RESPONDING TO VIOLENCE: LITURGY, AUTHORITY AND SACRED PLACES, c. 900–c. 1150

Abstract

The principle that church buildings constitute sacred spaces, set apart from the secular world and its laws, is one of the most enduring legacies of medieval Christianity in the present day. When and how church buildings came to be defined as sacred has consequently received a good deal of attention from modern scholars. What happened when that status was compromised, and ecclesiastical spaces were polluted by acts of violence, like the murder of Archbishop Thomas Becket in Canterbury Cathedral? This paper investigates the history of rites for the reconciliation of holy places violated by the shedding of blood, homicide or other public acts of ‘filthiness’ which followed instances such as Becket's murder. I first identify the late tenth and early eleventh centuries in England as crucial to the development of this rite, before asking why English bishops began to pay attention to rites of reconciliation in the years around 1000 ce. This paper thus offers a fresh perspective on current understandings of ecclesiastical responses to violence in these years, the history of which has long been dominated by monastic evidence from west Frankia and Flanders. At the same time, it reveals the potential of liturgical rites to offer new insights into medieval society.

The murder of the Archbishop of Canterbury, Thomas Becket, in his own cathedral on 29 December 1170 led to the closure of that church, and the suspension of all liturgical services there for almost a year.¹ Services only resumed some 357 days later on 21 December 1171, because the building itself needed to be ritually cleansed and restored from the violence which had occurred inside.² It took so long to reopen the church because advice about what to do in such circumstances was sought from the pope. The letter Alexander III (1159–81) wrote to his two legates on the subject is worth quoting:

We entrust to you how you are to reconcile the church of Canterbury: thus the sacrament of the original dedication ought not to be repeated, but just as it is the custom to do in the church of blessed Peter, blessed water should be splashed about.³

The papal legates consequently wrote to the prior and the community of Christ Church which served the cathedral, instructing them to call the abbots and bishops of their province to a solemn assembly so that they could ‘consecrate your church anew’.⁴ In a service conducted by Bartholomew, Bishop of Exeter, and attended by all the leading churchmen of southern England, the cathedral was, in the words of one later chronicler, ‘restored to its pristine state’ on the feast of Becket's namesake, St Thomas the Apostle.⁵

Becket's murder was shocking; it was also highly unusual.⁶ The texts cited above were all written close to the event and give the impression that what took place on 21 December 1171 was an unusual, even innovative, rite; advice had to be sought from the pope, who prescribed that a rite be conducted as at St Peter's in Rome. This impression of novelty is strengthened by the variation in the vocabulary used by different contemporary writers to describe the reconciliation service. The author of the Landsdowne Anonymous's account of the aftermath of Becket's murder described how the assembled clergy ‘reconciled the church of Canterbury (reconciliaverunt)’.⁷ Reconciliation is also the term used by Pope Alexander III in his instructions (cited above) to his legates that the Roman rite be followed. When it came time for them to pass these instructions on to the community of Christ Church, they chose instead to refer to the service not as a reconciliation but rather a chance to ‘consecrate (the church) anew (denuo ecclesiam vestram consecretis)’.⁸

The legates’ description of this rite as consecrating the cathedral ‘anew’ is reminiscent of contemporary debates in canon law about the circumstances in which a church might need to be reconsecrated. The influential twelfth-century compilation of canon law, Gratian's Decretum, included a canon that if the church has been violated, by either homicide or adultery, then it should be cleansed most diligently and consecrated anew.⁹ Gratian's account provided the foundation for the late thirteenth-century interpretation of this reconciliation rite by William Durandus, the southern French bishop of Mende, in his treatise on the symbolism of divine offices. There Durandus argued that it was not necessary to reconcile the church after all forms of bloodshed, such as nosebleeds or menstruation, but only after acts of violence.¹⁰ To date, scholarly considerations of this rite, to the extent they exist, have largely focused on the later Middle Ages. In fact, as this paper will suggest, there is good reason to suppose that the reconciliation service for holy sites polluted by violence was well known in England before Becket's murder, and indeed that the roots of this rite lie in the tenth century, if not earlier.¹¹ For a rite ‘for the reconciliation of altars or sacred places or cemeteries where blood has been spilt or homicide taken place or some other public filthiness (or superstition) carried out’ is already to be found in a bishop's service book associated with early twelfth-century Canterbury and now in Trinity College Dublin, which was probably made for Archbishop Anselm (1093–1109).¹² As this paper suggests, it is distinctly possible that that particular iteration of the rite was the basis for that used in Canterbury in 1171. Indeed, as will become clear, a rite for the reconciliation of violated holy places had already been a staple of southern English service books and church law for a century and a half before Anselm. All those involved in orchestrating the service after Becket's murder – the pope, his two legates, Bishop Bartholomew of Exeter, Prior Odo of Christ Church, Canterbury, and the assembled bishops and abbots – were already familiar with the rites for the reconciliation of those sacred places where blood has been spilt which are the subject of this paper.

The origins of this rite lie much earlier. Dedication rites to make a building into a church have deep roots in the early Church and were reserved to the bishop from the sixth century onwards, if not earlier.¹³ Dedication rites made ecclesiastical spaces sacred and set apart from the secular world, and consequently they have been the subject of a good deal of recent attention. Scholars have identified the period from the ninth to the eleventh century as a crucial time for the development of the church building (and its surroundings) as a separate, sacred space in the Latin Christian West.¹⁴ What has been much less noticed, however, is that the focus on developing a liturgy for making spaces holy was accompanied from at least the 790s by an interest in what to do if that space was violated by acts of violence.¹⁵ Yet some of the earliest surviving liturgical evidence for church dedication prayers is to be found in the late eighth-century Frankish world in three sacramentaries which also include prayers to reconcile a holy space which has been violated by blood being spilt.¹⁶

At around the same time the term ecclesia underwent a change in meaning. In late antiquity the word denoted a community of the faithful, but in the central Middle Ages it came to denote a particular physical site, enclosed and set apart from the secular world.¹⁷ Dominique Iogna-Prat found the origins of this spatial turn in the meaning of ecclesia in the writings of ninth-century Carolingian churchmen, arguing that the idea of the church as a separate, sacred space was first fully developed in this Frankish context; Miriam Czock has pushed this back into the eighth century.¹⁸ At the same time, the definition of ecclesia was expanded beyond the building in which worship took place to include cemeteries, the priest's house and land.¹⁹ This changed the understanding of church buildings so they became not just places reserved for worship, but also places set apart from the secular world, symbolising the community of the living and the dead but also offering a place of refuge from secular law.²⁰ This change was marked by the emergence in Frankia of liturgical rites for the consecration of churches from the seventh century, and their consolidation in the ninth century, and for the blessing of cemeteries in the tenth century.²¹ These were liturgical responses to a fundamental change in the ecclesiastical geography of the Latin West: the increase in local churches.²² Central to this development is the emergence and spread of parochial structures across the West, that is, the erection and maintenance of local churches funded by those living in the locality.²³ Alongside the proliferation in local church buildings came a shift in where the dead were buried. Between the seventh and twelfth centuries, the dead moved from their traditional Roman location outside the city walls to be situated next to the local church. Michel Lauwers termed this transformation inecclesiamento to capture the significance of this change in local landscapes.²⁴ Although the significance of the early medieval rite for the reconciliation of holy places has not been much noticed by modern researchers, either of Frankia or England, it has a good deal to tell us about how, when and why these changes came about.

In what follows I will start by investigating the evidence for the twelfth-century Canterbury rite available to Bishop Bartholomew of Exeter in 1171. I then consider how we should interpret this liturgical evidence by uncovering how the rite evolved between the time it was first recorded and the mid-twelfth century, with a focus on southern England. Tracing when and how this particular liturgical response to violent attacks on holy places came to be recorded in church law, I then ask how far these particular changes correspond with the developments in the liturgical record. In the final part, having established when and where the twelfth-century Canterbury rite first emerged, I will investigate the wider context in which it appeared, asking why this rite was recorded when it was, and what it reveals about the interaction of developments in ecclesiastical thought with wider social changes.

Quite which rite Bishop Bartholomew used to reconcile Canterbury Cathedral after Becket's murder cannot be known. The Magdalen Pontifical (Oxford, Magdalen College, Ms 226) is one of a group of pontifical manuscripts associated with Canterbury; it was copied in the mid-twelfth century and the rite it contains is therefore broadly proximate to the events of December 1171.²⁵ As the initial rubric makes clear, this rite assumes that the church building has previously been made holy through the rite of church dedication, but that that holiness has been contaminated by a violent act. It runs: ‘On the reconciliation of the altars and holy places or cemeteries where blood has been spilt or homicide committed or other filthiness perpetrated publicly.’²⁶ The rite in the Magdalen manuscript outlines the following stages:

• • It begins with bishop, clergy and people processing in front of the church, singing an antiphon from Psalm 76, ‘Thy way, O God, is in the holy place: who is the great God like our God’, followed by two prayers requesting the Lord put demons to flight and purify the minds of those present.

• • They then enter the church singing the antiphon ‘Peace be in this house and on all its inhabitants’, followed by prayer.

• • The text specifies that the clergy should then recite the same ‘litany as required in the dedication rite for an altar noted above’.²⁷

• • The bishop then makes the exorcism, blessing salt, ashes, water and wine as in the rite for the altar dedication, followed by a prayer requesting the reconciliation of this holy place or cemetery, and that it be consecrated or perpetually remain consecrated.²⁸ The aim is to avoid a repetition of the events which led up to this rite being performed.

• • Afterwards the bishop processes around the church or cemetery three times inside and three times outside, sprinkling blessed water and singing another antiphon: ‘Thou wilt sprinkle me, O Lord, with hyssop and I shall be cleansed.’²⁹ The prayers which follow request that the place be cleansed, restored and rebuilt: thus ‘this place will be purged of all pollution, sanctified and restored to its prior state, and reconciled as sacred’.³⁰

• • The bishop then goes around the church inside and to the doors and to the cemetery with incense whilst singing an antiphon and Psalm 130, and requests that the church (or cemetery) (hanc aecclesiam uel hoc cymiterium) be sanctified as it had first been sanctified (echoes here of the legates’ injunction to the cathedral community to consecrate the cathedral ‘anew’) and requests blessing and forgiveness of sins of both those who come to the oratory and those who are buried in the cemetery.³¹

• • The bishop comes back into the church singing a psalm and antiphon, ‘and the rest is as in the dedication of the church spelt out above.’³² He blesses the altar and commends the holy place to the Lord, requesting that all wickedness in the future be eliminated from this place. The prayers which accompany this stage again use the language of restoration and cleansing.

• • A solemn Mass follows immediately, in which the readings specified – the passage from the Book of Revelation (21:2–5a) about seeing a holy city, a new Jerusalem, coming down out of the heavens, and the parable from Luke (6:43–6) that no good tree bears bad fruit, nor a bad tree good fruit – allude to the importance of God's grace.³³ The prayers again continue the theme of cleansing from contamination and sanctifying place.³⁴

The language used throughout is that of disease and pollution, including spurcitia (filthiness), sordium (filth), culpa contagii (contagion of sin), maculatam (taint) and pollutio (pollution).³⁵ This is a state which can only be rectified by cleansing and exorcism, hence the continuous requests that the Lord purge and purify the contamination.³⁶ The Canterbury rite also deliberately echoes and repeats elements from the rite for the original dedication of the church, hence the repeated cross-references to various elements in the dedication rite: the recitation of the litany, the exorcism, and the blessing of the altar.³⁷

The rite is a very public and collective one. It begins outside the church with a procession, led by the bishop, but also involving the clergy and people: ‘First the bishop comes before his church with the clergy and people singing this antiphon with a sonorous voice.’³⁸ The prayer texts repeatedly emphasise the collective nature of the service. The deacon instructs those present to genuflect after the bishop utters the initial prayer and then rise up; the prayers include a request that purity ‘rise up in this place [since] this crowd of people agree to hold fast to their vows’. The rite ends by requesting the Lord to grant the request for restoration of purity to the crowd of people making it.³⁹ It is also a rite which emphasises the bishop's authority. Only he can preside over the rite and reconcile this space, just as church dedication rites were reserved to the bishop.⁴⁰ Thus through this collective rite, in which bishop, clergy and people all call on the Lord to cleanse this space and make it sacred again, the church (and/or cemetery) is restored as a space suitable for the community of the living and the dead.

The rite in the Magdalen Pontifical represents the culmination of a series of liturgical changes across the later eighth, ninth, tenth and eleventh centuries.⁴¹ The earliest evidence for a rite for the reconciliation of holy places is found in a group of three Gelasian sacramentaries compiled in France in the years around 800 ce.⁴² This rite consists of two prayers under the rubric ‘Reconciliation of the altar where killing has been carried out.’ The orator requests that God restore, cleanse and purify the altar which has been polluted by this act.⁴³ The final request in the second prayer, that the ‘pure simplicity of your church … that has been defiled after receiving grace return to its glory’, suggests that even at this early stage the altar already represented the whole church.⁴⁴ Rob Meens recently suggested that the appearance of this rite in late eighth-century Frankia should be linked to the emergence at the same time of a rite for the consecration of an altar;⁴⁵ and that both rites are a reflection, in the liturgy, of wider developments in Latin Christian thought and law over the course of the eighth century, which was when the idea that the church building should be regarded as sacred first became firmly established in ecclesiastical thought.⁴⁶

Both Carolingian rulers and churchmen promoted two versions of the Roman liturgy: first the eighth-century Gelasian and then, under Charlemagne, in an effort to return to what they believed was a more authoritative rendering, the Gregorian sacramentary.⁴⁷ The Gregorian text was soon found wanting and a Frankish supplement to it compiled which includes blessings for the consecration of an altar. Various later ninth-century manuscripts of the Supplemented Gregorian include further prayers and ordines for the dedication of a church.⁴⁸ Whilst at least one example of the Supplemented Gregorian sacramentary includes a Mass for the reconciliation of a violated church, none of the manuscripts in this particular group includes any indication of a separate rite for the reconciliation of an altar or church which has been contaminated by an act of violence.⁴⁹

Records of a full rite for the reconciliation of an altar or church contaminated by a violent act do not in fact reappear anywhere in the Latin West until the tenth century. When they do so, it is in England, and it is possible to trace the evolution of a specifically English rite across the next two centuries, from the mid-tenth to the mid-twelfth century. There are three main stages in the development of the written record of this rite across early English pontificals; these are collections of those rites reserved only to the bishop, such as clerical ordination as well as dedication of altars and churches.⁵⁰

The earliest two English examples of the rite are to be found in two manuscripts with connections to Worcester, now known as Claudius Pontifical i and the Egbert Pontifical.⁵¹ Claudius i is a pontifical copied c. 1000 at Worcester, but seemingly based on an early tenth-century copy compiled at Canterbury.⁵² The Egbert Pontifical was probably written for Oswald, bishop of Worcester (961–71).⁵³ The rite found in both these Worcester pontificals from the late tenth century is on the face of it a conservative text: the eighth-century Frankish Gelasian tradition of two prayers is accompanied by a ‘Mass for the reconciliation of the church’ drawn from the ninth-century Gregorian tradition.⁵⁴ The consequence was to extend the rite out from the altar to include other ‘holy places’, and in doing so move beyond the church building to incorporate also the reconciliation of cemeteries where blood had been spilt. We can follow this development in two places in the manuscript of the Egbert Pontifical. First the text of one of the original eighth-century Frankish Gelasian prayers, Deus cuius bonitas, has been amended to allow it to be used to reconcile a cemetery polluted in similar circumstances:

God whose goodness has neither beginning nor end hear our prayers that that of yours which is polluted may be cleansed, that which is neglected restored, that which is damaged, rebuilt, so that you may accept as reconciled the shelter of this church or cemetery and your altar should be cleansed of the infestation of devilish deceit.⁵⁵

Furthermore the noun chosen is not the usual coemeterium, but rather the archaic Frankish term poliandrum.⁵⁶ Secondly, the post-communion prayer in the Mass for the reconciliation of a church which follows was similarly amended later to request that the Lord ensure that this temple or this cemetery remains holy and protected from the defilements of barbarians and the unjust.⁵⁷ The ‘Egbert Pontifical’ has been linked on palaeographical grounds to Worcester c. 970, partly on the basis of the inclusion of texts also associated with Wulfstan, bishop of Worcester and archbishop of York (d. 1023), which include several archaic elements similar to the use of poliandrum here.⁵⁸

The second stage in the development of this rite is found in the pontifical made for Archbishop Dunstan (959–88), probably at Canterbury in the 960s.⁵⁹ This example includes for the first time an instructional rubric to the bishop on how to conduct the rite ‘for the reconciliation of holy places where blood has been spilt or killing perpetrated’: ‘the primate should sprinkle blessed water three times around and inside saying these prayers’.⁶⁰ The rite itself consists of the same two prayers first found in the late eighth-century Frankish Gelasian sacramentaries, followed by a new prayer in which the minister requests God to sanctify this church or cemetery which he has already made holy previously.⁶¹ The Latin here is the more usual cymiterium. Unlike the scribe of the Egbert Pontifical rite, that of the Dunstan Pontifical sought to include the cemetery as well as the church building in this reconciliation rite from the beginning. The Dunstan Pontifical is securely dated to Archbishop Dunstan's pontificate on grounds of script, contents (including a copy of the papal privilege granting him the pallium he collected from Rome on 21 September 960, in the same hand as the main text, and an ordo for consecrating an archbishop) and illustrations.⁶²

The third stage in the development of this rite is to be found in two later pontificals, both associated with Canterbury and copied c. 1000: the so-called Benedictional of Archbishop Robert and the Anderson Pontifical.⁶³ The Anderson Pontifical, in particular, bears various marks that it was read closely. These comprise firstly a number of marginal additions including a marginal gloss to the rite for reconciliation of holy places where blood has been spilt which extends the rite explicitly to include cemeteries violated in the same way.⁶⁴ The same annotator also made reference to the singing of litanies, and specific antiphons. The antiphons cited are those more commonly used in contemporary English rites for the dedication of a church and consecration of a cemetery.⁶⁵ These marginal additions were written in a very similar hand to that of the main text, so why were such chant texts added, and by whom? The most probable answer is a pragmatic solution: they reflect actual occasions on which this rite was staged. Similar additions were made to the rites for the reconciliation of altars and holy places and the consecration of cemeteries in a pontifical written at Canterbury or Winchester (now Cambridge, Corpus Christi College Ms 146) around the same time as Anderson.⁶⁶ Rites such as this were outside the normal temporal round of offices and Masses; the need to reconcile violated space was clearly not so routine that the community of Canterbury was able to undertake it without some planning. The cantor, the cleric tasked with staging a formal rite like this, is therefore the person most likely to have to consult the text ahead of time to consider how this rite should be enacted.⁶⁷ The later marginal notes of the antiphons to be sung – seemingly written in two stints – therefore probably signify that on at least two different occasions a cantor gave thought to how this rite should be delivered, and hint at the translation of written text into physical performance.

This evidence of revision, amplification and glossing suggests that these rites for the reconciliation of holy spaces where blood has been spilt were amended and updated across the second half of the tenth century and into the early eleventh century at Worcester and Canterbury. Each time, provision was made to expand the rite to incorporate the possibility that the sacred space which might be violated could include the cemetery as well as the church itself. The final major change in the evolution of the reconciliation rite in England came in the mid-eleventh century in a pontifical possibly made for the last pre-Norman archbishop, Stigand (1052–70).⁶⁸ This rite included new Continental elements from the Romano-German rite for the reconciliation of a violated church which seems to have developed independently, if contemporaneously, in tenth- and early eleventh-century east Frankia.⁶⁹ But the English rite departed from the Frankish Romano-German rite in one major way. The Frankish rite specified that, having blessed the water, salt, wine and ashes for exorcism, the bishop should go around the church three times on the inside, splashing the blessed water about, singing an antiphon and a psalm, in order to wash and purify the contaminated places.⁷⁰ The English version of this rite in Archbishop Stigand's mid-eleventh-century pontifical instead specifies that the bishop should go three times inside and three times outside the church or cemetery.⁷¹ This requirement – that the bishop go outside – reflects a peculiarity of the English consecration rites, which also specify that the bishop lead three processions around the exterior of the church, rather than the more common single occasion in the Frankish tradition.⁷² Indigenous English traditions clearly remained extremely powerful, even when Frankish elements were introduced into the reconciliation rites. This fourth version of the rite was in use for more than a century. It is almost identical to that in both the early twelfth-century pontifical made for Archbishop Anselm and the mid-twelfth-century Magdalen Pontifical.⁷³

This review of reconciliation rites suggests first that the cathedral communities of churchmen in southern England – not just at Canterbury but also at Exeter and Worcester – in copying rites for the reconciliation of sacred places extended the remit of those rites to include violation of cemeteries from the second half of the tenth century onwards. Second, these communities reworked this material again and again across the later tenth, eleventh and twelfth centuries. They incorporated new elements from the Frankish tradition and from the chant tradition; they even offered alternative singular and plural readings of sacred places – oratory/oratories, church/churches, cemetery/cemeteries – in the prayers in Cambridge Corpus Christi College MS 44; this was presumably to help readers understand the importance of correct grammar in prayer rather than for practical use, as similar changes were made to the subsequent rite for blessing the baptistery to allow for the consecration of multiple baptisteries at once, a situation which seems highly unlikely, not to say unfeasible.⁷⁴ Whatever the reason for these revisions, they are all testimony to an ongoing interest in developing the rite.

Their interest in the liturgy is echoed in church law. The medieval English church inherited a rich legacy of secular as well as earlier canon laws and penitentials.⁷⁵ The view that a place or building should be regarded as holy has deep roots in Roman tradition as well as early Christian doctrine and practice. In the Roman empire, the terms dedicatio and consecratio denoted a formal juridical procedure by which a site – a temple or a grave – was removed from secular use and became subject to divine law.⁷⁶ This idea that certain sites should be set apart as divine was carried over into early Christianity. The penitential associated with Theodore of Canterbury (d. 690) includes provision for sanctifying a church and cleansing it if a pagan is buried there.⁷⁷ At the same time, Roman law upheld the idea that such places should be inviolate; it condemned the violation of graves, for example, as a capital offence, and this teaching was taken over into medieval church law.⁷⁸ The contested nature of early Christianity also generated questions about the circumstances in which a church building might need to be reconsecrated. What happened if the building was destroyed by fire? What if the church had previously been in the hands of heretics? Was the original consecration valid? Should it be reconsecrated if the church was moved?⁷⁹

Two specific problems inform a pair of canons which are first recorded in the mid-tenth century. If the altar had been moved should the church be reconsecrated? If the church had been violated by killing or an adulterous act should be it reconsecrated? The answer in both cases was yes. The earliest surviving record of this duo is from mid-tenth-century southern Germany; they first appear in England half a century later in the early eleventh-century records of a Worcester canon law collection, seemingly made for Archbishop Wulfstan.⁸⁰ Despite their apparent novelty, they represent developments from earlier traditions. Scholars of central medieval dedication rites have shown how they evolved out of earlier customs which separated dedication rites for the placing of relics in altars from those for the dedication of a church. A canon on the need for reconsecration when an altar has been moved is therefore not surprising; indeed the issue was raised in Theodore's Penitential and remained a live issue in canonical circles four centuries later as is testified by letters written by Ivo of Chartres (d. 1115) and Archbishop Anselm (d. 1109) about whether a church should be reconsecrated when an altar was moved, or only when it was removed.⁸¹ Similarly, reconciliation rites for altars and holy spaces polluted by bloodshed and homicide survive, as we have seen, from the eighth century and are echoed in the provisions made in royal and ecclesiastical law in England and Frankia for punishing anyone killing someone in church from the ninth century onwards.⁸² What seems more novel is the reference to the need to consecrate anew a holy place after an adulterous act had been committed there; whilst this canon entered church law, appearing in two of the most popular collections in eleventh- and twelfth-century Europe – Burchard's Decretum (c. 1020) and Gratian's Decretum (c. 1140) – the focus of the liturgical rites remained on pollution from bloodshed rather than sexual acts.⁸³ To return to these two canons: they suggest that it is only from the second half of the tenth century that churchmen, first in Germany, then in England, became interested in prescribing that churches which had been violated by sinful acts should be cleansed and consecrated anew. When combined, the evidence of the liturgical rites and canon law indicates that English bishops and their clergy at Canterbury and Worcester became particularly interested in the rite for the reconciliation of holy spaces which have been violated in the later tenth century, and in extending this rite to include cemeteries around the year 1000. In the last part of this paper I therefore wish to consider why English churchmen began to pay attention to this rite at this time and in this way.

Bishop murder was extremely rare, but churches, the courtyards (atria) in front of them and cemeteries next to them were often sites for violent and destructive activities in the tenth and eleventh centuries. For example, the Anglo-Saxon Chronicle records how in 1002 King Æthelred II ordered that ‘all Danish men who were in England [were] to be slain’; this is an event known to modern historians as the St Brice's Day Massacre.⁸⁴ What this meant in specific terms is made clear in the records of St Frideswide's church in Oxford. They include a charter issued by Æthelred in 1004 in favour of St Frideswide's which records how as a consequence of his royal decree a group of Danes sought sanctuary in the church whereupon the townspeople burnt it down.⁸⁵ All the community's records were consequently destroyed, hence the need for a new charter. The discovery in 2008 in nearby St John's College of a mass grave of some thirty-four men who had died by violence sometime between 960 and 1020 has been linked by archaeologists to this massacre.⁸⁶ Their skeletons are tangible evidence of the reality of such an event. Unfortunately liturgical texts are not as amenable as bones to radiocarbon dating when it comes to dating them precisely.

Even if we cannot link rites to particular events, stories widely reported in miracle accounts, chronicles and letter collections from across England, northern France, the Low Countries and east Frankia point to how churches and their precincts were often the site of violent acts in this period. In 1060, monks from the Flemish monastery of Lobbes took the relics of their patron, St Ursmer, on a tour of Flanders in order to re-establish their authority over their scattered properties after the disruption caused by conflict between the emperor and the count of Flanders, and to raise funds for rebuilding their church.⁸⁷ In St-Omer they encountered two rival parties of armed men surrounding the church of Blaringhem:

The shields glinted red as the rays of the morning sun struck the courtyard (atrium) of the church as the steel of the weapons glistened, horses were snorting and whinnying, adding to the confusion.⁸⁸

On enquiry the monks discovered the cause of this tension: two knights had quarrelled and initially been reconciled in the presence of the local count, but one knight, still unhappy, killed the other and then sought sanctuary in Blaringhem church. The count and his men then returned intending to take the killer captive for challenging his authority, whilst the other group sought to defend him. The monks passed through the armed men, entered the church, where they celebrated Mass, and then processed outside, placing St Ursmer's relics in the midst of the crowd and provoking a mass declaration of peace in which, so the author of the Miracles of St Ursmer recorded, more than one hundred feuds were ended. Not all cases ended so peacefully. Around 1020 the sub-dean of Chartres was murdered in the atrium in front of the cathedral as he was coming to church one night; this became a cause célèbre as Bishop Fulbert of Chartres accused the Bishop of Senlis of having organised the assassination together with his mother and brother, because they thought the post should have gone to the brother.⁸⁹ This story reveals a good deal about the tensions over family control of cathedral offices, but both tales also suggest that the courtyards in front of churches were not only public spaces, but potentially dangerous ones.⁹⁰

Nor were the insides of churches safe places for members of the clergy and their households. In a case which shocked the north Italian clergy and the Ottonian court, Arduin, Margrave of Ivrea, killed Peter, Bishop of Vercelli, in his cathedral on 18 March 997, and burnt Peter's remains, incurring a sentence of excommunication.⁹¹ But, as we have already seen at Chartres, churchmen might also be responsible for perpetrating violence within the church itself. Some sixty-five years later, violence erupted as a consequence of a seating-plan dispute between two important Ottonian churchmen: Abbot Widerad of the wealthy Rhineland monastery of Fulda and Bishop Hezelo of the Saxon see of Hildesheim. The incident took place at the Whitsun royal court at Goslar. The bishop of Hildesheim was unhappy about an incident which had occurred the previous Christmas, in which his claim to sit closest to the leading churchman in the east Frankish Church, the archbishop of Mainz, at vespers had been dismissed in favour of the abbot of Fulda. In order to prevent the same seating plan being enacted again, the bishop therefore arranged for one of his counts, a certain Ekbert, to hide behind an altar with several knights. When it became clear that the bishop's claims were again being challenged by the abbot of Fulda, Ekbert and his men leapt out and began fighting the abbot's men with cudgels and drove them out of the church. The men of Fulda then gathered their weapons and returned, bursting into the church during vespers as the brethren were singing psalms in the choir, and started fighting, this time with swords, causing ‘rivers of blood’ to run through the church. Two men, one on each side, were killed in the ensuing melee: the standard-bearer of the abbot and one of the bishop's knights. The bishop's men were again victorious, and again the Fulda men regrouped occupying the churchyard drawn up in a battle line; only night led to an end to the stand-off. The next day an investigation was held under royal auspices. The abbot was blamed, being accused of having come prepared for a fight with a large force; he and his men were forced to pay punitive fines to the king to retain their offices, and these were so steep that they impoverished the abbey for years to come. This account comes from the monk Lampert of Hersfeld; he used the language of the Romano-German liturgical reconciliation rite, describing how the bishop ‘with so much bloodshed … avenged the injuries of the violated church (violatae aecclesiae)’, hinting at the ecclesiastical discourse which presumably followed these events.⁹²

These five cases are indicative of how in the tenth and eleventh centuries churches across the Latin West were the location for different sorts of violent clashes between armed men, attacks on unarmed men and attacks by angry mobs. Nor were such clashes confined to laymen; as both the Chartres and Goslar incidents suggest, churchmen themselves, even if they did not take part in the violence directly, could be accused of being its instigators. Stories such as these therefore provide one context for this ecclesiastical interest in recording and developing, if not creating, a liturgy for the reconciliation of such violated places. To put it simply, through administering such a rite, bishops were able to reassert their authority as arbiters of sacredness where it had been challenged by the violation of sacred space. The bishop's position as spiritual leader of the diocese was articulated from the moment he processed with the crowd in front of the church at the beginning of the rite, and reinforced as he physically cleansed and purified the boundaries of the holy spaces with holy water. Doing so helped him to reassert his own authority in setting the church and cemetery apart from lay society at a time of friction. Scholars have noted that other rites – the chrism Mass, public penance, church dedication rites, excommunication – which similarly articulated and promoted episcopal authority over the diocesan clergy and laity evolved around the same time (the later tenth and eleventh centuries) as that for the reconciliation of places polluted by violence, most noticeably in England, but also in north-east France, the Reich and Italy.⁹³

This interest in promoting episcopal authority over what constituted sacred space and how it should be restored following violent bloodshed in late tenth-century England fits with the model which Tom Lambert has recently proposed for the overall judicial landscape of post-Alfredian England:

We should, I believe, think of Anglo-Saxon England as a society in which violence … was regulated not by prohibitions imposed by a single central authority, but by a network of protections emanating from numerous different sources …⁹⁴

The tenth and eleventh centuries, in Lambert's view, witnessed an expansion of claims to royal protection, and at the same time, kings reserved the most heinous offences, including homicide within a church building, to their own jurisdiction. But this world of overlapping lordly protections, in which different authorities claimed jurisdiction over different spaces, and in which kings often asserted their jurisdiction in ecclesiastical spheres for the church's protection, provides a suitable context for, and explains, the seemingly precocious English interest in documenting – in law and liturgy – bishops’ control over reconsecrating spaces violated by violence.

To date, modern researchers’ focus has largely been on the ways in which west Frankish tenth- and eleventh-century churchmen responded to violent attacks. They have traced how monks responded to attacks on, and invasions of, church property both through the recitation of long, elaborate liturgical curses – the monastic clamor – and through an elaborate ritual in which they humiliated saints’ relics;⁹⁵ how bishops had recourse to excommunication of enemies in defence of their property;⁹⁶ and how bishops and monks came together in the Peace of God councils from the end of the tenth century to articulate and defend the defenceless – clergy and the poor, women and children – from violent attack on persons and property.⁹⁷ The rite for the reconciliation of sacred places where blood has been spilt therefore needs to be added to this list. The extent to which other ecclesiastical defensive actions are a result of, and reaction to, political transition and the weakness of royal authority – particularly because so many of them are concentrated in west Frankia – has been the subject of considerable debate.⁹⁸ The marked monastic bias of much of this evidence has been less discussed. Research on other forms of ecclesiastical defence against theft and violence relies on charter, legal and narrative evidence, or exceptional liturgical texts for the clamor and humiliatio, rather than texts routinely copied in liturgical collections. The rites for the reconciliation of sacred places where blood has been spilt help, therefore, to provide a useful corrective to this picture: for these rites are usually copied next to rites for the dedication of churches in sacramentaries and pontificals written across the Latin West.⁹⁹ They are thus normative and routine elements of these collections, but this should not mean they should be ignored. Rather their existence is a consequence of particular points in time in the legal and social history of the Latin medieval West. Additionally, the precocious evolution of this rite in late tenth-century southern English cathedrals offers a new perspective on a picture too long dominated by the experience of French monasticism.

Liturgy's formulaic and apparently timeless nature has for too long been regarded as an obstacle to its use as meaningful historical evidence, when in fact it is a strength as it allows, as we have seen, for the systematic comparison of changes in ecclesiastical thought and practice over both time and space.¹⁰⁰ Moreover we should not ignore a genre which constitutes the most plentiful category of texts to survive from England in the tenth and eleventh centuries.¹⁰¹ That entire codices, such as those considered above, represent only the visible surface of a large iceberg is indicated by the plentiful fragments preserved in later medieval and early modern book bindings and record wrappings which are still coming to light. But in a world in which parchment was expensive, and writing highly skilled, the decision to record any rite was always a self-conscious one. Asking why, in what was still a predominantly oral world, any rite was written down is an obvious, but curiously neglected, question. In a society which valued memory, these written records were not simple scripts for conducting worship, but instead served different and overlapping purposes: as instruments for meditative reflection on a theme, for instruction, and for the articulation of claims to authority.¹⁰² Tracing how particular rites such as this one were recorded and revised by anonymous copyists allows modern scholars to trace the evolution of ideas and the development of intellectual culture for the many known and unknown churchmen whose daily lives were shaped by the liturgy.

But let us return to where we began, with the reconciliation of Canterbury Cathedral eleven months after Becket's murder. That rite was not a Roman import but rather the latest iteration of a much longer tradition of a rite in which bishops reasserted their authority when church buildings were desecrated. The churchmen of the southern English province would already have been familiar with some form of this rite. The reason why this service was so delayed was because papal guidance was sought about what should be done in the absence of an archbishop to administer the rite. The legates and Canterbury community asked advice not because they did not know what to do when a church was desecrated by a killing, but rather because they were unclear as to who should lead the rite in this particular case; this rite was so indelibly linked in their minds to the articulation of episcopal authority that they were unclear who should conduct it in Becket's place.