I

The historiography of witchcraft is almost as turbulent and varied as its subject, and as revealing about modern times as it is about the period of the witch-trials. The flow of publications never abates and the debates refuse to settle, as each generation has its say. From the early eighteenth century, rationalists saw the middle ages as a time of barbarous superstition typified by witch-hunting, an interpretative bass-line upon which a range of melodic variations was played, from indictments of popish persecution to veiled meditations on twentieth-century genocide. The anti-clericalism of the first history of witchcraft – by law professor Christian Thomasius in 1701 – was revived in the 1840s by another German academic, Wilhelm Gottlieb Soldan, who argued that witchcraft was a non-existent crime invented by monks and inquisitors.Footnote ¹ This idea was most eloquently expressed in the early twentieth century by Joseph Hansen, an archivist in Cologne, whose work inspired American scholars Henry Charles Lea and George Lincoln Burr, both of whom took a dim view of religious politics. In Europe the so-called ‘witch-craze’, with its overtones of madness and zealotry, supplied polemical ammunition for emergent nation-states in their Kulturkampf with the Catholic church.Footnote ²

Alongside ‘liberal-rationalism’ meandered another train of thought: romanticism. Academic neglect of witchcraft from below had left a space into which some fantastic possibilities might be projected. In nineteenth-century Germany, mythologist Jacob Grimm put nationalism before rationalism, asserting that witches were wholesomely völkish wise-women persecuted by the church. His contemporary, French historian Jules Michelet followed a similar line, dressing witches up as anti-feudal rebels and progenitors of the Revolution.Footnote ³ Witches had become real again. In Britain this concept was elaborated by Sir Walter Scott, Harrison Ainsworth, and other novelists whose gothic flights of fancy profoundly influenced Victorian readers.Footnote ⁴ By 1900 a generation of nostalgic occultists, theosophists, and spiritualists had adopted romantic witches as their forebears, providing the fin-de-siècle counter-culture with its creation myth. The First World War dented faith in progress and boosted alternative religions, including the domestic cult of mediumship. In the early 1920s Sir James Frazer and Margaret Murray, avid folklorists, produced books which contributed to the idea that magical beliefs belonged to an ancient and venerable tradition; ‘witchcraft’, it seemed, had once been the authentic religion of the people.Footnote ⁵ The demolition of Murray in the 1970s, notably by Norman Cohn, did little to dampen the ardour of neo-pagan Wiccans and ‘radical feminist’ academics whose voices had grown louder in the preceding two decades; both movements were fixated by Murray's fantasies and execrated ‘the burning times’ of patriarchal, clerical tyranny.Footnote ⁶

For almost two centuries after the last execution of a witch (in the Swiss canton of Glarus in 1782) it was hard to know what had actually happened; the past had served the present better than present had served past. But things were about to change. Pagans and feminists were products of the age of the social sciences, disciplines that would help historians to challenge their tendentious views. In Britain the new direction was marked by Alan Macfarlane and Keith Thomas, who in the early seventies applied social anthropology to archival findings to transform understanding of English witchcraft; and in the United States by Erik Midelfort and William Monter who, in 1968 and 1972 respectively, published optimistic research forecasts followed up with Eurocentric monographs which, like the work of Thomas and Macfarlane, connected the intellectual dimension to the social and cultural.Footnote ⁷ At last the subject had found historians willing to embrace the ‘elementary village credulities’ shunned by Hugh Trevor-Roper in 1967.Footnote ⁸

And yet, as early modern demonologists had found, winning converts is not the same as creating consensus; renewed controversy is a more likely outcome. Critics of Macfarlane and Thomas objected that their pre-war functionalist anthropology seemed rather stale, especially in the post-colonial era when witchcraft and witch-hunting were acquiring new significance. By the 1980s the marriage between anthropology and history had begun to look like a brief encounter.Footnote ⁹ Another drawback of the Macfarlane–Thomas thesis was that however valid it may have been for England, it had limited applicability across the Channel, or for that matter the Atlantic or North Sea. In any case the juxtaposition of an ‘English’ model (concentrating on the accusatory prosecution of maleficium) with a ‘continental’ version (characterized by the demonological obsessions of inquisitors) did neither place much good.Footnote ¹⁰

These doubts were confirmed by empirical research in new academic centres extending north from western Europe into Scandinavia and Scotland, and east into Austria and Hungary. A symposium in Stockholm in 1984 produced a ground-breaking collection of essays edited by Bengt Ankarloo and Gustav Henningsen, published in 1990. Suddenly historians glimpsed the bewildering diversity of regional and local histories of witchcraft, peculiarities of which threatened to overwhelm the affinities on which the old models relied.Footnote ¹¹ At a conference at Exeter in 1991 the Stockholm essays felt like the livre de circonstance, setting the agenda for discussion from the outset. Robin Briggs's paper on the ‘many reasons why’ of witch-trials described the explosion of the former uniformity and holistic explicability of the subject. Stereotypes were questioned. I delivered my first conference paper, quaking in the presence of Keith Thomas as I reported archival findings at odds with his model. Thomas chaired a concluding session billed as ‘Where do we go now? Witchcraft studies in the 90s’, at which he puckishly announced that wherever we might be going he would not be coming with us.Footnote ¹² The way seemed clear. Within five years Jim Sharpe and Robin Briggs had produced major interpretative surveys of English and European witchcraft respectively, surpassing not just the rationalist or romantic models but the social science-based ones as well.Footnote ¹³

There were also signs at Exeter that history, tired of holding a flame for anthropology (which barely featured in the 1990 essays), had found new love with literary theory. Stuart Clark had offered a rousing ‘critique of positivism’, pointing to the ‘reality problem’ of witchcraft. In Religion and the decline of magic Keith Thomas had set out to show that witchcraft was not a delusional aberration but ‘a culturally acceptable form of reality’; now this sort of reality seemed tainted by twentieth-century theorization.Footnote ¹⁴ In Clark's opinion, contemporary language needed to be taken on its own terms, although it would not yield its meaning easily – an idea that resonated in the research of Lyndal Roper, Deborah Willis, Diane Purkiss, and Marion Gibson in the second half of the decade. The value of functionalism seemed passé now that matters of identity, subjectivity, and fantasy were demanding attention.Footnote ¹⁵ In 1997 Stuart Clark published a brilliant study of the idea of witchcraft which restored demonology's lost rationale, and the following year gathered scholars at a conference in Swansea, the papers at which became a volume of essays: Languages of witchcraft.Footnote ¹⁶ Another summit was convened at York in 2002, where a new cohort of young researchers (and some veterans) demonstrated the on-going vitality of witchcraft studies, and the ceaseless retreat of the conceptual and geographical boundaries that had enclosed the subject from Michelet to Macfarlane.Footnote ¹⁷

II

In a conclusion to the Ankarloo and Henningsen volume, Peter Burke observed that in the space of twenty years witchcraft had ‘moved from the periphery of historical attention to a place near the centre’.Footnote ¹⁸ Now, almost two decades after that, we need to take stock. It would be impossible to do justice to the rich complexity of witchcraft studies in the twenty-first century, so instead a handful of books has been chosen. This is highly selective, reflecting in part how this article grew out of a simple review to which more books were added as they arrived at the offices of the Historical Journal. I maintain, however, that the selection while far from comprehensive is broadly representative of current interests and ideas.Footnote ¹⁹ Given the difficulty faced even by specialists of staying on top of an endlessly expanding subject, the usefulness of Jonathan Barry's and Owen Davies's Witchcraft historiography is hard to overstate. In their introduction the editors survey the field, laying particular emphasis on interdisciplinary research. These days the study of witchcraft beliefs require the analytical tools not just of anthropology and sociology, but of folklore, psychology, and literary theory; a knowledge of community, society, and gender helps us to grasp the intricate mechanics of accusations; and science, medicine, religion, and politics benefit our understanding of trials. Barry and Davies explain why the history of witchcraft has been worked over, relentlessly and for so long, and why today it demands ‘to be understood in the context of the perceived realities of past societies’.Footnote ²⁰

The essays are written by leading authorities in their own sub-fields. Peter Maxwell-Stuart opens with a discussion of the relationship between magic and history in learned treatises. By 1600 demonological authors had many questions – for example, how should the law be used against witches? – but invariably posed them in a self-referential way, piously respecting assumptions that a century later would be overturned. Until then ‘a seamless continuity’ between truth and experience prevailed, each preserving the certainty of the other.Footnote ²¹ Christina Larner's framework is still helpful here. The witch-hunt began in an era of confessional state building, when the drive for political and religious orthodoxy converged and witchcraft was a kind of heresy, and declined in the age of absolutism after 1660 that saw the waning of moral paranoia, providentialism, and godly authority.Footnote ²² There was no avalanche of reason and disenchantment, rather a slowly evolving matrix of assertion, association, and meaning. By the eighteenth century, the theologian's defence of the world of spirits was struggling to make itself heard above the philosopher's defence of reason. And in the end, jurists were compelled to side with the latter.

The philosophy of witchcraft is developed in Peter Elmer's superb piece, which explains why the ‘witch-craze’ used to be harnessed to medievalism and the scientific revolution, and exactly why that is wrong.Footnote ²³ Rationalists from the eighteenth to the twentieth centuries found illustrations of intellectual progress in natural philosophy, medicine, and the demise of witch-hunting. In the 1930s Gregory Zilboorg hailed sixteenth-century sceptic Johann Weyer as the founder of medical psychiatry, a forerunner of Freud; in fact, Freud himself had saluted Weyer as, in Michaela Valente's words, ‘a physician affirming his right to cure his patients rather than burn them’.Footnote ²⁴ This, argues Elmer, is an anachronism. Allowed to speak in the idiom and intellectual context of his own time, Weyer seems less like a precocious philosophe and more like the Renaissance physician and demonologist that he really was: ‘a medical conservative, whose views were neither novel nor convincing, and whose approach to witchcraft differed little from that of his opponents’.Footnote ²⁵ Elmer stresses the importance of Stuart Clark, whose work has lowered the barrier between witchcraft and science, reconstituting demonology as the cutting-edge philosophy of its day – an inquisitive and empirical discourse that would talk itself out of existence precisely because of its appetite for truth and proof.Footnote ²⁶

This collection presents us with meanings of witchcraft as varied and nuanced as humanity itself, and exposes much of our knowledge of early modern witchcraft as the invention of a more recent past. Christa Tuczay describes the nineteenth-century historicization of witch-trials: the rational-liberal critique of medievalism, and its romantic-conservative counterpart at war with modernity. Both camps espoused nationalism. George Conrad Horst was a rationalist who believed that honest, enlightened Germans had fought the witch-hunt before 1700, while Frenchman Jules Garinet waved a similar flag for his own ancestors. Other writers, especially the romantic novelists, lionized witches not as innocents framed with an imaginary crime but as real witches undeserving of punishment.Footnote ²⁷ Juliette Wood's essay compares two writers for whom the reality of witchcraft was indisputable. Margaret Murray's quest to find a persecuted Dianic fertility cult in the archives was in the end no less fictive than the novels of Sir Walter Scott. But as late as the 1970s her influence was profound. Montague Summers (once described as a mixture of ‘spooks and sex and God’) believed that pagan votaries were satanists who presented as great a danger in 1930 as they had in 1630. He too had a lasting influence: on the horror movies of the 1960s and 70s.Footnote ²⁸

Marko Nenonen takes a long view of witch-beliefs in terms of acculturation and social control; Brian Levack surveys the history of witchcraft from the criminal-legal perspective; and Marion Gibson is illuminating on ‘new historicism’ and the ‘linguistic turn’, mapping the routes to the new feminist scholarship of Roper, Willis, and Purkiss, and to Clark's Thinking with demons. Until the 1990s ‘the language of witchcraft was unreadable’, explains Gibson, ‘because signifier and signified combined to produce a sign that – to the modern mind – meant nothing real at all’.Footnote ²⁹ Katharine Hodgkin's piece also deserves a mention here. She shows how attention paid to context and language helped to construct a contemporary reality for witchcraft – a reality that put gender back in the picture. In the previous generation, some historians (notably Macfarlane and Thomas) had neglected it, and, by ignoring empirical scholarship in preference for polemic, radical feminism had abused it. The fraught experience of motherhood and menopause now seemed more interesting than reductionist rants about misogyny, or structural-functionalist models that had little to say about emotion and experience.Footnote ³⁰

III

‘Witchcraft is a subject which puts the traditional conceptual frameworks of historiography under pressure’, writes Katharine Hodgkin, ‘making disciplinary boundaries more permeable’.Footnote ³¹ It is this dynamism that makes witchcraft such a historically rewarding subject, and yet it also generates endless contrasting images that resist intelligible unification. Not only is the sheer range of types baffling – the fairy-lore of the Scottish highlands, the shamans of Iceland, the werewolf trials of Franche-Comté, child-executions in Würzburg – but in many cases the historian is required to reconcile several distorted versions of the same event, like a painter attempting a portrait in a hall of mirrors. Compare the remarks quoted above from Keith Thomas in 1975 and Jonathan Barry and Owen Davies in 2007: the former spoke of the ‘reality’ of witchcraft, the latter of ‘perceived realities’. Plurals, perceptions, and multiple perspectives are the order of the day. But for students and experts alike, the question nags: how does it all hang together? We all need syntheses and it is to four such works – by Brian Levack, Gary Waite, Peter Maxwell-Stuart, and Wolfgang Behringer – that we now turn.

Levack's Witch-hunt in early modern Europe was first published in 1987, revised in 1995, and appeared in a new edition in 2006. Comparing these – or at least, reading their prefaces – is instructive. The book's raison d'être is unchanged: to offer a concise overview rather than a comprehensive history of witchcraft, which these days would probably have to be a multi-volume collaboration.Footnote ³² Levack is accessible to readers new to the subject, but in general avoids that common affliction of the textbook: blandness. The first edition emerged from many years of teaching, but Levack is also a leading archival researcher with ideas of his own. In the mid-1980s, between the Stockholm conference and the publication of its proceedings in English, he was already dissatisfied with the idea of a monolithic witch-hunt supported by a check-list of causes. He was also struggling to cope with the abundance of publications, blissfully unaware of what the next twenty years would bring. The preface to the second edition, dated November 1993, acknowledged not just the quantity of new studies, but their geographical and cultural diversity. A ‘grand unified field theory’ of witchcraft had never seemed more unlikely.

Like the second edition, the third is the product of bibliographical pressure. The preface demonstrates the emergence of an international community of scholars, a sign of the growing self-confidence of witchcraft studies. Despite the tidal wave of publications, Levack's main argument has not altered greatly. He believes that the impetus for witch-hunting was mainly local in origin, and that in most cases the centre was a force for restraint. He is also reluctant to attach too much significance to social and economic factors: these were present in communities that experienced no witch-trials (or had existed for years without incident), and were only faintly detectable or absent in many that did. Levack's thesis centres on the development of law courts in an age of state formation and religious reform. Social and economic conditions, he suggests, may have interacted with political, legal, and religious conditions to create ‘a mood of anxiety in all segments of society that made people more aware of the danger of witchcraft in the world and more eager to counteract it’. In this way, the latent and ill-defined concept of witchcraft took on a terrifying reality for both accusers and the people they accused.Footnote ³³

This remains the best undergraduate textbook: it is clear, succinct, and judicious, and masters recent historiographical trends. The final chapter of earlier editions has been expanded and divided to cover the decline of witch-hunting and its survival in the modern world – a reminder that neither Enlightenment nor judicial scepticism could remove this scourge. A fourth edition might benefit from more and shorter chapters, which students find useful. The contents page should at least list sub-sections, an improvement to the second edition now inexplicably reversed. For all its strengths, one wonders how many times a textbook like this can be reworked before the structure buckles under the weight of revision. Levack's title may also get in the way. Even the first edition suggested that ‘witch-hunt’ was a misnomer, implying a sustained campaign to eliminate witches. Objections could also be made about ‘Europe’. Levack includes material from colonial America and Scandinavia, but in the future a more global vision may need to be flagged up on the jacket. Finally: ‘early modern’. Nowadays historians recognize that the first major witch-trials occurred in the 1420s, even if the main body of prosecutions maps more directly on to the long Reformation. If your definition of ‘early modern’ brackets in most of the fifteenth century, there is no problem. Medievalists might demur.

Gary Waite's story begins in 1400, and emphasizes religious doubt, anxiety, and conflict. The repositioning of the witch-hunt to connect its ‘classic’ age (defined here as 1562–1630) with its pre-Reformation prologue is a welcome departure. Most impressive is Waite's fine shading of the medieval crimes of heresy and witchcraft, and the way that the former morphed into the latter. The perceived threat of heresy to the church in the later middle ages, especially its increasingly satanic and apocalyptic character, was an essential preparation for later demonology and judicial operations. By the mid-sixteenth century an elaborate concept of witchcraft had crystallized, an image crudely rendered in some general accounts.Footnote ³⁴ Like Peter Elmer and Stuart Clark, Waite suggests that ‘witch-hunters’ and ‘witchcraft-sceptics’ were much closer in outlook than rationalist retrospection would have it. Johann Weyer and Jean Bodin, usually depicted as ideological enemies, ‘contended primarily on whether or not witches were victims of diabolical illusions … or willing participants in a real, diabolical conspiracy’.Footnote ³⁵ As ever, our appreciation of meaning is only as good as our ability to understand context.

The decline of witch-trials was closely connected to their rise. More than Levack does, Waite follows Larner and Clark in his assertion that the twilight of confessional states, religious enthusiasm, and enforced orthodoxy helped turn opinion against witchcraft as a prosecutable crime. But then witchcraft had always occasioned debate: even Heinrich Kramer, author of the Malleus maleficarum (1487), had more critics than champions in his day.Footnote ³⁶ Indeed, the fact that witch-hunters were ready to argue, and to execute people to underline a point, would in the end prove more significant than the quality of their arguments. Ironically, by 1700 a doctrine that had helped rulers to purge ambiguity and stifle dissent had imploded: ‘the process of prosecuting witches itself’, observes Waite, ‘raised more doubts about the supernatural realm than it resolved’.Footnote ³⁷ As would happen again with spiritualism in the twentieth century, asserting the truth of interaction between the seen and unseen worlds depended on evidence: initially the arena was the courtroom, later the seance-room. In each case graphic proof attracted believers, then lost them when the validity of that proof was questioned.Footnote ³⁸ The apex of witch-hunting therefore marks the beginning of the end. The persuasive force of confessions, and ensuing judicial savagery, were matched by the intensity of scepticism in the aftermath.Footnote ³⁹

Establishing an empirical basis for the reality of witchcraft, then, troubled theologians and jurists long before it became a problem for historians. The issue of the witches' sabbat is a prime example.Footnote ⁴⁰ Its ideological refinement in the fifteenth century is lucidly described in Peter Maxwell-Stuart's short history of witchcraft. The author has a lot of ground to cover, but patiently reveals the archaeology of an idea dating back to classical antiquity and early Christianity. He makes the point that the demonization of heresy had the converse effect of encouraging clerical inquisitors to see all magical activity as implicitly heretical. The explicit diabolism of the nocturnal assembly, where witches were said to kneel before Satan (and worse), completed the picture. This intellectual evolution also strengthened the feminine witch-stereotype, paving the way for the excessive, if not necessarily eccentric, misogyny of Kramer's Malleus maleficarum.Footnote ⁴¹

Maxwell-Stuart is good on the iconography of witchcraft, still an under-researched area.Footnote ⁴² His book is brisk, well-written, vivid, and gently steers the novice away from persistent historical myths – some of the most damaging seeded by Margaret Murray in the 1920s and nurtured by pagans and feminists fifty years later. In line with current thinking, witchcraft is seen in the light of its own times, playing down the impact of the Malleus, and demonstrating that stereotypes could exist without necessarily dictating experience.Footnote ⁴³ Maxwell-Stuart also strips ‘sceptical’ treatises of their proto-rationalist status. Reginald Scot's Discoverie of witchcraft (1584), for example, was a blast of anti-Catholic and anti-clerical religious invective, but under scrutiny appears to be ‘not as straightforward a work of modernizing doubt as it may appear at a first or rapid reading’.Footnote ⁴⁴ These are helpful adjustments. Apart from a feeble index, the book's only real flaw is a structure which limits the discussion to western Europe and north America; the author's explanation for this – mainly that ‘the time is not yet ripe to undertake a comparison’ between European and extra-European witchcraft – is not entirely convincing.Footnote ⁴⁵

Wolfgang Behringer's title alone suggests that he would be unconvinced.Footnote ⁴⁶ This is a ‘global history’, a practical manifesto for the comparative approach. Drawing on history and anthropology, Behringer widens the focus on early modern Europe to display witchcraft as a human phenomenon expressed differently in different times and places. The first third of the book discusses historiography, witch-beliefs, and early persecutions, presenting us with a reality of witchcraft built on its universality. Witch-beliefs are found among educated Africans not because they are ignorant of some truth revealed to the developed world, but due to the stresses of political and economic transition. Behringer cites Peter Geschiere's ‘modernity of witchcraft’, as flagrant a rebuttal of the rationalist position as one can imagine.Footnote ⁴⁷ It is all the more intriguing, then, that the intellectual and judicial origins of early modern witch-trials lie in the earlier fifteenth century. Behringer makes a strong case for the medieval witch-hunt, even if the post-1560 period ushered in a ‘new style of witch-hunting … a paradigm shift’.Footnote ⁴⁸ At the same time, he is wary of the usual litany of causes: state-building, confessionalization, the growth of demonology, inquisitorial tribunals, and torture. These factors were not always applicable. Behringer is keener than Levack on economic factors, or more precisely the effect of climate: the so-called ‘little ice age’.

A large part of this original and thought-provoking book is taken up with witch-hunting in the post-1700 period, a continuity that may surprise many students. And yet the best chapter deals with the early modern period, although here the anthropological insights are less well integrated – in fact, not much in evidence at all. Later sections do glance backwards at the sixteenth and seventeenth centuries, but only intermittently and not always with obvious benefit to explaining the phenomenon. The partial and the particular still seem to trump the global and the universal, and the question that was asked of Macfarlane and Thomas in the 1970s remains a valid one: are we really comparing like with like?Footnote ⁴⁹ Ronald Hutton, another leading historian sanguine about collaboration between historians and anthropologists, is more realistic. ‘With so many variables present in the model it is not surprising that many scholars, with specific preoccupations, have found a comparative approach to the subject problematic or unhelpful.’Footnote ⁵⁰

Even so, Behringer's conceptual creativity and willingness to express opinions are the book's strengths. Its weakness is that it is not always easy to read: the long chapters have no sub-headings, the text is peppered with distracting italics, and the prose is sometimes awkward. There are mistakes, too. I will confine myself to an area I know well, but readers might be concerned if the degree of error on page 132 is representative. Matthew Hopkins died in 1647, not 1648, and was not from Manningtree in Essex but Great Wenham in Suffolk; swimming and ducking were not the same, nor was the water ordeal known as ‘the bath’: Behringer has misread the word ‘hath’ in John Stearne's account.Footnote ⁵¹ He also trusts the witchfinder's claim that he and Hopkins had executed 200 witches, given he would have ‘tried to minimize the numbers in his apology’. The opposite is the case: Stearne exaggerated because his work is a boastful riposte to critics. For him to speak of hundreds of witches was to broadcast a warning about their reality in the world and so vindicate himself. On the same basis, and because of some evidence to the contrary, it is unlikely that sixty-eight witches were condemned at Bury St Edmunds in 1645, as Stearne maintained.Footnote ⁵² The reality of the witchfinders' mental world is all the more difficult to grasp because they themselves struggled with it, mainly the hard task of persuading a contemporary audience of the truth of their experiences, perceptions, and beliefs.Footnote ⁵³

IV

A thorough understanding of a single episode suggests that the devil really does lie in the detail. In the last decade, there has been a move away from English county studies towards individual stories, meticulously explored at book-length where once an article would have sufficed.Footnote ⁵⁴ It is to Peter Maxwell-Stuart's credit that he devotes several pages to describe an isolated Scottish trial, demonstrating the role of the contingent and the special lethal energy of an accusation when circumstances were aligned precisely. Unlike England, Scotland has an abundance of legal records that permit this sort of research, and recently a number of scholars have been making the most of them.Footnote ⁵⁵ Two new works – a volume of essays edited by Julian Goodare, Lauren Martin, and Joyce Miller, and a monograph from Brian Levack – draw much of this together.Footnote ⁵⁶ Goodare begins by setting Scottish witchcraft in context, and concludes with some comparisons between the approaches of Robin Briggs, Ronald Hutton, and others. He suggests that the historical reality of witchcraft resides not just in interaction between elite and plebeian ideas, but also in the relationship between culturally embedded magical beliefs and inherent biological imperatives. This is witchcraft in the blood, not merely some construct of the interrogation chamber.Footnote ⁵⁷

The accent on beliefs in this collection represents witchcraft as a living part of early modern culture. Joyce Miller makes a brisk survey of satanic manifestations, observing that ‘in popular culture the devil was more effective disguised in domesticity than disgust’; similarly Owen Davies's contribution on cunning folk examines the power not of the agencies of reform and repression but of common practice. Unbroken adherence to charming, healing, second-sight, and so on ‘provides one of the most striking proofs of the resilience of popular culture in the face of excommunication and execution’.Footnote ⁵⁸ An essay about charms against witchcraft preserved in Scottish museums is particularly absorbing. These include bones, stones, and crystals, the significance of which was long clouded by the rationalist spirit in which they were collected and classified; according to Hugh Cheape this served mainly ‘to demonstrate the advances of Victorian society and to distance the citizens of the nineteenth century from the past’. He concludes that to invest an object with magical force is to reify a miasma of beliefs and feelings, the historical recovery of which the modern student of witchcraft is compelled to pursue.Footnote ⁵⁹ There are solid pieces by Edward Cowan (on the beliefs of the lowlands), Lizanne Henderson (on highland belief), and Michael Wasser, who, following modern trends in intellectual history, sees the decline of witch-beliefs as another Kuhnian ‘paradigm shift’.Footnote ⁶⁰

Brian Levack contributes a study of demonic possession, a version of which forms a chapter in his excellent new book, Witch-hunting in Scotland. Levack is a no-nonsense historian, and his subtitle ‘Law, politics and religion’ sets the agenda: these are the areas of activity in which, he believes, Scottish witch-trials need to be understood. He makes a convincing case, reinforced by detailed comparisons not with modern Asia, India, or sub-Saharan Africa, but with seventeenth-century England. This approach makes a lot of sense and is highly revealing; in fact, one wonders why it has not been done before.Footnote ⁶¹ A chapter entitled ‘Witch-hunting in revolutionary Britain’ stands out, exploring in detail Anglo-Scottish connections overlooked in my own book on English witch-trials in this era.Footnote ⁶² In exchange for Levack's major contribution I offer a single correction to a small but common error: the Sir Harbottle Grimston who helped persecute witches in Essex in 1645 is not the celebrated namesake critic of Archbishop Laud but his father.Footnote ⁶³

As in Levack's other work, the role of the state is key, not as the accelerator of witch-hunting but as its brake. Whereas Scottish religious politics had influenced English witch-trials in the 1640s, in the 1650s the assertion of English legal control, especially the introduction of circuit judges, reduced the number of convictions, which in turn must have deterred accusers. Political uncertainty after the Restoration led to another burst of persecution, but from 1662 onwards concerted action by the privy council restrained local authorities, principally by denying them the commissions of justiciary they needed to hunt witches. Challenging Christina Larner, Levack argues that the Scottish state was more likely to cause witch-hunts when it was weak than when it was strong; witch-hunting was a failing of absolutism, not its enabling policy or a sign of hegemony. That witchcraft was real in hearts and minds is beyond doubt; but its tangible social reality, most convincingly shown in judicial executions, was less consistent and substantial because executions were actually quite rare. The cause lies not in a shortage of reasons to initiate a witch-trial, but in a profusion of reasons not to initiate one.

Rothenburg ob der Tauber, an autonomous Franconian city answerable only to the Holy Roman Emperor, saw only eighteen cases of witchcraft between 1561 and 1652, involving forty-one suspects of whom just one was executed. This ‘restrained pattern of witch-hunting’, expertly investigated by Alison Rowlands, can be explained by the way that the sixteen-man council strove to prevent the calamity that befell other cities in southern Germany, notably Würzburg where 160 people were burned in 1629.Footnote ⁶⁴ As Levack suggests for Scotland, in Rothenburg the law worked mainly to inhibit trials. It hardly matters whether the city governors detested witchcraft less than elsewhere – they almost certainly did not – only that they valued the preservation of honour more. Indiscriminate accusations attacked the reputations of blameless people and threatened economic stability. Here crimen exceptum meant a crime to handle with exceptional care, not one exempt from normal evidential standards. Counter-suits for slander were encouraged, so that peasants were afraid even to talk about witches; hence people voluble about their suspicions were more likely to be taken seriously. But this policy did not necessarily mean that more reticent folk suffered in silence: a lack of trials promoted the feeling that witchcraft was not a problem, just as big trials advertised the opposite. All this reminds us that the hot-headed strife of the Reformation did not lead automatically to witch-hunting. There were only ever two executions on the Isle of Man, a place like Rothenburg where most cases of witchcraft surfaced as slanders and accusations were, in Jim Sharpe's words, ‘just as likely to lead to clergymen shaking their heads in dismay at the ignorance of their flocks, and at their inability to live in peace with each other’.Footnote ⁶⁵

Rich archives allow Rowlands to describe thickly the structures and textures of early modern Rothenburg and to reanimate its people. Life seen in the round obviates simplistic views of witch-hunting and undermines obdurate myths. Witch-trials were not concerned with hysteria but with politics: authority, loyalty, honour, discretion, alliance, and faction – matters of tactics and strategy, not the blind imposition of policy or belief. Like ourselves, these people did not always say what they thought, or did what they said: ‘it was possible to hold and express harsh opinions against witches in theory while remaining cautious in the handling of specific witchcraft cases’.Footnote ⁶⁶ Peter Elmer and Ian Bostridge have demonstrated how closely English witchcraft and politics were connected, work that requires uncommon delicacy and erudition.Footnote ⁶⁷ Rowlands skilfully reconstructs the political contexts of specific trials, describing conflicts over contested lordship in Rothenburg's hinterlands, plus battles both territorial and confessional. One case involving the revelations of some attention-seeking girls lays bare the competing claims of Catholicism and Lutheranism over the souls of the people.

Rowlands makes witchcraft utterly intelligible at macro and micro levels, and her account is all the more realistic because as an explanation for misfortune bewitchment was possible but never probable, still less inevitable. This is subtle, sophisticated history. Nor do her connections between state and people stop with neighbourhood and household: she reaches further into the psychic realm to consider personal experience – the turbulent politics of the individual constitution, including the complex feelings of the accused. In one episode from 1587, a six-year-old boy confessed to flying with his mother to the witches’ dance accompanied by a black man, plausibly suggested to be a sprightly surrogate for the boy's elderly father and a symbol of his departure from a feminized existence.Footnote ⁶⁸ Detail brings nuance, and with nuance a sense of authenticity: as with other histories of witchcraft we encounter ambiguity of meaning here, but also demonstrable ambivalence of feeling. Rowlands has no time for feminist reductionism, arguing that to see witches and accusers as ‘the cowed or brutalized mouth-pieces of all-powerful patriarchal elites over-estimates levels of elite enthusiasm for witch-hunts, ignores the importance of the motives and personalities of alleged witches and their accusers to the trial-process, and misunderstands fundamentally how the law worked’.Footnote ⁶⁹

So often our assumptions are wrong, insulting in their simplicity and superficiality. Like most European states in the later seventeenth century, the trend in Rothenburg was for increased judicial caution; but even there things did not go as one might expect. As hope of reaching clear conclusions about the guilt of witches faded, so the council took a harder line about their theoretical reality and the peril they represented. Another twenty-six people were tortured and tried after 1660, and two of them executed – in 1673 and 1692.Footnote ⁷⁰ The rationalist paradigm, where declining witch-trials are linked to an elite laughing off peasant superstitions, cannot stand. Brian Levack's account of the end of Scottish witch-hunting establishes this firmly. He admits that it is hard to say when Europe's ruling class gave up believing, ‘but we can be confident that very few of them did so until the trials were almost completely over’. Change came not in the salons and coffee-houses but in the courts, where witness testimony was distrusted and confessions discredited. In theory, maleficium might still explain misfortune, but ‘because the burden of proof was on the prosecution, all that was necessary to secure acquittal was evidence that natural causation was possible’. Meanwhile, legal defence for suspects became more common, leading to frequent dismissals of hearsay and exposure of malice.Footnote ⁷¹

Another historian who eschews reductionism is Emma Wilby. Her Cunning folk and familiar spirits aims to make magic look as real as it did to its early modern practitioners. She explains her methodology thus: ‘if an early modern familiar belief can be found to resemble a familiar belief found in a traditional non-European animist culture then this resemblance will be seen to support the likelihood that the former was rooted in pre-Christian British animism, that is, that it was of folkloric origin’.Footnote ⁷² This may be good anthropology; but is it good history? The third section of her book aims to recover ‘the experiential dimension’. To that end she draws on native American beliefs and Siberian shamanism, an approach that, ironically, jolts us out of the early modern world she is at pains to represent. Instead of rationalism and reductionism we get uncompromising relativism. Another irony about Wilby's method is that her aversion to the liberal-rational paradigm edges her towards the romantic tradition. It is not just that her witches are active agents rather than passive victims: she comes as close to establishing the reality of spirits as is possible to do without actually affirming their existence. Freud, Durkheim, and Weber considered witchcraft to be a kind of collective and individual neurosis, a perspective leading us into the dark caverns of the psyche. But again this is too reductionist for Wilby, as is my own analysis of the story of Margaret Moore; to locate Moore's visions in her emotions, Wilby believes, constituted a dismissive ‘pathologizing’ of real experience.Footnote ⁷³

Wilby's paradox is that she does not want to say that spirits were real, but cannot account for their presence in early modern culture without explaining them away. To seek the meanings of phenomena but not their causes may be possible, but feels ahistorical. Alison Rowlands achieves more by immersing herself in an archive, setting out a context, and exercising judgement. Wilby's more tenuous comparisons and wilder speculations risk driving students back to traditional empiricism, even back to the rationalist paradigm. Historians need to tread a middle way, constantly aware of the spiritual dimension of witch-beliefs but compelled to silence by lacunae in the evidence. Owen Davies, Jim Sharpe, and others reproached by Wilby are not wilfully uncooperative, just observant of the limits of historical interpretation. Even Diane Purkiss, a scholar not shy of pushing the boundaries of psychological meaning, is criticized here for only seeing the fantasy element of witch-confessions, something rendered superficial and unspiritual in the very act of its identification. Exactly what ‘spiritual’ means, we are not told.

Wilby makes some dubious claims. The idea of the familiar spirit imposed by demonologists, she argues, originated in the fairy lore of cunning folk. But most clerical writers on witchcraft had no time for the familiar, a numinous intermediary as redundant as priests in the relationship between man and God; like the Christian covenant, the compact with Satan took place in the heart, not by means of the obscene suckling of a cat or a rat. Wilby also asserts that formal religion was subordinate to fairy-lore; that most people understood little of God; that many cunning folk never attended church (how does she know?); and that ‘the guiding principles in the lives of many ordinary people in early modern Britain were essentially animist rather than Christian’.Footnote ⁷⁴ Keith Thomas is censured for his lack of spiritual sensitivity, but this involves reprising much of what endures in Religion and the decline of magic. Wilby also repeats herself, and needs a better editor. There is a general lack of context – she treats the Suffolk confessions of 1645 as if they are representative rather than virtually unique – and she makes mistakes. The Susan Swapper case of 1607 occurred in Rye (Sussex) not Wales: see the article by Annabel Gregory.Footnote ⁷⁵ William Perkins ‘observed’ nothing in 1618 because he was dead; Wilby may have meant ‘1608’ – the year when Perkins's treatise appeared, but even that would have been wrong because he died in 1602 and his work published posthumously.Footnote ⁷⁶ The publication date for Religion and the decline of magic is consistently given not as 1971 but as 1991 – not a decade Thomas dared to enter as a historian of witchcraft, as we have seen.

V

The last three centuries of witchcraft historiography can be summed up by an alliterative sequence: rationalism, romanticism, relativism, and realism. These are not discrete categories in neat chronological succession; but they do represent four key ways of thinking about witches. Of course, in its own way each aims at realism; even the phantasmagoria of the romantic novelists looked to convey moral truths. Too often, though, attempts to explain witchcraft have been exercises in rationalization, reconfiguring phenomena into familiar forms. As Marion Gibson observes, when feminists argue that witch-hunting was really women-hunting, or functionalists show it to be a way of alleviating guilt, they are exchanging an alien reality for one we understand.Footnote ⁷⁷ Witness testimonies and confessions get culturally translated. When people said they saw the devil, it is decided that they must have been mad or dreaming or hallucinating or lying; we know people imagine things and that they lie, but we cannot easily think about the reality of Satan. With regard to demonic possession at Salem in 1692, David Harley has highlighted ‘an epistemological problem at the heart of most histories of witchcraft that makes it difficult for historians to hear the explanations offered at the time’.Footnote ⁷⁸ Bewitchment has been written about as syphilis, encephalitis, rabies, food poisoning, and many other forms of ‘cultural materialism’, but remarkably rarely as witchcraft.Footnote ⁷⁹ What historians can forget is that their reality is also subject to relativism. When we write about early modern people, it is all too easy to end up writing mostly about ourselves.

The trick for historians is to be analytical and objective without being too reductionist or anachronistic; but it is hard to pull it off to everyone's satisfaction. For trying to understand a belief-system, functionalists might be called relativists, but seen another way they are reductionist rationalists because they make witchcraft conform to a rationale. Anthropology can avert the condescension of posterity and contaminate a complex and unique cultural environment with a modern social science. Hugh Trevor-Roper and Norman Cohn stood somewhere between rationalism and relativism, aware of the witch-hunt as an act to be condemned but eager to understand its intellectual mechanisms. The idea that all witch-beliefs were logical, however, was hard to stomach. This relativism, pioneered by Keith Thomas, pushed older historians closer to the sort of rationalism encapsulated by the title of Charles Mackay's 1841 work Extraordinary popular delusions and the madness of crowds. And yet the more extreme relativism deployed by Emma Wilby would surely vex Thomas for its empirical weaknesses. Her position recalls the ontological problem created by nineteenth-century novels: ‘the dual impulse … to highlight the chimera of witchcraft and condemn the persecutors of witches was compromised by the fictional device of the representation of witchcraft as a reality’.Footnote ⁸⁰ Historians can only travel so far into a psychic realm that cannot be substantiated before they lose touch with a social world of the past that can. At this point the history of fantasy becomes fantasy itself, and relativism slides into romanticism.

Today, most university departments offering courses on witchcraft will teach some kind of relativism, possibly with an eye on realism. By ‘realism’ here I mean history committed to the recovery of meaning and experience from above, below, and within, epitomized in the work of Stuart Clark, Robin Briggs, and Lyndal Roper respectively. It is striking, however, that while witchcraft studies are so vibrant and innovative public perceptions remain stridently rationalist and romantic. The Harry Potter phenomenon is testament to the latter, and the former is abundantly evident in the press, broadcast media, websites, museums, heritage trails, and guidebooks. A recent children's book, though well researched, still relates the witch-hunt as a story of ‘cruelty, stupidity and horror’.Footnote ⁸¹ Emotional bids to pardon and memorialize executed witches are likely to appear more frequently. In 2007 the canton of Glarus in Switzerland petitioned its regional government to pardon Anna Göldi, the last executed witch; and in Britain the campaign to exonerate Helen Duncan, the last woman punished under the 1735 witchcraft act – in 1944 – has received support from the Salem Witch Museum and a Scottish baron who has pardoned the eighty-one tenants from his estate executed for witchcraft in the sixteenth and seventeenth centuries. The word ‘injustice’ is used frequently; but without a measure of historical relativism (or sometimes even reliable history) we cannot reach the reality of what happened, why, and what that meant in the context of its own era.Footnote ⁸²

So where do we go from here? There can be no straightforward answer because those engaged with the history of witchcraft have different desires and destinations, even within academia. There is no agreement even about basic definitions, the term ‘witch’ having ‘undergone mutations of meaning that create a potentially hopeless confusion in common parlance’.Footnote ⁸³ Hermeneutic debates about the Hebrew name date back to the seventeenth century, a sign that this is not so much a world we have lost as one we have never quite left behind. The final impediment to the pursuit of reality is the fact that witchcraft never had a fixed reality in the past. Ideologues like Heinrich Kramer in the 1480s, Jean Bodin the 1580s, and Joseph Glanvill in the 1660s, all painted vivid pictures of the activity of witches precisely because the image in most people's minds was so indistinct. Opponents Bodin and Johann Weyer, or Glanvill and John Webster, were contemporaries breathing the same intellectual air, not the beginning and end of an arc of progress between credulity and scepticism.Footnote ⁸⁴ Stuart Clark's thesis may have started out as a critique of positivism, but there is something to be said for the old-fashioned notion of the historian keeping quiet while the sources do the talking. Only then can we hope to know what made demonologists, witnesses, and witches tick.

We need to respect the embeddedness of witchcraft, its organic dynamism, and connectivity to all other areas of early modern life. Alison Rowlands's elision of civic governance in Rothenburg and the neuroses of confessing witches is a model, as is Mary Beth Norton's In the devil's snare, which by similar means manages to say something original about Salem; we should also include Robin Briggs's outstanding new book, which looks beyond demonology and witch-hunting to show how suspects ‘formed an integral part of the therapeutic system of the times’.Footnote ⁸⁵ In all three cases we find deep learning, hard archival labour, a firm grasp of method, sensitivity to language and context, and great powers of historical imagination at work. We also see witchcraft as a political variable, part of and subject to local politics, individual conscience, and the vagaries of decision-making in communities forced to weigh up a plethora of contrasting ideas, attitudes, and intentions. This approach treats experience as reality, but the two are not equated in an exclusive manner – the pitfall of extreme relativism. In a critique of John Demos's analysis of the Salem witch-trials, David D. Hall pointed to the essential conflict between detailed historical contextualization and modern theorization. Hall doubted that events in 1692 were rooted in the psychic disturbance of puritan New England, and suggested that in the end most readers would opt for an explanation of witchcraft based on ‘a commonsense psychology of guilt and projection’.Footnote ⁸⁶

There is one final juxtaposition: against the fuzzy and fragmented reality of witchcraft in the early modern period must be set the graphic reality of witch-hunting in our own times.Footnote ⁸⁷ If anthropologists and historians of witchcraft do find a way to get along it might just be because the former have more to learn from the latter. Historical witch-hunts are more than just case-studies in paranoia and persecution, their abstract lessons resonating in our own age of contested ideologies, religious fundamentalism, international warfare, collapsing empires, and unrivalled state power over the individual. Early modern witchcraft trials actually help us to understand living witch-hunts in regions of Africa or India where the legal and political influence of the state is weak, ethnic conflict and civil war are endemic, animist beliefs are widespread, and poverty and disease are rife. In places and situations where anthropologists literally fear to tread we can, through the history of witchcraft, gain a strong impression of what might be happening there – the past imported into the present for a change. All of which is salutary, and a reminder of common cause between the two disciplines, even if in the process it exposes the perpetual human horror of witch-hunting. There is nothing archaic about witchcraft, whether from the rationalist, romantic, or relativist standpoints. Rather, as Brian Levack has observed, ‘many of the issues that early modern jurists debated with great intensity have not been resolved but have been restated in somewhat different geographical contexts’.Footnote ⁸⁸