We all know that church authorities often had writings burned if they were judged heretical. The procedure seems so simple as hardly to be worth detailed scrutiny. Den Irrtum liquidierien proves this assumption wrong. A close re-examination of more than 220 documented cases of book-burning, from the late fourth to the early sixteenth century, from all parts of medieval Europe, reveals the ritual as a focus of the deepest tensions and ambiguities in the Church of that period.

The most salient tension concerns law. Recorders of persecutions, then and now, have sometimes remarked insouciantly that so-and-so was ‘burned with his books’. Intricately, the author shows that this did not happen, or not until secular justice took over heresy when the Middle Ages were over. Nor, he explains, could it happen. An ecclesiastical court was entitled to burn as many books as it saw fit. But it could not burn people.The two procedures belonged to different jurisdictions. They used different formulae and ritual, enacted by different people, in different places and often on different days – this last, not least, because book-burning called for maximum publicity, best found on holy days on which judicial killing was forbidden. Book-burning slotted easily into the age-old ritual of public penance. A penitent's throwing of his own books onto the fire was meant to be his efficacious sign of a change of heart, in a ritual which became known in some places by the Portuguese term auto da fé. This ritual had to be conducted ‘coram clero et populo’, words never used about the burning of a person, where the only cleric regularly present was there to hear possible last-minute confessions. There were minor variations of detail, but the essential distinction held. The tension here was between theory and its ritual on the one hand, practice on the other. The ‘secular arm’, to which a stubborn heretic was ‘relaxed’, was urged to show mercy and avoid a sentence of death or mutilation. But everyone knew that this was double-think, since the secular judge was not only empowered to burn the victim but was obliged to, on pain of excommunication. It was called showing ‘animadversio debita’, to use Gregory ix's euphemism.

The public character of book-burning gave the ritual a second tension. The prime purpose of the ritual was not to destroy a book, but to show that destruction was what it deserved. Authorities sometimes thought it prudent to keep a copy for themselves, especially with the more original and technical heterodoxies produced by academics, as distinct from time-honoured popular enthusiasms which any inquisitor could look up in a heresy dictionary. A secret archive may have existed for copies of banned writings as early as 745, when Pope Zachary silenced the Frankish enthusiast Adalbert. There was certainly one when John Hus's works were burned at Constance: it was copies that were burned, while the originals were locked away safely. The only books of which copies were never kept were magical ones, best annihilated altogether. The ritual was meant, primarily, to send a message. Only this can explain why some documents, namely excerpts or précis of a heretic's writings. were officially produced for no other purpose than to be publicly burned. It also explains, as nothing else can, why there were gradations in methods of destruction. Burning had a quasi-supernatural, purifying quality, and not all error deserved it. Writings thought merely scandalous could get a lesser sentence. A book by Gilbert de la Porrée was on trial in Rome in 1141, and John of Salisbury reported that ‘some said it should be burned, but it was decided that tearing-to-pieces was enough’. The same excuse was made for John Falkenburg's denunciations of the king of Poland in the 1420s.

A third kind of tension to be detected in the ritual touched the question of what was to be burned. A work self-consciously ‘heretical’ made the answer easy. In 867 the Byzantine patriarch Photius held a synod in defiance of the pope; so Hadrian ii asked the emperor to burn the proceedings publicly. In 1059 Berengar of Tours had to burn a book that he had written to defend a particular doctrine, on the eucharist, which he knew was disapproved of in high places. But what of a book which was partly right and partly wrong? Was wheat to be burned with tares? The dilemma grew more acute with scholasticism, which in its very method, a dialectic of pro and contra, preserved tares and wheat together; so that from the 1270s the device grew in vogue of condemning lists of propositions. But the dilemma also grew more acute with the mere rise in volume of books, of all sorts. Jean Gerson inherited this phase of the dilemma, and said that whether a book deserved burning or not depended on a balance: did the good in it outweigh the bad, or vice-versa? No rule ever invited more flexibility in application, depending on a judge's perceptions of danger. Johan von Wesel, a venerable scholastic with many innocuous treatises to his name, was condemned in Mainz in 1479 for some recent utterances on the nature of the Church. He had to watch, not just these, but his entire life's output, burn before his eyes. In humanistic Florence, eight years later, a commission called to hear charges raised against Pico della Mirandola's 900 ‘theses’, condemned only eleven of them, sparing 889, only to learn, weeks later, that Innocent viii condemned all 900. Well might Hus's friends argue, as he himself did in De libris hereticorum legendis, that if a book were burnable just for containing erroneous statements, then the entire intellectual foundation of the Church was at stake: the Decretum, the Sentences, the Old Testament, everything.

The Church's hidden ambiguities are thus reflected in the fortunes of the ritual most expressly designed to remove them – a paradox all too familiar. The three ambiguities just identified are far from the only ones the ritual brings to light. Its trajectory distantly mirrors that of book production itself – with a gap from 849 to 1050, a surge after 1200 and the glimmerings of general censorship (shutting the stable door before the horse escapes) after 1369. Seen close-to, the same trajectory also reflects particular histories. Jewish books, immune before 1232, shot to first place with the mass burnings of Talmuds between 1242 and 1319. The Dominican mission, meanwhile, is reflected in a development of the aspect of book-burning as a trial by ordeal: a legend told how St Dominic confounded Albigensians when his own, orthodox, volume leapt miraculously from flames which had just consumed theirs. (Seven of Werner's fifty-six excellent plates illustrate this miracle.) The tension between conscience and external authority is nowhere better illustrated than in a debate with his conscience of a fifteenth-century heretic, pressed to condemn and burn writings he believes to be true. And underlying the whole story is the fact that books were costly property and, as such, attracted a second enemy as the Italian Renaissance made them more costly still. An excursus on those ‘bonfires of vanities’ set off by Observant friars in fifteenth-century Italy, shows how books, of all sorts, were immune early in the century, but joined these expensive sacrifices in the course of it – just in time to take their revenge on the most famous scourge of vanities, Savonarola, in 1498. There is a lot more. Den Irrtum liquidieren is a heavy book, in every sense, but it is full of gold, a heavy metal.