darkness.⁶⁰ Ovid produced a drunken hag called Dipsas, who understood the power of herbs and magical tools and could control the weather, raise the dead and make stars drip blood and turn the moon red, as well as performing the usual trick with rivers.⁶¹ Propertius’ equivalent was Acanthis, whose potions could make a magnet fail to attract iron, a mother bird abandon her chicks and the most faithful woman betray her husband. She also had the power to move the moon at her will, and turn herself into a wolf.⁶² Tibullus’ equivalent, called a saga, could perform the same feats with the moon, rivers, the weather and the dead, and (with an incantation) deceive the eyes: Roman witches inverted the natural as well as the religious order.⁶³ Apuleius, himself no stranger to charges of magic, put a range of magic-working women of different ages and degrees of wickedness and power into a novel. They are all murderous, lecherous or sacrilegious, and use magic to get their way: Apuleius remarks that women as a sex often do this. His most terrifying such invention, Meroe, can lower the level of the sky, stop the planet turning, melt mountains, put out the stars and summon the deities. Routinely, his witches can change their own shapes, and those of others, into animals.⁶⁴ In the same period, an older character occasionally got the same makeover: above all Medea, whom Roman poets and playwrights transformed into a darker figure, performing the kind of elaborate nocturnal rites, to dark powers, credited to these witch figures.⁶⁵ The satirist Petronius testified to the familiarity of the stereotype of the terrifying and mighty witch by the mid-first century AD, by sending it up in his own novel. The anti-hero, needing a cure for sexual impotence, which he has blamed on witchcraft, turns for help to an old priestess who boasts of possessing all the powers over nature attributed to Canidia, Erictho, Meroe and their kind, yet who lives in poverty and squalor and proves farcically to be a charlatan.⁶⁶

It may fairly be wondered whether any of this was intended to be taken as seriously at the time as early modern demonologists were later to take it. These are, after all, literary inventions appearing in genres equivalent to romantic fantasy, Gothic fiction, satire and comedy. A major element of preposterous exaggeration was plainly present. On the other hand, such images – of potent magic worked by evil women – would not have been chosen had they not resonated to some extent with the prejudices and preconceptions of the intended audience. Kimberley Stratton has plausibly linked their appearance to a concern with the perceived sexual licence and luxury of Roman women in the same period, combined with an ideal of female chastity as an indicator of social stability and order. The image of the witch, in her view, emerged as the antithesis of this idealized and politicized version of female behaviour.⁶⁷ However persuasive an argument, this still needs once again to take into account the likelihood that the image flowered so rapidly and luxuriantly because it was planted in soil made fertile for it. After all, the Romans who produced and consumed it had a historical memory of having put to death almost two hundred women in their city, centuries before, for having deliberately produced a major epidemic that claimed huge numbers of lives, by using veneficium. According to medical realities, all of them would have been innocent of this offence, and so their society would already have needed to believe in the capacity and will of women to commit it. An unknown number of women, perhaps the majority, would have been among the thousands of victims of the mass trials for the same crime in the 180s BC. Rome therefore already had a sense of wicked women as agents of murder and social disruption who used hidden means. Likewise, though with much more muted consequences both in social reality and in literature, it must be significant that when the Greeks conceived of divine or semi-divine figures who used dangerous magic, like Circe and Medea, these were female. It seems that cultures which had defined magic as an illicit, disreputable and impious activity, and in which women were excluded from most political and social power, such as the Greek and Roman (and Hebrew and Mesopotamian), were inclined to bring the two together into a single stereotype of the menacing Other. In the Roman case, however, the results, both practical and literary, were the most dramatic.

It is not clear how much a belief in the ‘evil eye’ tempered, or meshed with, Rome’s fear of witchcraft. Romans certainly had one. Virgil described a shepherd who blamed ‘the eye’ for a sickness in his flock, while Pliny and Varro both wrote of the use of amulets to avert it from children and Pliny of the efficacy of spitting three times to break its power. Pliny and Plutarch discuss the details of belief in it, and testify that it was believed to be wielded both deliberately and inadvertently, and the willingness to recognize the latter effect should in theory have damped down one to blame witches for injury. They also record, however, that it was thought to be a special property of foreigners, fathers and women with double pupils (a rare condition). This would not have affected the great majority of people suspected of witchcraft, and so the belief would probably not, in practice, have mitigated very much.⁶⁸ One further indication that the Romans took witchcraft seriously is that they became the first people in Europe and the Near East since the Hittites, over a thousand years before, recurrently to make accusations of it as a political weapon. These studded the first two reigns of the first imperial dynasty, as a feature of its attempts to establish its authority and the stability of the state. The first emperor, Augustus, linked that stability to a suspicion of magicians by declaring war on all unofficial attempts to predict the future, which might serve to encourage people in disruptive political ambitions. One decree early in his reign ordered the expulsion of all goētes and magoi from the city, and the retention only of the native and traditional forms of seeking oracles from the deities and natural world. This measure was repeated nine times in the following hundred years (testifying either to its continued importance or to its lack of effect), and Augustus was said to have ordered the burning of more than two thousand books of prophetic writings by unauthorized persons.⁶⁹ The next emperor, Tiberius, drove a prominent senator to suicide by investigating charges against the man of consulting magi and Babylonian astrologers and trying to summon underworld spirits with ‘incantations’. Another leading senator then resorted to suicide when accused of implication in the death of the heir presumptive to the throne, Germanicus, by having remains of human bodies, curse tablets, charred and blood-stained ashes and other instruments of evil magic concealed about the prince’s house. It was claimed that the empress dowager Livia, Tiberius’ mother, had accused a friend of her hated step-granddaughter Agrippina of veneficia, while Agrippina’s own daughter, who bore her name, charged three rivals with using magic, one of whom was also driven to suicide. A noblewoman was accused of driving her husband insane with ‘incantations and potions’. After that, such charges disappeared from high politics, only to reappear spectacularly in the fourth century, when two more dynasties struggled to establish themselves, the Flavian and that of Valentinian. They seem to have been a feature solely of that occasional phenomenon in Roman history, the protracted stabilization of a new imperial family.⁷⁰

It remains to consider the evidence for the actual working of destructive magic, and the fear of it, among people in the Roman Empire. By its very nature, this is relatively sparse, but it exists; and that is even when excluding for the time being the texts of complex ceremonial magic, which will be the concern of a later chapter. Curse tablets persisted in their original Greek homeland, and spread out from it across much of the empire.⁷¹ An inscribed lead tablet placed in the tomb of a woman at Larzac, in southern France, alleged the existence of two rival groups of ‘women endowed with magic’. One had cast malicious spells against the other by ‘sticking’ or ‘pricking’, perhaps of images of their intended victims, and been foiled by counter-magic aided by an additional ‘wise woman’: the person in the tomb may have been one of the presumed casualties of the conflict, or one of those who had triumphed in it, or the tablet may simply have been lodged in a grave as part of a rite to work with the dead or their deities.⁷² Andrew Wilburn has conducted a study of the material evidence for the working of magic, and especially of curses, at three sites, in the Roman provinces of Egypt, Cyprus and Spain respectively. His conclusion is that the cursing of opponents or oppressors, often using the services of experts, was a regular and important aspect of life under imperial rule, though not a respectable one.⁷³ It seems to substantiate the famous declaration of the Roman scholar and administrator Pliny, in the first century, that ‘nobody is unafraid of falling victim to an evil spell’.⁷⁴

It is also clear that people identified individuals whom they loved as having fallen victim to such spells. In the 20s AD, the princess Livia Julia, daughter-in-law of the emperor Tiberius, left an inscription to lament the loss of her little slave boy, whom she believed to have been either killed or abducted by a saga, one of the Roman words for a female magician.⁷⁵ This is vivid testimony that the accusations of witchcraft made by the imperial family during that reign were not merely the product of cynical political opportunism. One of the leading scholars of Greek and Roman magic in recent decades, Fritz Graf, has made a systematic study of epitaphs similar to that commissioned by Livia Julia, for people, usually young, who were thought to have been killed by magic. He found thirty-five, most from the eastern, Greek half of the empire and from the second and third centuries. They were not common among tomb inscriptions for young persons, suggesting that untimely death was not usually ascribed to witchcraft, and they called for divine vengeance upon the assailants, who were sometimes unknown and sometimes suspected and named (women being only slightly more common than men among these suspects). Graf suggested that this tactic of appealing to deities averted the need for legal accusations.⁷⁶

How far this is true is hard to tell given the lack of surviving legal archives from the pagan period of the empire. We can be reasonably certain that there were no mass hunts and trials, because these would surely have left traces in historical records. It is much harder to judge whether or not individual accusations often reached the courts, and if so how seriously they were treated. Occasionally one is revealed by a chance survival of evidence, such as the Egyptian papyrus which records that a farmer in the Fayum had denounced neighbours to the local governor for using magic to steal his crops. The outcome of the case is unknown.⁷⁷ Livia Julia’s passionate outburst on the death of her slave boy included an appeal to Roman mothers to protect their own children against such evil spells. It is only possible to wonder whether they needed such a warning, and if they shared the same reaction when their children died or disappeared, and what steps, if any, they took against the presumed killers.

The Night-demoness

Some societies in different parts of the world have held two concurrent concepts of the witch, one taking the form of a theoretical being, which operates by night and performs effectively superhuman feats, and one representing genuine human beings who are suspected and accused of witchcraft in day-to-day life. The Tswana of Botswana, for example, distinguished ‘night witches’ from ‘day sorcerers’. The former were supposed to be evil old women, who gathered at night in small groups and went, naked and smeared with white ash or human blood, around homesteads to harm the inhabitants. They were said to get through locked doors, having thrown the inmates into a deep sleep. In practice, these beings were treated as more or less fictional, few claiming to have seen them and many openly refusing to believe in them. The ‘day sorcerers’ were ordinary members of the tribe who were supposed to try to harm personal enemies with combinations of spells and material substances. Everybody believed in them.⁷⁸ On the far side of the Old World, in the Trobriand Islands, the inhabitants spoke of women who flew around at night naked, but invisible to victims, met on reefs in the sea to plot evil, and removed organs from living humans for cannibal feasts, thus blighting their victims with weakness and illness. They also believed that certain male members of the community learned how to use a combination of magic, natural materials and animal helpers to inflict illness and death on chosen victims. It was the latter who were feared in everyday life, while the former were held responsible for occasional major catastrophes such as epidemics.⁷⁹ Such dual belief systems have been fairly common, though not ubiquitous, among societies that have believed in witchcraft.

The ancient Romans were one people who possessed such a thought system, and in doing so tapped into another well-scattered aspect of human belief, the tendency to associate witches with owls. This family of birds has, after all, five features that people often find sinister: nocturnal habits, silent movements, predation, a direct stare and an ability to turn the head completely round. In the Native American languages of the Cherokee and the Menominee the word for the owl and the witch is the same, and the belief that witches could take the form of owls was found from Peru to Alaska. Even more widespread is the idea that owls, or humans in their shape, were responsible for the ubiquitous human tragedy of the sudden, unexpected and mysterious sickening and deaths of babies and small children. It was found among many North American peoples, but also in Central and West Africa and Malaya.⁸⁰ This was also a feature of Roman culture, but as one corner of a complex of ideas spanning the Near East and Mediterranean, which also allows us some opportunity to penetrate the thought world of pagan Germany.

This complex is first revealed in Mesopotamia, where by the early second millennium BC the lists of demons or ghosts compiled in the rites of purification and exorcism include a closely related group with seven different names that share the component lil. The first four were female, the last three male. They seem to have been erotic spirits who coupled with humans in their dreams, wearing them out and tormenting them. By the first millennium they also appear to have been regarded as dangers to women in childbirth, though the Mesopotamian demoness who was the special enemy of infants, pregnant women and new mothers was a lioness-headed one called Lamashtu.⁸¹ A Phoenician exorcism text of the seventh century BC calls a lili ‘flyer in a dark chamber’, which would suit these roles, and portrays her as a winged sphinx.⁸² In the Hebrew Bible there is a famous reference to a lilith (at Isaiah 34:14), in a list of beings that haunt a land devastated by divine anger; but it has been suggested that as the others in the list are genuine wild animals, the lilith may here signify a night bird, probably, indeed, the ‘screech owl’ of the King James translation.⁸³If so, the linguistic connection between nocturnal demons or malevolent ghosts, and a nocturnal bird, is suggestive for what is to follow.

These demons or ghosts represent the strongest and most convincing continuity between the belief system of ancient Mesopotamia and that of the incantation bowls fashioned in the same region between (probably) AD 400 and 800. It has been mentioned that 90 per cent of the protective spells upon them were aimed at demons, rather than humans, and about half of these were ‘liliths’ and lilin. The former were female, and preserved the dual character of the older lil- spirits, of coming to men in erotic dreams and endangering women, as virgins, during menstruation, and at conception, pregnancy and childbirth, along with their infants: this was because a ‘lilith’ regarded herself as the true paramour of the man on whom she preyed sexually, and so treated his human wife and their children with murderous jealousy. The male lilin brought erotic dreams to women. A drawing and a few inscriptions indicate the appearance of a lilith, as a young naked woman with long dishevelled hair and prominent breasts and genitals: in her aggressive and immodest sexuality and unkempt, wild state, the antithesis of the well-behaved Jewish wife or daughter of the age. Sometimes the same figure is mentioned on the bowls as a single being, Lilith, who also features in texts of the Talmud, which agree upon her long hair, and one of which also credits her with wings. In an eighth-century Jewish text, the Alphabet of Ben Sira, Lilith suddenly took a quantum leap in her mythological persona, being given a back story as the first wife of Adam and integrated into the Hebrew Bible. She was on her way to becoming the most feared demon of Judaism and one of the great imagined figures of the Western world.⁸⁴

The Greeks spoke of various child-killing demons abroad at night, called (in the singular) mormō, mormoluke, gellō and lamia, who were, as in Mesopotamia, also dangerous to young women, on the eve of marriage or while or after giving birth. The lamiai were furthermore credited with being predators on young men, whom they tempted sexually before devouring. Most – in the Greek style – were made into personalities with their own myths, in which they generally featured as human women who had died prematurely or lost their own children tragically.⁸⁵ Their similarity to the spirits of Mesopotamia may be the result of direct transference, as Lamashtu may be the original of lamia, and a type of Mesopotamian demon called a galla may stand behind gellō, though this is not conclusively proved.⁸⁶ It is the Romans, however, who are most significant for the present enquiry, for the most commonly mentioned of their child-murdering horrors was the strix (plural form striges or strigae), a figure which they had given to the Greeks by the last centuries BC. What was distinctive about the strix was that, whereas the Greek monsters were like ugly humans or serpents, it strongly resembled an owl, or (to a lesser extent) a bat, being a winged, clawed creature, which flew by night and had a hideous screeching cry. The resemblance was the stronger in that the Romans sometimes seemed to use the same name for a species of actual screeching owl.⁸⁷The actions and role of the strix varied from teller to teller, but everybody agreed that it was bad news. To some it was a creature of ill omen, a harbinger of foreign and civil war, which hung upside down. Its main function, however, was to prey on young children at night, weakening or killing them by feeding on their blood, life force or internal organs. When a victim was dead, it could eat the corpse. Unlike the child-killing monsters from further east, it seems to have had no inherent connection with sexuality. The would-be scientist Pliny was not sure whether striges were genuine or fictional creatures, while the poet Horace mocked the belief in them; and indeed they feature almost wholly in works of imaginative literature, and not in law codes or histories. On the other hand, in the seventh century AD John Damascenus could still note that the common people of his time, despite the teachings of Christianity, still believed in ghosts and striges, which slipped into locked houses and strangled sleeping infants.

One thing that was distinctively Roman about the strix was its connection with witchcraft, which was based on a quality of Roman witches, mentioned earlier, that was not shared by their Near and Middle Eastern counterparts: they were shape-shifters, able to change into animal forms in order to go abroad. This opened the possibility that striges were in fact temporarily transformed witches. Ovid’s Dipsas possessed, in addition to all her other powers, that of flying at night clad in feathers. In another work, Ovid left open the question of whether striges were actual birds, or crones who had been turned into bird form by spells.⁸⁸ By the first century AD, the grammarian Sextus Pompeius Festus could define strigae simply as ‘the name given to women who practise magic, and are also called flying women’.⁸⁹ Novelists subsequently took advantage of the idea, as both Lucian (or somebody writing like him) and Apuleius produced accounts of a woman undressing, rubbing her body with ointment while carrying out particular rites, and then turning into an owl and flying out through a window into the night.⁹⁰ As in each case she is going in quest of a lover, and is a highly sexed woman in her prime, fond of young men and inclined to destroy those who reject her, this image supplies the link between striges and predatory sexuality which was missing earlier; and so fits them more neatly into the wider pattern of belief in such figures stretching as far as Mesopotamia. Apuleius, indeed, collapsed distinctions further, by twice referring to one of his human witches as a lamia.⁹¹

This evolving complex of beliefs may also supply a key to unlock the attitudes to witchcraft of the pagan Germanic tribes that lived to the north of the Roman Empire, and broke into it from the late fourth century onwards to conquer its western half and substitute successor kingdoms of their own.⁹² To do so involves stepping far across the conventional boundary between ancient and medieval, to consider some later texts which cast light on earlier belief. One of the steps taken in the formation of those successor kingdoms was usually the proclamation of a law code, in the Roman manner and in the Roman language, of Latin. The oldest of these to survive, and possibly the first to be issued, is that of King Clovis of the Franks, for the northern part of his kingdom, later France, in 507–11. It is accordingly the least Romanized and, as it was created only a couple of decades after the Franks accepted Christianity, retains many echoes of pagan culture. Two of its clauses apparently concern bad magic. One prescribes a huge fine for anybody who commits maleficia against another person, or kills with a potion: both acts were probably regarded here as magical, as they are in the Roman codes being copied. The other imposes the same heavy penalty on a stria who ‘eats’ a person. This is a version of the word strix and suggests a native belief in a night-roaming female who consumes a person’s life magically. Two more clauses provide further information. One fines anybody who falsely accuses anyone of being a herburgius, and glosses the term as meaning ‘one who carries a cauldron to where striae do their cooking’. This suggests that the nocturnal women concerned were believed to meet to cook and eat the fat or organs they subtracted from their victims: a tradition also found in parts of Africa. A larger fine was directed for anybody who falsely called a freeborn woman a stria, which was clearly a very serious insult.⁹³ Other Germanic tribes had the same fear of the same figure. The law code of the Alamanni, from the early seventh century, ordered fining for a woman who called another a stria.⁹⁴ In the same period the Lombard king Rothari fined anybody who called a young woman over whom they stood as guardian a striga or masca, presumably at times to lay hands on her inheritance: the second term, ‘masked one’, could be another word for the same being, or for a different magician who harms by stealth. Another fine was inflicted on anybody who killed another person’s female servant or slave for being a striga or masca, ‘because in no way should Christian minds believe that a woman can eat up a living human being from within’.⁹⁵ In 789 the first Holy Roman Emperor to emerge from the Germanic peoples, Charlemagne, informed the recently conquered Saxons, who were undergoing a forcible conversion to Christianity, that, ‘If anyone, deceived by the Devil, believes, as is customary among pagans, that any man or woman is a striga, and eats men, and shall on that account burn that person to death or eat his or her flesh, he shall be executed.’⁹⁶

It may be noted that the successor kingdoms had adopted the attitude of learned Romans: that beings such as the strix probably did not exist. The codes are themselves, however, testimony that in pagan times all classes of German society had believed in them. This was a significant difference from the Roman situation, and, moreover, while the Roman strix was primarily a danger to children, the Germanic equivalent allegedly attacked people of all ages. That belief continued as a tenacious popular tradition far into the Middle Ages.⁹⁷ Around the year 1000 the Swiss monk Notker Labeo commented that whereas savage foreign tribes were said to practise cannibalism, ‘here at home’ witches were said to do the same.⁹⁸ Soon afterwards, Burchard of Worms prescribed a penance for the belief among women that while their bodies lay in their beds at night they could go out as spirits through closed doors to join other women of the same kind. They would then band together to kill people and cook and eat their organs, restoring them to a brief and weakened life by substituting straw or wooden replicas for the parts that had been taken. He thought this a devilish delusion.⁹⁹ In the early thirteenth century Gervase of Tilbury dismissed as the product of hallucinations a tradition found in Germany and France that women known as lamiae, mascae or striae flew by night across great distances to enter the homes of chosen victims, dissolve their bones inside their bodies, suck their blood, and steal their babies.¹⁰⁰

Older sources may help to reconstruct more of the cultural context in which such ideas played a part. A succession of Roman authors noted that the Germans credited women with especial powers as diviners and prophetesses. Julius Caesar heard that a German army had not attacked him as expected because the ‘matrons’ in it had pronounced that battle would be unfortunate if joined before the next full moon: Caesar added that it was customary for such women to make such decisions by ‘lots and divination’.¹⁰¹ In the next century the Roman historian Tacitus reported that Germans regarded women as ‘endowed with something celestial’, which gave them the power to see the future, and that several of their most famous prophetesses had been venerated almost as goddesses.¹⁰² A succession of such figures is recorded among the Germanic tribes and kingdoms by other ancient and early medieval historians, causing the nineteenth-century folklorist Jacob Grimm to suggest that ancient German culture had invested females with greater inherent powers than males, both for divination and for magic in general – which would certainly make a fit with the fear of night-roaming, flesh-eating women.¹⁰³

What is missing from this picture is any sense of how good and bad magic in general were conceived and gendered among the pagan Germanic peoples. Likewise, there is far too little information on how magic and its practitioners were characterized, deployed and averted among ordinary people in the pagan Roman Empire. We have some striking images of certain kinds of it, and them, which are significant and allow some conclusions to be drawn, but large areas of relevant knowledge are absent.

Summary

It may be argued from all the data above that when magic is the subject under scrutiny, the ancient European world can indeed be divided into different regions, with contrasting attitudes and traditions. The Egyptians made no distinction between religion and magic, did not distinguish demons as a class of supernatural being, and had no concept of the witch figure. The Mesopotamians feared both demons and witches, and the Persians combined this fear with a division of the cosmos into opposed good and evil powers, the Hittites introduced it into high political life, and the Hebrews blended it with a belief in a single, good, deity with a single permissible cult. The Greeks (or at least some of them) made a distinction between religion and magic, to the detriment of the latter and some of its practitioners, but do not seem to have had an idea of witchcraft. The Romans made the same distinction, and accompanied it with a vivid concept of both witchcraft and witches, which extended to a criminalization of many forms of magic. The Germans feared a mythical sect of night-flying cannibal witches, which projected into real life, and criminal prosecution, a much more widespread mythology – found as far as Mesopotamia – about nocturnal demonesses.

It is possible to draw a simple and crude conclusion from all this: that the early modern witch trials derived ultimately from the fact that Western Christianity managed to blend the Mesopotamian belief in demons and witches, the Persian one in a stark cosmic dualism, the Hebrew one in a single true, jealous and ultimately all-powerful deity, the Greek one in a difference between religion and magic, the Roman one in witches and (perhaps) the need for witch-hunts in times of especial need, and the German one in night-flying, murderous human cannibals who were mostly or wholly female. There would be some truth in such an idea, but it would ignore a whole range of complications and subtleties which are needed to explain why the early modern trials took so long to occur after the triumph of Christianity across most of Europe, why they were relatively short-lived and why they happened when and where they did. Instead, ancient traditions played an important role in the formation of European witchcraft beliefs in more complicated and subtle ways, and over a long period of time: and this process will be the subject matter of most of the rest of this book.