Introduction: the ubiquity of magic?

It is usually assumed that belief in magic was ubiquitous in the early Roman empire,^[1] that, in the words of Pliny the Elder, “there is no one who does not fear to be spellbound by curse tablets”.^[2] One needs only read the accounts of the famous trials for sorcery of Apollonius of Tyana^[3] and Apuleius of Madaura,^[4] or the magical explanations given for the untimely demise of Germanicus, Tiberius’ popular heir,^[5] to see how significant magic appears to have been. Indeed, the only fully extant novel in Latin that we possess,^[6] Apuleius’ Metamorphoses, is concerned with the consequences^[7] of meddling in such things.^[8] Homer’s Odyssey, one of the formative texts for most of the inhabitants of the empire,^[9] could be thought to be “composed of nothing else”.^[10] There are also a number of practical magical writings that seem to confirm much the same picture, including not just those that constitute the well-known Papyri Graecae Magicae^[11] but such works as the amulet grimoire of Cyranides^[12] or the Testament of Solomon — a handbook for controlling demons potentially responsible for everything from migraine to death.^[13] Early Christian literature, such as the canonical Acts of the Apostles^[14] and the apocryphal Acts of Peter,^[15] depict an empire preoccupied by magic, a world in which those spreading the new faith are forced to battle with magicians^[16] and magical books are burnt in public by those that they convert.^[17]

The material culture of the empire likewise seems to provide copious, tangible evidence of the vitality of belief in magic. Artefacts, such as the myriad of defixiones (binding spells),^[18] incantation bowls,^[19] “voodoo” dolls,^[20] magical lamellae and amulets,^[21] brought together in the extensive collections by the likes of Campbell Bonner, John Gager, Roy Kotansky, Simone Michel, Daniel Ogden, and Hanna Philipp,^[22] appear compelling evidence of magic’s significant place in the lives of most of its inhabitants.^[23] And we could easily go on: from the presence of the paradigmatic witches Circe and Medea on Roman oil lamps, gemstones, murals and sarcophagi,^[24] to the plethora of apotropaic representations of the evil eye found on everything from mosaics and amulets to ear-rings,^[25] the salience of magic in the Roman Empire seems to be anything but illusory. Even epitaphs appear to bear witness to its importance. Here, for example, is one from Rome itself which dates from the 20s CE:

Iucundus, the slave of Livia the wife of Drusus Caesar, son of Gryphus and Vitalis. As I grew towards my fourth year I was seized and killed, when I had the potential to be sweet for my mother and father. I was snatched by a witch’s hand, ever cruel so long as it remains on the earth and does harm with its craft. Parents, guard your children well, lest grief of this magnitude should implant itself on your breast.^[26]

Indeed, the moral and legal prohibitions placed upon magic,^[27] not least the fact that the practice of magic was deemed a capital offence in Roman law,^[28] combined with its prominence in early Christian heresiological literature, where it functioned “as the discourse of alterity par excellence”,^[29] appears to confirm that magic was indeed a dynamic and potent force in early imperial culture. It is, perhaps, so hard to resist the intrinsic allure of an amulet depicting an anguipede, cockerel-headed Abrasax,^[30] or Solomon, on horseback, spearing a demon,^[31] that to conclude otherwise seems unimaginable.^[32] In the face of the data we have just surveyed it could be judged perverse not to agree with Hans Dieter Betz that “Magical beliefs and practices can hardly be overestimated in their importance for the daily lives of the people.”^[33]

However, the picture just drawn at best only indicates the presence of ideas about magic and magical practices of some kind, and we need to determine a defensible definition of magic before we can say even this with any confidence. Gauging the character and prevalence of magic requires a more sustained and rigorous analysis of sources that shed light on the early Roman empire, and one that, importantly, attends not just to the apparent presence of magic but its absence too. We need to note not just where it appears but also, tellingly, where it does not. Before we address these two elements of our analysis, let us begin, however, with the question of the definition of the term “magic”, something that is necessary if what follows is to have any value.

Now you see it, now you don’t: defining magic

Although “magic” at least has the advantage of being a “native category of thought” for those who lived in the Roman empire,^[34] something that is not necessarily the case for the inhabitants of other cultures in the past and the present,^[35] what exactly constituted “magic” for them is far from self-evident. To eschew a definition of “magic”, as some classical scholars do,^[36] is not advisable because it tends to result in the conflation of “magic” with a variety of other things that might strike some modern scholars as manifestly magical but were, in fact, everyday and uncontroversial elements of religious life in the empire and not considered such by any of its inhabitants.^[37] For example, divination, the attempt to determine the will of the gods and the likely outcome of future events, was not in itself something that would be judged magical by those living in the early Roman empire. It was not only ubiquitous^[38] but was a central part of most religions in antiquity,^[39] and especially the religious life of the Romans.^[40] It is not, for example, helpful to label the activities of haruspices, many of whom were key religious officiants in the public cults of the empire, as practitioners of “oracular magic”, as some have done.^[41] Such divination did not constitute magic but a respected and necessary religious act,^[42] something undertaken, for instance, after most public sacrifices.^[43] The same could be said of amulets or, indeed, incantations, the use of neither of which was thought in itself to be magical. For example, every freeborn male, before reaching maturity, wore a bulla, a locket hung around the neck, as an apotropaic device, often containing a representation of a phallus, but none would have considered such a thing magical.^[44] Similarly, incantations were not necessarily magical activities to Romans; their use in the healing of fractures was, for example, recommended by no less a figure than Cato the Elder^[45] and clearly considered by such a respectable authority to be quite distinct from magical practices proscribed under Roman law.^[46]

Failing to provide a definition of magic can also lead many to inadvertently miscategorise some data, to see magic where it was patently not thought to be. For example, invocations of gods other than the Olympic pantheon and closely associated deities have often been seen as “magical” because of a historical tendency within the field to protect a dominant but narrow understanding of classical religion, to fall victim to what has been termed “Classicity”.^[47] So, as Attilio Mastrocinque has demonstrated, the cult of the Askalon Asklepios has often been labelled “magical” out of ignorance of the iconography of a cult which was regarded as a local manifestation of one of the most widely dispersed and supported of all the cults in the empire, second only in significance, perhaps, to the imperial cult itself.^[48]

However, avoiding a definition is perhaps understandable, if not entirely forgivable. Sarah Iles Johnston is surely right to observe that:

Endless theorizing about how magic was or was not different from religion (or anything else) had stalled our progress toward examining and understanding some fascinating ancient material.^[49]

And there is good reason to sympathise with Matthew Dickie’s “dismay combined with a sense of foreboding”^[50] upon encountering yet another attempt to define magic. The literature can be quite overwhelming, not least because within anthropology, the field in which most contemporary thinking on the subject of magic has taken place, magic has been “at its epistemological centre”^[51] since its inception, and continues to generate extensive debate.^[52]

There are well-known strengths and weaknesses to the different kinds of definition of magic that have been proffered,^[53] however we categorise these, whether the definitions could be said, for example, to be essentialist,^[54] functionalist,^[55] locative-relational,^[56] evolutionary,^[57] developmental,^[58] intellectualist,^[59] instrumentalist,^[60] linguistic,^[61] performative,^[62] emotionalist,^[63] existential,^[64] phenomenological,^[65] mythopoetic,^[66] or sensory.^[67] For example, essentialist or substantivist definitions of magic have proved notoriously problematic. “Magic” and “religion” cannot be easily distinguished by differences between them in, for instance, intention, attitude, action, or social and moral evaluation,^[68] nor even, as Jonathan Z. Smith has suggested, scale;^[69] no criterion is effective in making a clear distinction between the two.^[70] Functionalist definitions of magic suffer from the failing common to functionalist definitions more generally: they tend, in practice, to be dependent upon an implicit, substantive definition of something to which a function is ascribed.^[71] They are also often procrustean, indeed many radically so, only capturing one aspect of a phenomenon in their definition, effectively amputating a great deal that is vital, and sacrificing “historical context in favor of taxonomic purity”.^[72] For example, it seems unlikely that magic should be viewed solely as a response to risk, something that is found wherever there is “a hiatus in knowledge or practical control”, as Bronislaw Malinowski maintained.^[73] Such an understanding is impossible to square with ethnographic data^[74] and does not do justice to the range of motivations, emotions and practices most cultures associate with magic. Those who have argued that magic is a locative or relational category, something that, for example, distinguishes between those labelling and those labelled,^[75] to designate a form of deviance against which a dominant discourse defines itself,^[76] have to deal with the problem that such definitions are, at best, once again, only partial. The association of magic with specific subjects, places, practices and practitioners (some of whom may even self-identify as magicians) indicate that there is more to magic than just a way of creating and condemning alterity.^[77] Within some cultures, including those in antiquity, magic clearly has an identifiable, agreed — if contestable — existence; it had a presence that was more tangible than mere rhetoric,^[78] and was not necessarily understood in relation to central, sanctioned and normative forms of religious life and practice.^[79] And considerably more could, of course, be said.

The business of definition has not been helped by the inconsistency of some major contributors within the field. For example, as Ronald Hutton has noted, although Dickie eschews essentialist definitions of magic in his comprehensive and influential Magic and Magicians in the Greco-Roman World, by the final third of his work he regularly uses the term in just such a manner.^[80] It has also not helped that some major theorists, such as Max Weber, though they regularly discussed magic, and had a substantial impact on subsequent definitional debates, never themselves attempted to define it.^[81]

Given the failure of scholarship to arrive at an agreed definition, Alfred Radcliffe-Brown famously suggested that there should be a moratorium on the use of the term “magic”.^[82] However, this is not a way out of the impasse. In practice, it has just resulted in a proliferation of unhelpful circumlocutions, or forced and ungainly synonyms. For example, some scholars of religion in antiquity refer to magic as “ritual power”,^[83] a designation that fails to take seriously non-ritual aspects of the phenomenon they are attempting to study. It precludes, for instance, analysis of the evil eye which could be cast inadvertently without any recourse to ritual.^[84] Where magic can reasonably be argued to be a native category, as is the case in the early Roman empire, such circumlocutions tend to obfuscate and hamper rather than aid analysis.

Rodney Stark is right to observe that, generally speaking, “the term magic has been a conceptual mess”,^[85] and this is especially true amongst those concerned with the study of magic in antiquity. Even though we have near universal belief in its significance, we do not have anything approaching a consensus about what it is or how it should be studied; instead we have “a confusing spectrum of divergent theories”.^[86] Indeed, recent debates amongst those who study magic in the ancient Mediterranean have “trodden what appeared to be a reasonable amount of scholarly common ground into a quagmire”.^[87] However, things are not as intractable as they might appear. A definition of “magic”, for our purposes, need not be one that is ahistorical nor universally applicable. Though such definitions can be useful to “think with”, or said to be sensitising^[88] — that is, they can assist us in scrutinising the phenomenon more carefully by helping us to ask questions about both the subject and our own analysis of it — they can also be misleading and are unnecessary for interpreting imperial culture. All we require is a definition that fits this particular context. It does not need to extend to making some kind of sense of the world of the Azande, Trobriand Islanders or practitioners of contemporary Wicca.

However, deriving a definition that is rooted in first-century conceptualisations of magic is still a challenging task. Perhaps surprisingly, given that it carried a capital penalty,^[89] “the Romans produced no precise definition of what magic was and what was not”.^[90] Indeed, Apuleius raised the matter of definition when defending himself against the charge of witchcraft (an occasion when it was clearly of some consequence), asking a deceptively simple but devastating question of the lawyers representing his accuser: “I should therefore like to ask his most learned advocates how, precisely, they would define a magician?”^[91]

Whatever definition we arrive at will, clearly, have its limitations, particularly given the range of different ethnic and regional cultures encompassed by the empire. Nonetheless, a definition derived from those things which can reasonably be assumed to have been considered magical by most people in the early Roman empire, largely, but not solely, indicated by the presence of a cluster of key Latin and Greek terms related to magical practitioners (Latin: magus, lamia, saga, maleficus, praecantrix, veneficus; Greek: μάγος, γόης, φάρμακος) and the practice of magic itself (Latin: magica, veneficia; Greek: μαγεία, γοητεία, φαρμακαεία), seems reasonable, even if, as the famous trial of Apuleius indicates, the meaning of such terms was both malleable and contestable.^[92] Such a definition could, in Ogden’s taxonomy, be termed “linguistic”.^[93]

However, I would also like to propose a definition that is polythetic,^[94] to borrow a concept from a form of classification employed in biology, but also familiar in the study of religion in general as well as the study of religion in antiquity.^[95] Such a form of definition allows it to reflect the multivalent interpretations of magic in the early Roman empire. That is, the definition that follows is based upon a set of characteristic properties regarded as indicative of magic, many of which need to be present for us to identify its presence in our sources (and then undertake the business of gauging its saliency), though none of which is either sufficient or necessary. It is useful to think of those things that were identified as magic in antiquity as possessing what Ludwig Wittgenstein referred to as a “family resemblance”, something that allows considerable variety whilst also allowing for identifiable commonality.^[96] The definition I would like to use is also one that is dependent, as far as it is possible, upon the emic perspective of inhabitants of the first century,^[97] or better, given disagreements and differences over what exactly merited the label “magic”, as we can see in Apuleius’ trial, emic perspectives of inhabitants of the early Roman empire.^[98]

So, in brief, I believe it is both useful and legitimate to think of magic in the early Roman empire as something associated with characteristic:

(a) Practices. Magic was often thought to involve nocturnal and secret rites,^[99] the use of incantations, spells and voces magicae,^[100] as well as abnormal sacrifices, including the sacrifice of humans.^[101]

(b) Practitioners. Although non-specialists could carry out magical acts,^[102] a range of identifiable experts were associated with the practice of magic, from sorcerers and magicians to witches and root-cutters.^[103]

(c) Places. Particular locations, especially those places connected with the dead and death, such as cemeteries, battlefields or places of execution,^[104] and places that were secret or isolated, such as caves, ruins, or woods,^[105] were regularly associated with magic.

(d) Times. Magic was especially associated with the night,^[106] a full moon^[107] or an eclipse.^[108]

(e) Materials and artefacts. Specific plants and gemstones, as well as animal and human body parts, were thought to be necessary for the practice of magic.^[109] Certain objects, such as amulets, magical books, voodoo dolls, lamellae and defixiones, and knotted threads,^[110] were believed to be tools employed by those utilising it.

(f) Knowledge. Magic was usually thought to involve the possession and application of distinctive, specialist and secret knowledge. This could be of both a technical and propositional kind. In the case of the former, it could include such things as knowledge of specific rituals and practices, and, in the case of the latter, such things as knowledge of supernatural realms and their inhabitants, or the true natures of, and potential causal relationships between, animate and inanimate objects.^[111]

(g) Gods and spirits. Magic was particularly associated with infernal, chthonic gods of the underworld, especially Hecate,^[112] and the spirits of the dead, especially the restless dead, those who had died too early, or too violently or who had not received the appropriate burial rites, or had been killed by magical practitioners themselves.^[113]

(h) Effects. Magic was usually thought to be something that was harmful to at least one of the parties involved.^[114]

There are other traits that regularly appear in depictions of magic that were prominent in the early Roman empire.^[115] Magic was, for example, regularly associated with particular geographical locations, such as Babylonia,^[116] Egypt^[117] or Thessaly,^[118] and cities such as Ephesus^[119] and Memphis,^[120] or ethnic groups, both real and imagined, such as Chaldeans,^[121] Hyperboreans,^[122] Persians,^[123] Egyptians,^[124] Jews,^[125] and the Marsi.^[126] It was also usually deployed in specific agonistic contexts where the practitioner or client often had much to lose or gain, such as trade, law, sport and love.^[127] It was sometimes spoken about in terms of compulsion, with the magician assumed to have the power to be able to compel even a god to act against their will.^[128] However, the key characteristics I have just adumbrated are a useful distillation of the central features of magic in the early Roman empire, at least for most of its inhabitants (there were, of course, variations within some groups, notably Jews, and later Christians, who, in addition to sharing many of these general notions about magic, tended to equate the religious practices of others with magic).^[129]

So, using our definition, perhaps unsurprisingly, the famous depiction of the witch Pamphile in Apulieus’ Metamorphoses could be said to contain (a) Practices, (b) Practitioners, (d) Times, and (e) Materials and Artefacts, that to inhabitants of the early Roman empire were characteristic of magic:

As night began … she arranged her deadly laboratory with its customary apparatus, setting out spices of all sorts, unintelligibly lettered metal plaques, the surviving remains of ill-omened birds, and numerous pieces of mourned and even buried corpses: here noses and fingers, there flesh-covered spikes from crucified bodies, elsewhere the preserved gore of murder victims and mutilated skulls wrenched from the teeth of wild beasts. Then she recited a charm over some pulsating entrails and made offerings with various liquids …. Next she bound and knotted those hairs together in interlocking braids and put them to burn on live coals along with several kinds of incense.^[130]

Similarly, the description of events surrounding the death of Germanicus, as recounted by Tacitus, has (a) Practices, (c) Places, (e) Materials and Artefacts, (g) Gods and spirits, and (h) Effects, associated with magic by most in Graeco-Roman culture:

Explorations in the floor and walls [of the building in which Germanicus died] brought to light the remains of human bodies, spells, curses, leaden tablets engraved with the name Germanicus, charred and blood-smeared ashes, and others of the implements by which it is believed the living soul can be devoted to the powers of the infernal deities.^[131]

However, using our definition, the much-discussed Isis theophany that is central to the climax of Apuleius’ Metamorphoses^[132] and which leads to the protagonist’s return to human form, would not be considered an example of “magic” because it fails to possess any of its possible characteristics (other than it taking place at full moon, a time which, in any case, had specific non-magical associations for worshippers of Isis).^[133] Whilst modern commentators, such as Stavros Frangoulidis, are entitled to label it magical,^[134] depending upon what kind of definition of magic they are employing,^[135] such a designation would have made little sense to its original readers.

Certainly, if we look at the implied definition of magic found in Roman legislation, our operational, polythetic definition appears congruent with what is assumed there. Sulla’s Lex Cornelia de sicariis et veneficiis of 81 BCE, the chief law relating to magic that was in force in the early Roman empire,^[136] contains all of the elements of our definition (with the exception of a clear reference to characteristic (b) Place). Although we do not have the text of the Lex Cornelia de sicariis et veneficiis itself, this can be seen in excerpts from Pseudo-Paulus’ famous commentary on this law:

15. Those who perform, or arrange for the performance of, impious or nocturnal rites, in order to enchant, transfix, or bind someone, shall either be crucified or thrown to the beasts.

16. Those who sacrifice a man or obtain omens from his blood, or pollute a shrine or a temple, shall be thrown to the beasts or, if honestiores,^[137] be punished capitally.

17. It is agreed that those guilty of the magic art be inflicted with the supreme punishment, i.e., be thrown to the beasts or crucified. Actual magicians, however, shall be burned alive.

18. No one is permitted to have in their possession books of the magic art; anyone in whose possession they are found shall have their property confiscated, and the books publicly burnt, and they themselves shall be deported to an island; humiliores shall be punished capitally. Not only is the profession of this art but also the knowledge prohibited.^[138]

Of course, there was another side to magic in the Roman empire to that which we have discussed so far. For some, there was a respectable and venerable form of magic. So Apuleius, for example, initially defended himself against the accusation of sorcery by confirming that he was happy to be called a magus — as long as it was understood that by this he meant someone in the line of the ancient Persian magi,^[139] priests of Zoroaster who were considered especially skilled in such things as oneirology, astrology, and additional forms of divination, including the ability to undertake otherworldly journeys.^[140] And this was clearly distinguishable from the corrupt form that was popularly thought to be “magic”. As Calasiris, an Egyptian priest in Heliodorus’ Aethiopica declared:

Of our wisdom there is one kind that is common and — as I may term it — creeps on the ground, which is concerned with ghosts and occupied about dead bodies, using herbs and addicted to enchantments, neither tending itself nor bringing such as use it to any good end …. The other, my son, which is the true wisdom, from whence the counterfeit has degenerated.^[141]

In the light of such material it might appear useful to speak of a variety of magics co-existing in the Roman empire, as Richard Gordon has suggested.^[142] Indeed, forms of magic developed and changed over the centuries, and it is possible to see the increasing elaboration of practice from relatively simple Greek techniques of the classical period to the involved esoteric forms that are more common in the empire (evidenced in the increasing complexity of curse tablets and the growing popularity of a new genre of physica, works such as that of Cyranides that detail the occult forces of nature).^[143] If we accept Fritz Graf’s analysis, we can see a shift from an essentially instrumental interest in magic to an epistemological fascination with what knowledge it might be able to provide about the supreme God. The latter was especially manifest in the various Hermetica that flourished from the mid-second century CE^[144] and the theurgy of the Iamblichus that became prominent in the third,^[145] though it might also have been present in the possible neo-Pythagorean revival associated with Nigidius Figulus which appeared in the late Republic.^[146] However, whilst it is certainly important to note rarefied discourses of magic, and, indeed, different regional and ethnic traditions and emphases, this should not preclude us from identifying and scrutinising the significance of what most people judged to be magic, of making judgements on the saliency of something that constituted the generally held, shared culture of the empire. Our definition is one that reflects the dominant and most widespread understanding of magic in the early Roman empire, the kind that Calasiris calls “common”; a kind of magic identified by most commentators as taking a surprisingly similar form across the empire by at least the second century CE,^[147] though present in most forms of Graeco-Roman culture sometime before that.

Just an illusion? Evaluating evidence for the presence of magic

Before we evaluate the evidence for magic in the early Roman empire, we need to begin by abandoning the fundamental assumption of many working in the field, or dependent upon work in this field, that magic must, of necessity, have been significant because the Roman empire was a pre-modern culture. In approaching the empire and its inhabitants we need to do something analogous to that which Mary Douglas, some decades ago, advocated anthropologists should do, and “ditch the myth of the pious primitive”.^[148] We need to be aware that the salience of magic needs to be proven rather than assumed, however much some may have invested in the subject. Magic was not necessarily a constant or significant feature of all pre-modern societies, and we should not presume that it must have been for the inhabitants of the early Roman empire.^[149]

Indeed, when we look at the evidence rather more closely, some perplexing things emerge and reasons for believing that magic was a significant element of early imperial culture and the day-to-day lives of its inhabitants appear less compelling. For example, the interest in magic in literary sources is a far from unproblematic indication of its saliency. Despite its centrality, Homer’s representation of magic is actually somewhat ambivalent and cannot be presumed to have contributed to its alleged importance in the empire. As well as providing the paradigmatic literary representations of magic, in the depiction of figures such as Circe^[150] and Calypso,^[151] Homer was also capable of demonstrating a sustained disinterest in it,^[152] something that did not escape the attention of his readers: whilst the Odyssey is replete with references to magic,^[153] there is, as Pliny the Elder observed, no mention of it at all in the Iliad.^[154] The engaging depictions of magic by the likes of Apuleius, Lucian and Petronius,^[155] with their accounts of such things as haunted houses and human sacrifice, are heavily stylised and formulaic and, as Graham Anderson has argued, reminiscent of folktales or better fairytales that predate these texts.^[156] Such works tell us that stories about magic were considered entertaining and had an audience, but little else. Magic might “matter” but not in the sense that is usually assumed: the inhabitants of the early Roman empire could well be like the Dani of Papua New Guinea who show “more fear of ghosts in stories than they do in their everyday activities.”^[157]

The reservations we have about the value of literary works as evidence for the widespread significance of magic in the early Roman empire should also extend to legal sources too. The existence of laws aimed specifically against magical practices and practitioners, such as the Lex Cornelia de sicariis et veneficiis, do not, in themselves, tell us much about the saliency of magic in the empire. Such laws do not necessarily reflect the sustained assumptions and anxieties of the wider cultures within which they operate. Indeed, laws against magic are often the residue of short-lived moral panics.^[158] In this sense, laws like the Lex Cornelia de sicariis et veneficiis (and the earlier laws from which it was constituted),^[159] may well be similar to such things as the Garrotter’s Act of 1863, which remained on the statute books in England and Wales for almost a century, and was a legal response to the sudden appearance of foreign stranglers who, albeit briefly, gripped the imagination though not the throats, of Victorian Londoners.^[160]

Indeed, the limited number of prosecutions for witchcraft in the early Roman empire supports such an interpretation of the nature of such legislation and, in itself, is indicative of a general lack of interest in magic. Few people were tried and even less executed for magic in the empire (nor is there evidence of the extra-judicial or de facto killing of magical practitioners). Relative to the population of the empire as a whole, the numbers put to death appear to have been extremely small, and when judged against practices in other cultures, strikingly so. For example, although the data is not entirely unproblematic and cross-cultural comparisons can be invidious, the number of witches executed in only two years in the English region of East Anglia between 1645 and 1647 appears to be roughly comparable to the total number executed in the first few centuries of the Roman empire^[161] — and the former had a population of less than one per cent of the latter.^[162]

There is also a famous paradox, well known in antiquity, evident in the actions of those that did bring prosecutions against magical practitioners, which makes it difficult to believe that they really credited magic with the kind of power that is often assumed: as Apollonius of Tyana allegedly remarked, “If you think me a sorcerer, how will you chain me? And if you chain me, how will you think me a sorcerer?”^[163] Indeed, not only would it be impossible to punish someone who had such power but, as Apuleius pointed out in his own defence, it would also be suicidal: “the man who believes in the truth of such a charge as this is certainly the last person in the world who should bring such an accusation.”^[164]

The material culture associated with magic which can be dated to the early Roman empire is also a far from reliable indicator of the ubiquity of assumptions about its efficacy even though it is tempting to interpret such evidence in this way.^[165] Of course, many artefacts associated with magic are, by their nature, ephemeral and unlikely to leave much of an impression on the archaeological record — one thinks, for example, of the magical threads that were used as charms or to effect binding spells^[166]— but magical artefacts, or references to them, are surprisingly thin on the ground. For example, no objects that Romans would have considered unequivocally magical were discovered at Pompeii or Herculaneum,^[167] and references to magic do not appear, even obliquely, in the abundant graffiti from these sites, material that allows “an attempt to define a popular culture of the time”.^[168] As Andrew Wilburn has observed in his study of the archaeology of magic in Roman Egypt, Cyprus and Spain (a study predicated upon a much more expansive definition of magic than the one employed in this paper):^[169]

The preserved evidence of enacted magic such as curse tablets is comparatively small when juxtaposed with other corpora of textual artifacts such as public inscriptions and ostraca. The number of published curse tablets stands at approximately 1,600 which derive from over a period of approximately one thousand years and the full geographic extent of the Roman Empire. In contrast, over one thousand ostraca have been published from the University of Michigan excavations at the site of Karnis alone.^[170]

Even when we do discover objects that can, with reasonable certainty, be categorised as magical, what we can deduce from them about the significance of magic in the early Roman empire is far from self-evident. Although it is common to see such things as having “attendant beliefs and assumptions”^[171] what exactly these might be is not easily discerned. What can we say about the “attendant beliefs and assumptions” possessed by an amulet that was claimed to render the wearer invisible?^[172] Did those manufacturing and using such an object really think that it worked? Did they imagine it was as efficacious as, say, those amulets that were declared, rather more modestly, to relieve indigestion or alleviate a hangover? Or to make the wearer more popular or lucky? (all claims that allowed a rather more subjective assessment of their veracity).^[173] What can we say about the kind of beliefs that “attended” to the defixio found in Hadrumetum (Sousse) in which a man sought to make four women fall in love with him? Does the large number of potential lovers tell us merely about the ambition of the man or does it tell us that he did not hold out much hope of the likely efficacy of such a practice in relation to any of the women named?^[174] And what of a bracelet made up of over forty different “charms” found at Herculaneum?^[175] Should it be considered evidence of the significance of magic in the life of the wearer? Or was it primarily decorative, sentimental, or even a form of mnemonic device, providing a means of exercising control over the universe, in a limited but effective way, though not through the supernatural power of magic but through the process of collecting to which it bears witness^[176] and the autobiographical structuring of memory such an activity can facilitate?^[177] Of course, none of these alternatives need be the sole “meaning” of the charm bracelet for the wearer or others creating or encountering it, and none need preclude the possibility that magic was, indeed, a constituent part of its variegated “attendant beliefs” but they do alert us to the possibility that magic might, at best, be just one, perhaps inconsequential element in the meaning ascribed to an object, even an object that some might assume must be understood in such a way.

Indeed, we should be careful not to mistake the presence of an object with the simple presence of particular ideas, magical or otherwise. Although artefacts may have the capacity to “symbolise the deepest human anxieties and aspirations”,^[178] such as those associated with the agonistic obsessions of love, sport, law and business, that are, for example, the stuff of ancient magic, and such objects might relay “a cultural image of the way in which the universe works”,^[179] they also have “social lives”^[180] and “biographies”,^[181] determined locatively and temporally, and we should not overlook what Woodward calls the “idiosyncrasies, incoherencies and sheer mundanity of the user’s perspective.”^[182] We know, for example, that some who wore amulets (which were, as we have noted, not necessarily understood to be magical), had little interest in their supposed effects,^[183] and others recommended their use for psychological benefits but completely disavowed any “worldview” implicit in their manufacture.^[184]

Even tombstones do not provide us with evidence of the saliency of magic in the everyday lives of inhabitants of the Roman empire that is quite as solid as it might at first appear. We have tens of thousands of epitaphs, often recounting the manner in which the person commemorated met their death, but the epitaph mentioned at the outset of this paper is one of only a handful that speak of someone being killed by witchcraft.^[185]

In short, the data often taken as evidence for the cultural significance of magic in the early Roman empire, even when examined on its own, in isolation from the wider social context to which we shall now turn, is not as unequivocal or necessarily as substantive as is often assumed.^[186]

An empty box? The absence of magic

Although it is generally correct to say that absence of evidence is not evidence of absence, if magic were something of significance in the early Roman empire, we would expect to find evidence of its presence in sources that shed light on the day-to-day lives of its inhabitants,^[187] that is, to find evidence of it in those texts and artefacts which, however imperfectly, could be said to be indicative of popular culture. If we leave aside those sources that are directly concerned with magic, such as the Papyri Graecae Magicae or Apuleius’ Apologia — compelling though they may be, and particularly so when grouped together in collections dedicated solely to the subject of magic in antiquity^[188] — and instead look at those sources that reveal the general preoccupations of the time, we discover a near silence concerning all things magical. The lack of interest is striking and unequivocal. Witches, sorcerers, and spells warrant virtually no mention or none at all in, for example, the popular ethical literature common in the early Roman empire, the collections of proverbs, fables, gnomai, and exempla.^[189] The only appearance of a magical practitioner in Aesop’s Fabulae, for example, a body of literature that was culturally omnipresent and popular across all strata of Graeco-Roman culture,^[190] is one in which the powers of a witch are ridiculed (with no untoward effects):

One of the spectators, seeing her [a witch] being dragged out of the court said to her: “How is it that you claim to be able to avert the gods’ anger, that you were not even able to persuade human beings?”^[191]

Nor is magic of consequence in the Vita Aesopi either, the comic biography of the fabulist, composed around the second century CE.^[192] The collection of exempla by Valerius Maximus, from the reign of Tiberius, and a useful window into common assumptions and obsessions, likewise, contains no clear reference to magic.^[193]

Magic and magicians also play little part in popular paradoxographical literature of the period, such as Phlegon of Tralles’ de Mirabilis, texts that seem to have had a wide readership across social and cultural groups in the Principate.^[194] Nor do they feature in Artemidorus’ Oneirocritica, a handbook of dream interpretations that provides an extremely valuable repository of the anxieties of the time and which has been likened to an ethnography of the second-century Mediterranean world.^[195] Whilst the Oneirocritica indicates that those who lived in the early Roman empire were fearful of such things as disease^[196] and poverty,^[197] and dreamed of a host of subjects, from having sex with their mother,^[198] to being crucified,^[199] or getting dressed the wrong way in the morning,^[200] they did not dream of magicians or spells. Nor, from the range of interpretations given, was magic one of the things that they believed that their dreams were really about.^[201] Magic is also not a subject that appears in Roman joke books, such as the Philogelos, again a useful source for identifying the general preoccupations of the time and which, instead, finds humour in such perennial topics as sickness, sex and intellectuals’ lack of common sense.^[202] And it is not a concern of the popular do-it-yourself oracle books such as the Lots of Astrampsychos.^[203] Although this text does have an exotic quality to it — it took its name from a mythical Zoroastrian priest^[204] — when we scrutinise the wide variety of questions that could be asked of the oracle (of which there were 92), and the answers given (of which there were 1030), it is clear that magic was of no consequence.^[205] Other things preoccupy the text and, one assumes, those using it, such as employment, health, love, fertility, travel, business and death. Nor is magic amongst the causes of fortune and misfortune assumed. Similarly, the popular Homeromanteion, an oracle which consisted of 216 lines of Homer that provided possible answers to whatever questions were put to it, makes no direct reference to magic or witchcraft even though many of the excerpts from Homer were taken from the Odyssey, a text which, as we have noted, has a considerable interest in magical themes.^[206] Such material appears to indicate that most people were unconcerned by magic, most of the time. They clearly did not think it had explanatory power in making sense of their lives or obtaining their goals. Nor was it something perceived to be a threat. Nor did they ascribe to it any symbolic significance. They were, it appears, at best, indifferent to it. From these popular cultural texts it is fair to conclude that it had little saliency in the early Roman empire.

In the light of the preceding discussion, it is apparent that Betz’s assertion that, “Magical beliefs and practices can hardly be overestimated in their importance for the daily lives of the people”,^[207] is untenable. It is clear that the significance of magic in the lives of those in the early Roman empire can, in fact, all too easily be overestimated and, indeed, regularly is. To put it crudely, and I am aware the distinction has its limitations, for most of the inhabitants of the Roman empire, for most of the time, magic appears to have been largely the stuff of stories and not of life.

Explaining indifference

It is not necessary to explain why inhabitants of the early Roman empire had such limited interest in magic in order for our conclusions about its lack of saliency to stand. Nonetheless, given that it is often, even if erroneously, assumed that magic was a significant preoccupation of pre-modern cultures, this unusual finding does invite further comment and I would like to posit some tentative, partial, explanations for this phenomenon. Indifference is under theorised in the study of religion in antiquity (although it is of increasing interest for the study of contemporary religion)^[208] nonetheless I would like to suggest three possible reasons for the absence of interest in magic in the lives of most inhabitants of the empire, most of the time. I believe that it is probably, in part, a consequence of the existence of widespread scepticism of two kinds, which whilst related are not synonymous: (a) scepticism concerning the supernatural and (b) scepticism concerning magic.^[209] In addition, it is also likely to be (c) a function, on those limited occasions when it was indeed used, of the agonistic contexts within which magic was deployed in the early Roman empire, something that we shall return to at the conclusion of this essay.

It is important to emphasise that the term “scepticism” is used here both in the modern, popular sense of active disbelief, as well as the related sense of the necessary suspension of judgement where a valid conclusion is impossible, for example, about the causation of a phenomenon. I am not using it with Pyrrhonism and formal philosophical Scepticism in mind.^[210] It is also important to emphasise that scepticism about magic does not necessarily imply scepticism about the power of the gods, although the reverse is not the case.^[211]

However, the use of the concept “scepticism” requires some defence. It could be said to be misguided, to be both unhelpfully polarising^[212] and to approach the subject with unwarranted, anachronistic, presuppositions about the necessary significance of “belief” in the study of religion generally, and the religions of antiquity more specifically.^[213] As Ken Dowden quite rightly says:

One of the hardest features of ancient religion for the modern student is the sheer unimportance of belief. […] The ancient religions are not dead faiths, they are obsolete practices.^[214]

It could also be said to be an idea that does not do justice to the mutually contradictory ways of talking about the gods that were common and allowable in the empire that resulted from “the different kinds of assent and criteria of judgement”^[215] applied in different contexts; an approach to religion characterised by what Veyne calls “mental Balkanization”.^[216] Such a view is most clearly evident in the three very different theologiae of poetry, politics and philosophy identified by Varro.^[217]

Nonetheless, whilst it is true that public, elective, and domestic cults of the empire did not have any place for instrumental or soteriological conceptualisations of belief,^[218] and nor did magic, religion and magic in the empire were both predicated on certain assumptions, such as the efficacy of ritual and the power of the gods that underpinned their workings.^[219] Such “beliefs” (or, perhaps better, “ideas” or “convictions”) were not the kind that required active assent — they were not beliefs “in” but rather beliefs “that”^[220] — they were not of a soteriological but of an epistemological kind.

However, even beliefs of this sort can be the subject of dissent (rituals, for example, can be left undone) and so it is not unreasonable to speculate on the role of scepticism in making sense of the lack of interest in magic in the empire. And whilst it is true that most of those in the empire operated with a number of different, apparently mutually contradictory, theologiae of the kind identified by Varro, this does not preclude us from talking about scepticism, although it does require us to be sensitive to the situational articulation of such beliefs so that we do not misread the evidence.

Sceptical attitudes towards the supernatural

There is evidence of a significant degree of scepticism concerning the supernatural in the early Roman empire, particularly in relation to the possibility of direct intervention by the gods or other supernatural powers in human life (something that is not necessarily the same as scepticism about the existence of the gods per se). Such an argument is not dependent upon the number of those who identified themselves with philosophical schools that were hostile to supernaturalism, such as the Epicureans, Cynics, and Sceptics, something that, relative to the population as a whole, is unlikely to have been large.^[221] We should not overlook the attempts by members of these movements to disseminate key doctrines beyond their core adherents, seen, for example, in the remarkable inscription at Oenoanda in Lycia which gave passers-by access to an extensive collection of Epicurean treatises,^[222] or the notorious behaviour of Cynics that was intended, in part, to both embody and communicate their ideas to a wide audience,^[223] but their success appears to have been limited.^[224]

Rather scepticism towards the supernatural went beyond such circles and was not necessarily associated with strong philosophical commitments or philosophical identities of any particular kind. This is evident, for example, in historiographical and medical discourses prominent in the early Roman empire in which the supernatural was not a causative agent in the lives of humans. Some historians of the period excoriated those that believed it was,^[225] whilst most seem to have been studiously “ambivalent”^[226] about the direct intervention of the gods in human history, and it is common to find naturalistic explanations for allegedly supernatural events,^[227] even if many were not always consistent in their approach.^[228] Naturalistic explanations of disease were also dominant in professional medical discourses of the empire that were indebted, directly or indirectly, to the Hippocratic tradition that effectively demythologised supernatural aetiologies.^[229] Of course, such rational approaches to disease and healing should not be crudely contrasted with those that allowed room for intervention from the gods (even the physician Galen could believe that the god Asklepios had saved him from the plague and that he was only a doctor because the god had appeared in dreams to his father),^[230] nor should we assume that they were dominant in popular culture^[231] but they were well known^[232] and contributed to the normalisation of discourse in imperial culture which was sceptical of the supernatural.^[233]

Although no one in the Roman empire achieved the notoriety of the infamous “atheist” Diagoras of Melos of the fifth century BCE who not only mocked the Eleusinian mysteries but, after his prayer for the return of a lost manuscript went unanswered, boiled up some turnips on a fire kindled with a wooden statue of Heracles,^[234] it is also the case that there were some who, at least on occasion, showed a comparable lack of concern for the supernatural power of the gods. The general Claudius Pulcher, for example, famously drowned the sacred chickens who refused to eat when offered grain, and so failed to provide a positive omen for his forthcoming (and unsuccessful) campaign, quipping “If they will not eat, let them drink.”^[235] And he was hardly alone.^[236] According to Suetonius, Roman crowds, grief-stricken at the death Germanicus despite their prayers, “stoned the temples, and toppled the divine altars, while others flung their household gods into the street”, in part, no doubt, in an attempt to punish the gods but also an indication that the gods were judged to be powerless.^[237]

It was not unusual to doubt whether gods were capable of intervening in human affairs,^[238] and such a position was not limited to moments of collective crisis or disappointment.^[239] We find plenty of examples of popular, everyday scepticism in the period. So for example, one of Babrius’ Aesopic fables reads: “Since the gods do not know who steals from their own temples what is the use of appealing to them for help in finding any other lost property?”^[240] In the Enchiridion, Epictetus is reported as observing that those who did not obtain what they expected in life were prone to abuse the gods and accuse them of being uninterested in human affairs, something that was particularly true of farmers, sailors, merchants and those who had been bereaved.^[241]

On occasion, the gods, both new and old, could be the subjects of ruthless satire^[242] and irreverent behaviour: their festivals^[243] and oracles^[244] mocked, their sacred groves cut down,^[245] sacrifices stolen,^[246] and cult images abused.^[247] People could even dress up as gods for fancy dress parties^[248] and make the condemned parade as gods for sport before their execution.^[249] It is, perhaps, no surprise that there was such widespread concern in the empire about the danger of impietas (“denying the gods the honours and rank that were rightfully theirs”)^[250] and, in particular, impietas that was deliberate, with malicious intent, rather than accidental (prudens dolo malo rather than imprudens), something that was inexpiable.^[251] Clearly, there were at least some in the empire more than willing to behave in a manner that showed no fear of supernatural retribution, to the concern of their contemporaries.^[252]

In addition to scepticism of the supernatural evident in the behaviour of some towards the gods, there are also indications that other supernatural powers could be approached with significant scepticism. Epitaphs, for example, could mock the existence of ghosts^[253] and interest in demons could be ranked alongside interest in quail fighting, as a frivolous waste of time.^[254] Even those traditionally believed, at least amongst the elite, Roman males who dominate our literary sources, to be particularly receptive to such beliefs, had, according to Cicero, albeit writing from the context of the late Republic, become more rational:

Who now credits that the hippocentaur or the Chimaera ever existed? Is there a single old woman to be found who is so unhinged as the be sorely afraid of those monsters in the nether world in which people once believed? Time obliterates the falsehoods of common belief.^[255]

The existence of scepticism towards the supernatural in the early Roman empire, whether of an intellectual or apparently more visceral kind, is certainly not key, nor even, necessarily significant, in explaining the lack of saliency of magic, but it undoubtedly had a part to play in this phenomenon.

Sceptical attitudes towards magic

There is considerable evidence that magic in the early Roman empire was regularly denounced as fraudulent. As Gordon has effectively demonstrated, important representations of magic in antiquity “conceived of it not as powerful for harm but, on the contrary as vacant show, as empty nonsense.”^[256] And such scepticism was not just an elite perspective: “Although this view is associated generally with the educated elite, it was also a view widespread in the population at large: for most of the time, under most circumstances, many people considered … it absurd”;^[257] something that played on people’s foolish and extravagant hopes. Those writers, such as Petronius, that made extensive use of magic in their narratives did so “to enthral and entertain in their own right, but at the same time they serve to convey the gullibility and feeble-mindedness of their tellers”.^[258] And they were not alone. The hostility towards magical practitioners evident in the Aesopic fable to which we earlier referred^[259] is a sentiment that recurs elsewhere.^[260] The failure of magic to achieve results was infamous. The inefficacy of love magic, for example, is a recurring topos in literature.^[261] In Ovid’s Heroides even Medea has to admit she cannot be successful at this.^[262] The idea that magicians and witches were frauds who prayed on the vulnerable is a recurring motif in a range of texts.^[263] It can be seen, for example, in Tacitus’ account of the story of the young Servilia, tried before the Senate for using magicians to determine the future fate of her family after it had fallen foul of Nero, and forced to commit suicide as a consequence.^[264]

Stinging criticisms of magical claims can be found in medical writing too. Galen mounted a savage attack on Pamphilius, composer of a treatise on herbs which included extensive discussion of their magical properties, denouncing it as “long-winded Egyptian sorcery” so incredible that not even a child could believe it.^[265] And for the encyclopaedist Pliny the Elder the fact that Nero had reportedly sought to become a magician but, despite all the means he had at his disposal, had failed, was evidence that magic was fraudulent, “ineffectual, vain”.^[266] Cynic criticisms of the claims of magicians were also common. According to Lucian, Demonax confronted a magician who claimed to be able to obtain whatever he wanted by means of incantations, and offered to go to the nearest baker and turn a coin into a loaf of bread.^[267]

For some critics, magic was no more than trickery. For example, Plutarch mentions a witch using her knowledge of the occurrence of an eclipse to achieve the so-called Thessalian trick:^[268]

Aglaonice, a Thessalonian woman — though being thoroughly acquainted with the periods of the full moon, when it is subject to eclipse, and knowing, beforehand the time what the moon was due to be overtaken by the earth’s shadow, imposed upon the (other) women, and made them all believe that she was drawing down the moon.^[269]

Indeed, a number of authors appear to have written works containing rational, reductionist explanations of the secrets of magic. These evidently circulated widely in the empire as Philostratus can mention in passing that several individuals “who have laughed out loud at the art”, had written books on how its effects were manufactured, and seems to assume that these would be familiar to his readers.^[270] Such rationalisations were of various kinds. Some seem surprisingly modern whilst others are wedded to specific ideas about causation that might seem implausible to us.^[271] The plausibility of such rationalisations to us is, of course, of no consequence — the issue is the plausibility of such rationalisations for those who lived in the early Roman empire.

Although no texts of the kind alluded to by Philostratus have come down to us, Hippolytus’ Refutatio omnium haeresium^[272] does include a substantial section that appears to be dependent upon a source of this kind, and gives us our most extensive exposé of the fraudulent techniques of magicians. In this we hear, for example, that magicians demonstrated their powers — such as drawing down the moon and reading sealed letters — in mostly darkened rooms, a context conducive to deception, and used such staples of modern stage magic as misdirection, prestidigitation, and ingenious stage props.^[273] A skull could be made to speak, for example, by the surreptitious use of the long windpipe of a crane;^[274] the clever deployment of rocks, planks and sheets of brass, could create the illusion that the magician is able to summon up thunder.^[275]

Such books may well have made public the secrets of a particular genre known as Paignia or “trifles”, of which our most extensive, surviving fragment, ascribed to Democritus, can be found, somewhat tellingly, in the Papyri Graecae Magicae.^[276] These works seem to have given specific recipes to create dramatic effects, akin to childhood chemistry experiments, some of which were designed to liven up dinner parties^[277] but others of which, such as those found in the Paignia of Salpe, or the collection by Anaxilaus of Larissa, were evidently intended to be employed in other contexts,^[278] and “could be used to impress the gullible with the superhuman powers of the magician”.^[279]

Some provided rational explanations of the apparent effects of magic of a somewhat different kind. Rather than expose the techniques of its practitioners, they attacked the non-falsifiable nature of its claims. Such criticisms had a long pedigree. The author of the Hippocratic work, De morbo sacro, for example, said of magicians that if a patient recovered, they would claim the credit but if they died, they would “have a sure fund of excuses, with the defence that they are not at all to blame, but the gods.”^[280]

And a similar argument is made by Philostratus who provides a surprisingly modern-sounding explanation for the apparent success of magic: to those committed to its use it can never fail, the believer will always provide technical or other excuses to justify whatever outcome occurs; an observation strikingly reminiscent of Malinowski.^[281]

The vulnerability of magic to rational criticism in the early Roman empire is perhaps no better seen than, somewhat paradoxically, in the defence used by some of those tried for practising it. As Pliny recounts, a farmer accused of achieving outstanding yields by magical means defended himself by explaining that toil, not magic, led to his abundant harvests.^[282] Scepticism about magic was clearly vibrant in the empire and may also have contributed to its lack of cultural saliency.

The deployment of magic

The lack of significance of magic in the day-to-day lives of inhabitants of the early Roman empire was probably not only a consequence of scepticism about the supernatural and scepticism about magic itself. It may also have been, in part, a consequence of the context of its deployment, on the limited occasions when some made use of it, something that, as we noted earlier, appears to have been primarily agonistic. There is good reason for thinking that such agonistic use accompanied conceptualisations of magic in which it would be understood to be insubstantial; something ephemeral, equivocal and transitory.

The approach taken by Galina Lindquist,^[283] is particularly useful for identifying the nature of such magic. Magic accessed in contexts characterised by deep uncertainty and lack of control,^[284] according to Lindquist, is a form of materialised “hope” conjured up by frustrated agency, “where the uncertainty of life calls for methods of existential reassurance and control that rational and technical means cannot offer.”^[285] However, the use of magic is not just an attempt to stack the odds in one’s favour through supernatural assistance but has other, more substantive effects. For example, Lindquist usefully suggests that it can redefine a situation, taking away responsibility and accountability for misfortune by transforming “risk” (something dependent upon the decision of an individual) into “danger” (something that can be attributed to the environment).^[286] As she puts it, “When one risks and loses, one has only oneself to blame. In danger, if one is struck and hit, one is an unwitting victim, unfortunate but not guilty.”^[287] There is a temporal and contingent dimension to belief of this kind, and it is not useful to think only in terms of what someone “believes” when a curse is written or spell cast but also about the subsequent form this takes (as Jean-Claude Schmitt has rightly said, “a belief is never a completed activity”).^[288] Once the challenge has passed, Lindquist found that the need for magic or even the recognition of its efficacy often diminishes or vanishes.^[289] Clients create post hoc rationalisations of events, similar to Arthur Kleinman’s “explanatory models” familiar from medical anthropology and which reflect the plural, indeterminate, and mutable character of potential interpretations over time.^[290] Although we lack first-hand accounts to confirm this reading for the early Roman empire, I would suggest that narrations of magic in this period, for most of the limited numbers that seem to have accessed it, would have taken a similar shape to that found in the lives of Lindquist’s contemporary informants: it would acquire a degree of potential saliency at time of need but rather less or none in retrospect as the individual returns to a society in which magic, when it was thought about at all, was viewed as an unsanctioned and problematic activity — whether because it was something shocking and subversive or something embarrassing and risible.

Conclusion

There is a great deal more that can be said about the nature and place of magic in the early Roman empire. It would, for example, be useful to explain why magic did have considerable and unusual saliency for the early Christians, and the factors that led them to conjure up a useful, oppositional illusion of an enchanted and enslaved world.^[291] The alleged significance of magic in the empire is not solely a matter of smoke and mirrors, but by arriving at their estimations of its importance by focusing solely on evidence of its presence, of being too quick to fall under the spell of texts such as the Papyri Graecae Magicae, scholars in the field could be said to have unwittingly been guilty of the classic magician’s trick of misdirection, and have themselves missed perhaps the feature of magic in the early Roman empire that is its most surprising: its lack of significance in the day-to-day lives of its inhabitants. Whilst they clearly enjoyed stories about magic, magic itself seems to have been largely inconsequential and ephemeral, of only fleeting importance, and the subject of the most attenuated and sporadic interest except amongst a handful. We have made some suggestions as to why this might be so, but the necessary process of revision and re-description has only just begun. Despite the plethora of publications in the field, substantial work, some of the most fundamental kind, clearly remains to be done.