on March 22, 2026
Gustav Henningsen, 1980
Stream Three · The Basque Library · Entry 06
The sentence that brought Alonso de Salazar Frías to the attention of historians was recovered from the archive in Madrid by Gustav Henningsen after more than three centuries of obscurity. It appears in Salazar’s fifth and final report to the Supreme Council of the Inquisition, completed in 1614, summarising three years of investigation into the witch panic that had swept across Navarre and the Basque Country. He had travelled through the affected region for eleven months, collecting testimony from nearly two thousand people who confessed to attending witches’ sabbaths, examined hundreds of items of physical evidence submitted as proof of witchcraft, subjected the confessions to systematic cross-examination against independent witnesses, and arrived at a conclusion that the evidence did not support. His sentence reads, in translation: There were neither witches nor bewitched until they were written and talked about.
It is one of the most remarkable sentences in the history of European legal reasoning. A man operating inside the institution that had organised the terror, holding the power to prosecute thousands, states plainly that the entire category of offence does not correspond to any reality he has been able to find, and that the act of naming it — the sermons, the pamphlets, the testimonies extracted under torture, the public spectacle of the auto de fé — had itself produced the phenomenon it purported to describe. He is not saying that Basque folk belief contains no unusual elements. He is saying that the specific organised witch-sect — the sabbath, the pact with the devil, the systematic maleficium — is a confabulation: generated by suggestion, sustained by procedure, and now running on its own momentum with no connection to any underlying reality.
The man who found this sentence, and who reconstructed the story it belongs to from an archive whose original documents had largely been destroyed, was Gustav Henningsen — a Danish folklorist who spent years in Madrid working through the surviving inquisitorial papers before publishing The Witches’ Advocate: Basque Witchcraft and the Spanish Inquisition, 1609–1614 in 1980. This post is about the book. But it is also, necessarily, about the events the book documents, and about the methodological lessons Henningsen drew from those events — lessons that changed how historians think about witch trials, confessions, and the relationship between institutional procedure and the reality it claims to adjudicate.
The fire from France
The Basque witch panic of 1609–14 did not begin in the Basque Country. It crossed the border from France, carried by fear.
In early 1609, King Henri IV of France commissioned a Bordeaux judge named Pierre de Rostéguy, lord of Lancre — his family had Basque ancestry, a fact that seems to have intensified rather than moderated his feelings about Basques — to investigate reports of widespread witchcraft in the French Basque region of Labourd. De Lancre arrived in July 1609 with broad powers of arrest and execution. Over four months he worked through the parishes of the coast and the mountains, interrogating hundreds of people — many of them children, many of them women whose husbands were away fishing on the Grand Banks of Newfoundland — and conducting executions with a speed and certainty that reflected his prior convictions rather than his evidence. He had, before leaving Bordeaux, already concluded that the Basques were peculiarly susceptible to diabolical influence: their language was impenetrable, their customs were strange, their dance was lascivious, their priests were corrupt. His Tableau de l’inconstance des mauvais anges et démons, published in 1612, is a masterwork of demonological self-conviction — a text so saturated with certainty about what he was looking for that it found it everywhere it looked.
By the time de Lancre was recalled, in late 1609, he had executed somewhere between sixty and eighty people. The precise figure is uncertain because his records were destroyed in the French Revolution, and the estimate of six hundred executions that circulated in nineteenth-century histories turns out to have been a misreading. Sixty to eighty is the consensus of current scholarship. It was enough. The Basque Country was a single community divided by a border that had only recently been formalised, and news of what was happening on the French side produced immediate panic on the Spanish side. Refugees from the Labourd trials crossed the Pyrenees, carrying their accounts of sabbaths and denunciations. A young woman who had been in France returned to the village of Zugarramurdi, near the French border in Upper Navarre, and announced that she had attended the sabbath; she named names.
Zugarramurdi is a village of a few hundred people in a narrow valley, separated from its French neighbours by mountains and connected to them by the same language, the same customs, the same landscape, and now the same terror. Its name, as the philologist Koldo Mitxelena suggested, probably means something like place of the ruined elms, though he acknowledged uncertainty about this; the Basque suffix -di denoting abundance is the clearest element. It had been unremarkable before 1609. It has been unremarkable in most respects since. In between, it hosted the largest witch trial in European history in terms of the number of people investigated, and it gave its name to a legal and cultural episode that is still being argued about.
The auto de fé at Logroño, November 1610
The first phase of the Spanish trials moved quickly. The local village priest, seeing the accusations multiply, sensibly tried to contain them by offering a parish-level pardon to anyone who confessed; but the Abbot of the nearby monastery at Urdax, presumably with an eye on his own visibility, overrode this by sending the accused on the hundred-kilometre walk to Logroño to submit themselves to the formal inquisitorial tribunal. The Inquisition of Logroño had jurisdiction over Navarre and the Basque provinces — it was based in Logroño, a city to the south in La Rioja, and operated from a distance that gave it limited understanding of and even more limited sympathy for the specific cultural context of the people it was trying.
Forty people from Zugarramurdi were eventually taken to Logroño. The interrogation of the prisoners, conducted with the usual instruments of inquisitorial procedure including torture where deemed necessary, produced confessions of sabbath attendance, pacts with the devil in goat form, the making of ointments from the bodies of murdered children, weather magic, vampirism, and necrophagy. These confessions should be read not as evidence of what the accused actually believed or did, but as evidence of what the interrogators were looking for: the testimony was shaped by the categories provided to the witnesses, who were given a vocabulary of sabbath practice derived from demonological manuals and then asked to confirm or expand it.
The auto de fé was held on 7–8 November 1610. Thirty thousand people came to watch — a figure that requires context: this was an enormous public spectacle in a society where such events served simultaneously as judicial ceremony, theological demonstration, and popular entertainment. Twenty-one penitents were dragged through the streets in the traditional procession. Eighteen who had confessed and shown contrition received sentences ranging from imprisonment to public humiliation. Eleven were condemned to the severest penalties: six who had refused to confess or had recanted their confessions were burned alive; five whose deaths in prison had forestalled burning were represented by effigies and their exhumed remains, burned alongside the living.
Those six who burned — Domingo de Subildegui, María de Echachute, Graciana Xarra, Maria Baztan de Borda, Maria de Arburu, and Petri de Joangorena — were condemned, in the logic of the procedure, precisely because they would not confess. Those who confessed and showed remorse were spared execution. This is the inquisitorial paradox at its most stark: the truth of the accusation was confirmed by confession, and the refusal to confess was taken as evidence of obdurate guilt. The silence or denial of the accused was itself incriminating. There was, in strict procedural terms, no position from which innocence could be demonstrated.
After the auto de fé, the panic intensified rather than subsiding. The public spectacle had, precisely as Salazar would later diagnose, generated further accusations. Within months, the estimate of active witches in the region had grown to thousands. The second phase of the trials began.
Salazar and the inductive method
The Inquisition of Logroño had three judges: Alonso de Becerra Holguín, Juan de Valle Alvarado, and Alonso de Salazar Frías. Salazar was the youngest and most junior, a trained lawyer who had been educated at the University of Salamanca and ordained as a priest before taking up a career in inquisitorial service. He had participated in the first phase of the trials without dissent. What changed him was the scale of the second phase.
As the accusations multiplied across the region in late 1610 and 1611, the Supreme Council in Madrid — the governing body of the Spanish Inquisition, which had historically been more cautious about witch prosecutions than the local tribunals — issued an Edict of Grace: a promise of pardon to all those who voluntarily confessed and denounced their accomplices. The edict was the instrument Salazar was given to work with. In January 1611 he set out on an eleven-month journey through the Basque Country and Navarre, travelling from village to village, hearing confessions and collecting denunciations.
What he collected was extraordinary in scale. Nearly 1,802 people confessed to attending sabbaths. They named approximately five thousand others as fellow participants. The majority of the confessors — 1,384 of them — were children between the ages of seven and fourteen. The total number of people named in the process, by the end of Salazar’s investigation, was somewhere between seven and eight thousand, in a region whose total population was perhaps one hundred thousand.
The scale should have been self-confirming: surely so many confessions, from such varied sources, in so many villages, constituted overwhelming evidence. Salazar thought otherwise. He had been trained in the inductive method, which held that legal proof required not confession but corroboration — independent evidence that the confessed events had actually occurred. He set his secretaries to work systematically verifying the claims in the confessions. Had the events been witnessed by sober observers who were not themselves accused? Were the physical marks on the bodies of the accused consistent with the injuries they claimed to have sustained? Did the ointments and substances submitted as evidence contain anything identifiable? Was it possible, in the cases where confessors named specific dates and specific locations for sabbaths, to establish that no unusual gathering had occurred?
The answers were uniformly negative. Salazar’s secretaries visited sites where sabbaths had allegedly been held, on the nights when they were allegedly held, and found nothing. Witnesses who claimed to have seen a neighbour flying through the night sky turned out, on examination, to have been asleep themselves, or standing in a doorway, or looking at a distance in bad light. The ointments contained no active ingredients beyond some herbs. The physical marks on the bodies of the accused were consistent with ordinary skin conditions, insect bites, and the effects of torture. The confessions, when examined against each other, contained inconsistencies that would have been impossible if the sabbaths were real events attended by multiple people who could verify each other’s accounts.
Salazar’s five reports to the Supreme Council, delivered between 1611 and 1614, built this empirical case systematically. He did not deny that witchcraft was theologically possible. He did not claim that the accused were lying. His argument was more precise: that the confessions were the product of a specific social and psychological process — suggestion, fear, the desire to please interrogators, the dream-like vividness of certain states of mind — and that this process produced testimony that felt, to the people giving it, genuine, while corresponding to no events in the external world. The dream epidemic, as Henningsen would call it, was real; the sabbaths were not.
The key phrase — there were neither witches nor bewitched until they were written and talked about — does not mean that no one in the region believed in witchcraft before the trials. It means that the specific, organised witch-sect, complete with sabbaths and pacts and devil-worship, did not exist as a social reality until the machinery of accusation and confession created it. The trials manufactured their own evidence.
Henningsen’s reconstruction
The problem Henningsen faced in writing The Witches’ Advocate was archival and fundamental. The original trial records from the Logroño tribunal — more than five thousand folios of testimony, interrogation, and evidence — had been destroyed in the early nineteenth century during the Napoleonic wars and the subsequent political upheavals that closed the Inquisition. What survived was the correspondence between the local tribunal and the Supreme Council in Madrid, held in the Archivo Histórico Nacional: Salazar’s five reports, the letters of his co-judges arguing for prosecution, the responses and instructions of the Council, and a fragmentary selection of individual case papers.
Working from this correspondence, Henningsen reconstructed the narrative of the trials with a patience and precision that the book’s 607 pages only partially convey. He cross-referenced inquisitorial administrative records, episcopal archive material from Pamplona, notarial records, local parish registers, and the surviving portions of the testimony. He mapped the geography of the accusations, tracing the spread of the panic from Zugarramurdi outward through the valley systems of northern Navarre with the attention of an epidemiologist tracking a disease. He established the chronology of the panic, identifying the specific mechanisms by which accusations spread from village to village: a preacher’s sermon, a printed account of the auto de fé, a child who had heard an adult’s account of the sabbath and incorporated it into a vivid dream.
The book is both a narrative history and a methodological argument. The narrative is gripping by any standard: the opening of the panic in Zugarramurdi, the flight of refugees from the French trials, the gathering of the accused in Logroño, the auto de fé and its terrible processional pageantry, Salazar’s solitary journey through the countryside, his systematic demolition of the confessional evidence, and the eventual victory of his position in the 1614 Instructions that effectively ended witch prosecution under Spanish inquisitorial jurisdiction. The cast of characters is vivid: the three judges in permanent disagreement, Becerra and Valle pressing for prosecution and Salazar pressing for caution, their dispute reaching the Supreme Council with both sides accusing the other of bad faith; the Bishop of Pamplona, Antonio Venegas de Figueroa, who had independently reached similar conclusions to Salazar’s and whose letters are among the most incisive documents in the archive; de Lancre burning witches on the French side of the border while Salazar was releasing them on the Spanish side.
The methodological argument is equally important for anyone who wants to understand what inquisitorial archives are and what they can tell you. Henningsen showed that the archives must be read against the grain: that a confession in an inquisitorial record is not evidence of the event confessed but evidence of a process of interrogation, and that the value of the document lies in what it reveals about that process rather than in any claim about the external reality to which it purports to refer. This is now a commonplace of early modern historiography, but it was not in 1980, and The Witches’ Advocate was one of the texts that made it commonplace.
The destruction of the original trial records is also part of the argument. Henningsen noted with careful restraint that the documents Salazar had compiled — the most systematic empirical investigation of witchcraft claims in early modern European history, resulting in the exoneration of thousands and the prevention of executions that would otherwise certainly have occurred — no longer exist in their original form, while the demonological writings of de Lancre, who burned between sixty and eighty people in four months on the French side of the same border, were published and preserved and went through multiple editions. The archive does not favour the sceptic.
What the trials reveal about Basque belief
A question that runs through The Witches’ Advocate but is never fully resolved — and that connects this entry to the library’s first, on Barandiaran’s Mitologia Vasca — is what the trial testimonies reveal about actual Basque popular belief, as distinct from what they reveal about inquisitorial procedure.
Henningsen was careful about this distinction, but some of his contemporaries were less so. The confessions contain details that correspond closely to elements of the Basque mythological tradition: the goat-form deity (Akerbeltz, the black goat, appears in Barandiaran’s materials as a figure of folk belief quite independent of any Christian demonology), the gathering in the meadow (akelarre — the word for the sabbath field, which in Basque means literally the field of the goat), the association of caves with nocturnal assemblies, the role of women in the transmission of magical knowledge. These correspondences have sometimes been used to argue that the trials were, beneath the demonological overlay, documenting a genuine residual pagan tradition — an argument advanced most enthusiastically by Margaret Murray in the early twentieth century and largely rejected by subsequent scholarship.
The consensus position, which Henningsen’s work supports, is more cautious. The folk beliefs are real — Barandiaran’s fieldwork collected them in their living form, entirely independent of the trial records, and they are clearly old and specifically Basque in character. The inquisitorial category of the organised devil-worshipping witch sect is not real — it is an overlay applied to these folk beliefs by interrogators working from a demonological framework imported from outside the region. The two things are not the same, and conflating them produces a confused reading of both.
The sorginak of Basque tradition — the attendants of Mari, the female figures associated with the transmission of healing and magical knowledge — are a pre-Christian category, embedded in the mythological system Barandiaran documented, that has no essential connection to the Satanic sabbath of the demonologists. That the inquisitors used Basque words (sorgin, akelarre) for the categories they were imposing does not mean the categories were indigenous. It means the interrogators were working through interpreters who supplied local vocabulary for concepts that the local population did not, in most cases, actually hold.
The trials are therefore useful to the student of Basque religion and culture primarily as a negative document: they show what happened when an external framework of belief about witchcraft was applied with force to a population whose actual beliefs were different, and they show how that application generated the very thing it purported to find. Salazar’s conclusion is the most reliable guide to what the evidence means: there were neither witches nor bewitched until they were written and talked about.
The cultural life of Zugarramurdi
The village today has embraced its history with an equanimity that is partly commercial and partly, it seems, genuine. The population is small — well under three hundred at the last count. The cave system known as the Sorginen Leizea (Cave of the Witches), where the Orabidea river once ran through a natural tunnel and where the alleged sabbaths were said to have been held, is now a major tourist site and draws tens of thousands of visitors a year. The cave is genuinely extraordinary, independent of any association with the trials: a vast natural cathedral of limestone, its walls wet with the mountain’s seepage, its floor the meadow where in summer the solstice festival is held.
The Witch Museum, opened in 2007, addresses the trials with more historical honesty than many such institutions. Its permanent exhibition covers the demonological framework of the period, the social conditions that produced the panic — the absence of men at sea, the tensions of the border zone, the importation of fear from the French side — and the figure of Salazar as a voice of institutional reason. It does not romanticise the trials as evidence of genuine paganism or frame them as a feminist narrative of women resisting patriarchal power, though both of those interpretive frameworks have their advocates elsewhere. It presents the events as what they were: a catastrophe rooted in institutional cruelty and ideological delusion, brought to an end by one man’s systematic insistence on evidence.
Every year at the summer solstice, a festival called Akelarre Eguna fills the cave with bonfires, lamb roasted on stakes, music and dancing. The irony is deliberate: the very word the inquisitors used for the diabolical sabbath has been reclaimed as the name of a party. The lamb on the stake is a more literal version of the same reclamation. The darkness of the cave and the light of the bonfire are doing the same work as Salazar’s archive — reversing the procedure that manufactured terror from folk tradition, restoring the tradition to itself.
How to read it
The Witches’ Advocate is 607 pages long and repays close reading throughout, but it is not an equally compelling read throughout. The chapters that cover the actual events — the spread of the panic, the auto de fé, Salazar’s journey — are narrative history of the highest order, told with the attention to human detail that distinguishes the best archival scholarship from mere antiquarianism. The chapters that deal with the administrative structure of the Inquisition, the procedural dispute between the three judges, and the technical analysis of the confession evidence are denser and require more sustained attention, but they are not less important: the procedural material is where Henningsen’s methodological argument lives, and skipping it produces a less useful reading.
The recommended approach is to read Part One — the narrative of the trials from the French outbreak through the auto de fé — as a continuous sequence, then to read Salazar’s reports in summary (Henningsen provides extensive quotation) before reading the analysis of the confession evidence. The 2004 companion volume, The Salazar Documents, edits the key papers in bilingual Spanish-English form and is invaluable for readers who want to engage with the primary sources directly; it is available from Brill and at specialist libraries.
Reading The Witches’ Advocate alongside Barandiaran’s Mitologia Vasca — the library’s first entry — produces a specifically useful double perspective. Barandiaran shows you the Basque mythological tradition in its living form, collected from people who held it as part of their ordinary understanding of the world. Henningsen shows you what happened when that tradition was intercepted by an external framework of accusation. Together, the two books let you hold both the tradition and its persecution in the same field of vision, and to see clearly that they are not the same thing.
Access and further reading
The Witches’ Advocate: Basque Witchcraft and the Spanish Inquisition, 1609–1614 was published by the University of Nevada Press, Reno, in 1980. The University of Nevada’s Center for Basque Studies has been the primary anglophone publisher of Basque studies scholarship, and the book remains in print in their catalogue (ISBN 978-0874170061). It is available through major academic booksellers and on the secondary market.
The Salazar Documents: Inquisitor Alonso de Salazar Frías and Others on the Basque Witch Persecution, edited and translated by Henningsen, was published by Brill in 2004 (ISBN 978-9004131866). It contains the key primary documents in bilingual format with substantial introduction and commentary, and is the essential companion to the main study.
The Archivo Histórico Nacional in Madrid, which holds the surviving inquisitorial papers, has made portions of its collection digitally accessible through PARES (pares.mcu.es), the Spanish archival portal. Inquisitorial documents are searchable by tribunal location and date; the Logroño tribunal papers are partially available. The navigation requires patience and some Spanish palaeographical skill.
The Witch Museum (Museo de las Brujas) in Zugarramurdi has a website at museobrujas.navarra.es with visiting information. The museum is open year-round and admission is modest. The cave system is separately accessible at the edge of the village, managed by the local municipality.
For a scholarly overview of the broader European context, Brian Levack’s The Witch-Hunt in Early Modern Europe (4th edition, Routledge, 2015) is the standard reference, and places the Basque case within the continental pattern. Gerhild Scholz Williams’s English translation of de Lancre’s Tableau de l’inconstance des mauvais anges et démons (Arizona Studies in the Middle Ages and the Renaissance, 2006) makes the key French source accessible for the first time in English.
Previous in Stream Three: Del sentimiento trágico de la vida — Miguel de Unamuno, 1913 Next in Stream Three: Oteiza’s Sculpture — Jorge Oteiza, 1950s–1959