Medieval Archaeology, 53, 2009 WINNER OF THE 2009 MARTYN JOPE AWARD The Architectural Setting of the Mass in Early-medieval Ireland By TOMÁS Ó CARRAGÁIN1 SURVIVING CHURCHES AND DOCUMENTS are analysed for what they may reveal about the architectural context of the mass in early-medieval Ireland. This shows that there is no evidence to support the widely held view that the congregation stood outside. Instead, the variable but relatively small size of these churches expresses the fact that they served smaller and more diverse communities than their high-medieval successors. The altars in large episco- pal and/or monastic churches seem positioned further west than those in relatively small, pastoral churches. In part, this was probably to facilitate relatively complex eucharistic liturgies. Externally defined chancels appear for the first time in the late 11th century ad in response to an increased emphasis on the real presence of Christ in the Eucharist. Signifi- cantly, they occur at a handful of important sites whose clerics and patrons were in direct contact with Lanfranc of Canterbury, a key exponent of this doctrine. About 180 pre-Romanesque churches survive in Ireland. One-fifth of these are drystone churches with corbelled roofs of Gallarus-type, some of which are as early as the 8th century. The rest are of mortared stone and date mainly from about ad 900 to 1130. They are a remarkably uniform group: all 180 are unicameral with short proportions (1:1.56 on average), a single W doorway and almost invariably just two small windows, one in the E wall and one in the south. The literature so commonly cites well-preserved examples at minor sites like Gallarus (Co Kerry) (Fig 1) that they are seen as typical, while larger churches at more important sites, which often survive only as vestiges incorporated into parish churches, tend to be overlooked. As a result, it is widely believed that Irish churches were too small to have been congregational.2 One of the main aims of this paper is to counter this view. In fact these churches vary greatly in size from tiny structures that could house no more than a handful of people to cathedrals 200 sq m or more internally. The average size of the principal church at sites that went on to be parish centres is 60 sq m and only a quarter of these are less than 40 sq m. This compares well with the naves of contemporary 1 Archaeology Department, University College Cork, Ireland. t.ocarragain@ucc.ie 2 Macalister 1935, 247; Henry 1940, 25–7; Leask 1955, 59; O’Kelly 1958, 127; Sharpe 1995, 368–9; Hunwicke 2002; Laing 2006, 217; O’Keeffe 2006, 128. For other criticisms of this argument see Ó Carragáin 2006, 112–14 and Ryan 2007, 520–1. 119 © Society for Medieval Archaeology 2009 DOI: 10.1179/007660909X12457506806207 120 tomás ó carragáin fig 1 Irish sites mentioned in the text. Map by T Ó Carragáin.  architectural setting of the mass 121 proto-parish churches in England that, according to Morris, averaged 20–30 sq m in the 10th century and 60–80 sq m from the later 11th century.3 Nobody has systematically studied these buildings for what they may reveal about how people celebrated the mass. Furthermore, while a lot has been written about the few surviving Irish liturgical texts, and especially on the Gallican and Roman influences evident in them,4 the documentary evidence for the architectural context and spatial organisation of the eucharistic liturgy has been neglected. In this paper I aim to make good this deficiency by looking in turn at the archaeological and documentary evidence for the location of the congregation, the celebrant and the altar. I will also briefly consider how developments in eucharistic theology and liturgy are expressed in the architecture of a few important sites around 1100. LOCATING THE CONGREGATION The most recent exponent of the idea that early Irish churches were not congregational is J W Hunwicke. He believes that they were designed as sanc- tuaries in which the mass was celebrated while the congregation stood outside.5 His starting point is an incident in Adomnán’s Vita Columba in which the saint enters the church on Hinba, a daughter-house of Iona (Argyll and Bute), with four visiting saints ‘after the Gospel had been read’.6 He concludes that the saints conducted the liturgy of the Word outside with the congregation but left them to celebrate the Eucharist inside. They did this, he argues, because the church on Hinba was not designed to accommodate a sizeable congregation, being similar in size to the drystone oratories of peninsular Kerry (Fig 2). Following on from this he argues that the 8th-century Stowe Missal, the most complete early Irish liturgical book we have, was written with tiny church-sanctuaries like this in mind. This is very doubtful, however.7 There are two possible alternative explana- tions for the episode in Adomnán. The first is suggested by a feature of the Gallican eucharistic liturgy that was practiced across much of Western Europe outside of Rome before the Carolingians made the Roman liturgy standard across their empire. In the Gallican mass, after the liturgy of the Word some of the clerics (practice seems to have varied considerably) exited the church in order to re-enter it in solemn procession with the gifts: a practice analogous to the ‘Great Entrance’ of the Byzantine rite.8 It is widely accepted that the Irish 3 Morris 1989, 287. 4 For example Schneiders 1996; Ó Néill 2000. 5 Hunwicke 2002. 6 Adomnán, Life of Columba, Book III, chapter 17 in Sharpe 1995; also Sharpe 1995, 368–9. The full sentence is as follows: ‘He obeyed their command, and with them he entered the church as usual on the Lord’s day after the Gospel had been read’. This does not necessarily mean that it was usual for him to enter the church after the gospel, as Hunwicke (2002, 1) implies. In fact we know that this was not usual on Iona where all of the mass was celebrated by the whole community inside (below). ‘Usual’ may simply refer to the celebration of mass on Sunday, rather than the practice of entering the church after the Gospel. Note that in Book III, chapter 17 we are told that ‘mass was being celebrated as usual on Sunday’. 7 He further suggests that the washing of the utensils and priests hands took place outside ‘because there would be little room. . .inside’ (Hunwicke 2002, 6). However an ablution drain for this purpose has been excavated in one small Irish church: Caherlehillan (Co Kerry) (Sheehan forthcoming). 8 Smyth 2003, 197–9; on the Byzantine Great Entrance see Taft 1978. 122 tomás ó carragáin fig 2 Drystone church at Templecashel (Co Kerry). Such churches generally occur at minor sites, accommodating small communities, often no more than a single family or a handful of monks. Photograph by T Ó Carragáin. liturgy had a strong Gallican flavour. It is therefore possible that this incident in Adomnán is evidence that a version of this ritual was practiced on the island of Hinba. The second possibility relates to the fact that Hinba incorporated a penitential colony.9 We know that in Ireland (unlike elsewhere in Europe) peni- tents were prohibited from partaking of the Eucharist for the full duration of their penance.10 At Hinba, then, there was a particular reason why part of its community had to be excluded from the Eucharist.11 In contrast, Adomnán makes clear that on Iona, which allowed no penitents,12 the whole community celebrated the mass inside the church.13 Even if neither of these alternative explanations is accepted, this single incident would not provide a sound basis on which to build a general model about the function of early Irish churches, as Hunwicke attempts to do. 9 Adomnán, Life of Columba, Book I, chapter 21 in Sharpe 1995. 10 Adomnán, Life of Columba, Book II, chapter 39 in Sharpe 1995; Bieler 1963, 278; Herren and Brown 2002, 124–5. We are told that the four saints went into the church with him but this does not necessarily mean the monks did not go in also, leaving the penitents outside. In another eucharistic incident in Book I, chapter 44 the only people mentioned within the church are Columba and a visiting bishop Cronán; the presence of the brethren is taken for granted until the end of the chapter. 11 Adomnán, Life of Columba, Book I, chapter 21 in Sharpe 1995. 12 Adomnán, Life of Columba, Book III, chapters 22, 27 and 30 in Sharpe 1995. 13 Adomnán, Life of Columba, Book I, chapters 22, 37 and 44, Book II, chapters 42 and 45 and Book III, chapter 23 in Sharpe 1995. This was also normal practice at other well-documented sites like Kildare (Co Kildare) (Connolly and Picard 1987, 26–7) and Armagh (Co Armagh) (Bieler 1979, 185–6) where even penitents stood inside the church though they did not partake of communion.  architectural setting of the mass 123 According to Hunwicke, ‘the most secure evidence’ in support of his hypothesis is the fact that ‘so much music is provided to ‘cover’ the Communion in the Stowe Missal that even the largest Irish church . . . could not conceivably have contained more than a tiny percentage of those communicating’.14 One might have argued this in a late-medieval context in which ‘taking communion’, which now comprised a pre-cut disc unaccompanied by wine, was seen as the consequence of the consecration, which had become the most sacred moment in the mass. However, Thomas O’Loughlin has shown that when the Stowe Missal was in use the mass was still conceived of as a communal meal in which the dynamic process of moving, eating, drinking, tidying up, cleaning and reflec- tion afterwards was ‘the high point of the whole celebration’.15 The Eucharist comprised a single loaf divided up according to the number of people present in a complex and time-consuming ritual, on a large paten like that from Derry- naflan (Co Tipperary). The wholeness of the loaf was ‘a symbol of the unity of the Church’ that was made up of the congregation and the wider Christian community.16 This was also expressed by the fact that they were standing under one roof or, as Cogitosus put it, ‘in one basilica, a large congregation of people of varying status, rank, sex and local origin . . . but one in spirit’.17 An even clearer expression of this idea is found in the 7th-century Second Synod of Patrick: Of the sacrifice. On the even of Easter, whether it is possible to carry it outside. It is not to be carried outside, but to be brought down to the faithful. What else signifies it that the Lamb is taken in one house, but that Christ is believed and communicated under one roof of faith?18 This idea of the church as a metaphor for the Christian community is central to an understanding of how these buildings were conceived. In contrast to Roman Temples, and indeed the Jerusalem Temple, the church was, of its essence, a space for the gathered people, for the ecclesia.19 It would be very surprising if Irish churchmen developed an architecture that ignored this fundamental symbol and in fact there is no evidence to suggest that they did. Apart from this theological objection, it is hard to imagine that the secular patrons of these sites were happy to stand outside when they were well aware that their peers abroad were protected from the elements. By the early 11th century the Stowe Missal was a prized relic of the impor- tant monastery of Lorrha (Co Tipperary) and there is no reason to doubt that it was always associated with this site.20 Lorrha features the third largest early- medieval church in the country (14.85 m by 8.75 m internally) and its wooden predecessor may not have been any smaller (Fig 3). The position of its S window suggests that the congregation was restricted to the western two-thirds of the church (further below), but even so it could comfortably have accommodated 14 Hunwicke 2002, 12. 15 O’Loughlin 2000, 143–5; also 2003, 9–10; Warren 1881, 134–6. 16 O’Loughlin 2004, 232; also de Lubac 1944, 19, 27. 17 Connolly and Picard 1987, 26. 18 Bieler 1963, 188–9, emphasis original. See also O’Donoghue 2006, 351 who recognised the significance of this passage. 19 Jungmann 1954, 47–8. 20 Ó Riain 1991, 294–5. 124 tomás ó carragáin fig 3 The early eleventh-century pre-Romanesque church at Lorrha is the third largest surviving example in Ireland. The Stowe Missal comes from this site and may occasionally have been used for the celebration of the eucharist in this church and its wooden predecessor(s). Note the original robbed-out W doorway. Photograph by T Ó Carragáin. a congregation of 200.21 In a context such as this the communion prayers included in the missal are not at all excessive. It is possible that the missal was originally intended for a priest serving a number of non-monastic churches, perhaps within the paruchia of Lorrha.22 We can speculate that these included community churches with sizeable congregations and much smaller family churches. As O’Loughlin has pointed out, in a situation like this the communion prayers could have been curtailed or prolonged as required.23 Thus, while it is not beyond the bounds of possibility that the Stowe Missal was used in churches as small as the drystone ones of peninsular Kerry, it cannot be used to support Hunwicke’s interpretation of them. The burial record negates his inter- pretation, for there is a close correlation between church size and cemetery size in early-medieval Ireland. For example, the small numbers interred around most drystone churches clearly indicate that these were for one or two families or, in the case of single-sex cemeteries, very small monastic communities.24 No 21 A formula of three people per sq m seems to be appropriate for maximum estimates (Parsons 1996; Ó Carragáin 2006, 114–15). The estimate of 200 suggested here is on the basis of a little over two people per sq m. 22 Meeder 2005. 23 O’Loughlin 2000, 143. 24 See for example Fanning 1981, Marshall and Walsh 2005 and Sheehan forthcoming.  architectural setting of the mass 125 doubt, on occasion it was expedient for part of the congregation to stand outside, but we will proceed with the assumption that early Irish churches usually accommodated the diverse communities they were built to serve. That is not to say that the congregation had full access to the interior of the church. We know from Cogitosus that the nave of the principal church in 7th-century Kildare was divided longitudinally with women on the left side and men on the right.25 Carol Neuman de Vegvar suggested that longitudinal divisions were not ‘part of early Irish church design in general’ because she felt Irish churches were too small for such subdivisions to be feasible.26 However, at just 4.5 m wide, the 8th-century church excavated at Whithorn (Dumfries and Galloway) was much narrower than the principal churches at most important Irish establishments and yet at least part of its nave was divided longitudinally into three sections (see Fig 9).27 Her other objections to the possibility of longi- tudinal divisions are also unconvincing. The fact that the doorways of most Irish churches were in the centre of the W wall is of little consequence for these putative partitions could have terminated a short distance east of the doorway; after all, western doorways are ubiquitous in the Continental basilicas that were subdivided in this way. Another piece of evidence she cites is a passage in the tract on church consecration in the Leabhar Breac in which two alphabets are written diagonally across the floor of the church from corner to corner.28 However, the same ritual also formed part of 9th-century Continental consecra- tion ceremonies despite the fact that the naves of most Carolingian churches were subdivided.29 Thus, we should be open to the possibility that, within some Irish churches, the laity were formally segregated according to sex and possibly according to status. In addition, a number of documentary sources indicate that a partition of organic material often separated them from the sanctuary. For example Cormac’s Glossary, which is largely 10th-century but with additions, defines cancella and its Irish equivalent crann-chaingel, as follows: Caincell, a cancella, ie a latticed partition. Crann-chaingel, ie a wooden partition, a beam-hurdle there ie a hurdle in the beam between laity and ecclesiastics (eter laocha 7 cléirc[h]u), after the likeness of the veil of the Temple. for cliath (ie hurdle, wattle panel) is its name with its fochra claraid (ie a partition or boards) ut dicitur cro-chaingel ie cro-cliath.30 Crann-chaingel is a compound of ‘wood’ and ‘cancella’, while cro-chaingel, the term used near the end of the entry, is a compound of ‘enclosure’ and ‘cancella’. This latter term is also found in the 9th-century Book of Armagh where a legal testament is made ‘between the crochaingel and the altar’ at Drumlease 25 Connolly and Picard 1987, 25–6. 26 Neuman de Vegvar 2003, 160. 27 Hill 1997, 150. 28 Stokes 1901, 370–1. 29 Repsher 1998, 145–6. 30 O’Donovan 1868, 46–7; Meyer 1912, 31. The translation is largely that of O’Donovan with minor changes made by Etchingham (1999, 316) in his partial translation. Translations in DIL of words and phrases that O’Donovan did not translate are given in brackets. This is from the version of Cormac’s Glossary in the early 15th-century Yellow Book of Lecan. The transcriber adds that a crann-chaingel is ‘a cancella of lattices or hurdles’. 126 tomás ó carragáin (Co Leitrim).31 This particular spot, directly west of the altar, was clearly an especially sacred one probably because, as we will see below, it seems to be where the celebrant stood during the Eucharist. There is further evidence for its sanctity in the following passage from the 9th-century Rule of Tallaght, which also introduces yet another term relating to the sanctuary partition: It was not customary among them to pass between the altar and the clais tarsna which is in front of the altar, and if anyone so passes, he is held to have incurred a penance. They were unwilling to kill any creature whatever between the crann-saingeal (a variant spelling of crann-chaingel) and the altar, for by custom only the body of Christ and his blood might be sacrificed in that space.32 Clais tarsna is translatable as ‘crosswise groove’ and one suggestion is that it might be a groove for a movable (sliding?) partition providing access to the sanctuary.33 We should note that this text does not necessarily imply that monks were prohibited from the sanctuary as a whole, but simply from the sacred spot between the sanctuary partition and the altar. An 8th-century poem in the Hisperica Famina describes a church with an ‘extensive porticum’ formed from ‘an assembly of planks’.34 While porticus can mean anything from a side-chapel to an atrium, Brady has argued that in this case it probably refers to the sanctuary.35 In his famous description of the 7th-century church at Kildare, Cogitosus tells us that the N/S ‘board wall’ delimiting the sanctuary (sanctuarium) ‘is painted with pictures and covered with wall hangings, stretches width-wise in the E part of the church from one wall to the other. In it there are two doors, one at either end’.36 The rectangular buildings at both Ballygarran (Co Waterford) and Dunmisk (Co Tyrone) have a single posthole on their axes that might have been for a traverse sanctuary division, but other explanations are possible, and caution is advisable as neither of these structures is definitely a church (see Fig 5). The Kildare partition seems to have bisected the entire church from north to south, but we should bear in mind the possibility that in some of the other large churches a square enclosure fully surrounded the altar, in which case some of the congregation may have stood at the sides as well as in front of the sanctuary, a practice attested on the European mainland and further afield.37 Indeed, in Santa Maria Maggiore in Rome some of the laity was accommodated, not only to either side of the sanctuary and presbytery, but also behind them, as illustrated by an episode in the Life of Paschal I (ad 817–24) in which the pope is disturbed during the liturgy by women standing behind him.38 Given the relatively small size of Irish churches, it seems unlikely that space existed for 31 Bieler 1979, 172–3; Bhreathnach 2001, 118. 32 Gwynn 1927, 10–11. The translation of chrann-saingeal is that of Thomas 1971, 16. 33 Macalister 1935, 182; Neuman de Vegvar 2003, 160. Note that clais can also mean choir though the headword is clas rather than clais (DIL). 34 Herren 1974, 109. 35 Brady (1997, 330–2) reasons that that because the author only refers to one porticus it was probably the sanctuary, which would have been the most important and perhaps the only one. On the wide range of uses of the term in the early Middle Ages see Ó Carragáin 1999. 36 Connolly and Picard 1987, 25–6. 37 See Lara 1994, 218–220 who cites both European and North African evidence for this. See Duval 2005, 14 for a general discussion of the development of the screened-off sanctuary beginning in the 6th century. 38 De Blaauw 1994 vol 1, 350–5.  architectural setting of the mass 127 the laity behind the sanctuary. However, the possibility that screens entirely surrounded some Irish altars is perhaps supported by the use of the term cro- chaingel (chancel enclosure) instead of crann-chaingel (wooden chancel) in two texts quoted above. Waist-high, drystone enclosures of this sort surround the 10th- or 11th-century outdoor altar-leachta on the perimeter of the island of Inishmurray; it is possible that, like the leachta themselves, they are adaptations of features found within churches to an outdoor environment (Fig 4).39 While sanctuary screens may have varied considerably in form, most of the evidence from Western Europe suggests that they would not have been opaque like a modern Orthodox iconostasis.40 In fact, Taft has shown conclu- sively that even in Byzantium opaque cancelli were unknown prior to the 11th century: in the early Middle Ages ‘the chancel enclosure was intended to set off and reserve space, not to hide it’.41 In the West, notwithstanding the diversity of arrangements that existed, early-medieval congregations were usually able to fig 4 This probably 10th- or 11th-century outdoor altar and waist-high sanctuary enclosure on Inishmurray (Co Sligo) may be adaptations of features found within churches to an outdoor environment. Photograph by Jerry O’Sullivan. 39 On their likely date see O’Sullivan and Ó Carragáin 2008. 40 Contra Sharpe 1995, 368 and Hunwicke 2002, 3. 41 Taft 2006, 40–9. See also Bolman 2006 who comes to a similar conclusion about the early-medieval chancel screens of Christian Egypt. 128 tomás ó carragáin clearly see the celebrant and the sacrament and, more often than not, the sanctuary screen was only waist high.42 In light of this evidence, Hunwicke’s argument that drystone churches like Gallarus were conceived as versions of screened-off sanctuaries in large churches cannot be sustained. Notwithstanding the presence of cancelli, the short and wide early-Christian proportions of Irish churches, often coupled as we shall see with an altar posi- tioned quite far forward from the E wall, meant that the priest was relatively close to the congregation compared to his high-medieval counterpart, or even compared to his counterpart in the long, narrow churches of Anglo-Saxon England.43 It is also important to remember that the congregation was not stationary throughout the mass. According to Adomnán, lay people could ‘approach the altar and receive the sacrament’, at communion.44 It has been argued that in contemporary Rome the congregation remained in the nave and aisles and communion was distributed to them across low chancel screens.45 That this was sometimes the practice in Ireland also is suggested by the passage from the Synod of Patrick quoted above that specifies that the host should ‘be brought down to the faithful’. However, other Irish texts hint at a greater degree of lay mobility within the church. For example, Cogitosus implies that at Kildare ‘the faithful of the male sex,’ as well as ‘the abbess and her nuns and faithful widows,’ went into the sanctuary to ‘partake of the banquet of the body and blood of Christ’.46 This is also, perhaps, suggested by the Penitential of Finian which states that, after completion of penance, the layperson will once again ‘be received to communion . . . he shall be joined to the altar’.47 Before considering exactly where the altar was located, we should try to determine where the celebrant stood in relation to it. LOCATING THE CELEBRANT The position of the principal celebrant during the mass is even more difficult to determine archaeologically than that of the congregation, except in exceptional cases like St Mary’s in Winchester where wear-patterns on the floor indicate it.48 It is still a matter of considerable debate among archaeologists and historians of the liturgy alike. Nussbaum argued for a gradual but quite uniform shift in the West from celebration facing towards the congregation to celebration facing away from the congregation between the 6th and 9th centuries; but increasingly the validity of this model is being brought into question.49 Parsons 42 For example Klauser 1979, 66; Neuman de Vegvar 2003, 164. 43 Taylor (1979, 185–6) argues that the lengthening of church proportions during the Middle Ages contributed to the sense of separation between clergy and congregation. 44 Adomnán, Life of Columba, Book II, chapter 39 in Sharpe 1995; also Bieler 1963, 278. 45 Doig 2008, 93–4. 46 Connolly and Picard 1987, 26; see also Warren 1881, 136–9. It was decreed at the First Council of Braga (Portugal) in 561 that lay people should be excluded from the sanctuary. As Dodds (1990, 23; also Lara 1994, 215) shows, Iberia is unusual in the degree to which the altar and celebrant are physically and even visually removed form the congregation. She (1990, 25) brilliantly argues that this reflects a strategy on the part of the Church to emphasise the divinity of Christ in opposition to the Arianism of the Visigothic elite. No such motivation existed in Ireland. 47 Adomnán, Life of Columba, Book II, chapter 39 in Sharpe 1995; Bieler 1963, 86–7. 48 Biddle 1970. 49 Nussbaum 1965, 414–21; for criticism see Duval 2005, 8.  architectural setting of the mass 129 suggested that in England the shift did not take place until later still: that celebration versus populum was still the norm during the 11th century.50 However, this hinges on a description of a secondary altar at the W end of Canterbury Cathedral (Kent) at which the celebrant faced east across the altar towards the congregation. In fact, Continental parallels such as St Gall (Switzerland, c 830) and Mainz (Germany, c 978) suggest that the celebrant at Canterbury faced east whether he was at the western or eastern altar.51 The fact that he faced the congregation when celebrating at the western altar shows that a desire to separate celebrant from congregation did not motivate eastward celebration. Rather the main concern was that he face east (versus ad orientem) towards the sun: a symbol of Christ, who would ‘come like lightning from the East’ (Mathew 24:27).52 A far more convincing model is that developed by Cyrille Vogel. He showed that the principal of eastward celebration has its origins in Jewish and Roman practices and was firmly established in the early Christian period.53 In a signifi- cant proportion of important early-Christian churches, especially those in Rome, the main apse was at the west, which meant that the celebrant faced both east and towards the congregation.54 But in this arrangement the congregation itself faced west, so increasingly churches with eastern apses were built in which they naturally faced east, unlike the celebrant who now had to be inconsistent: he faced west when seated and had to walk around the altar to face east for the mass.55 Notwithstanding the early-medieval emphasis on the Eucharist as communal meal, there was also a strong sense that it was an oblation (offering) through which, rather than closing in on itself, the pilgrim Church led by the celebrant progressed a little further on its journey to God.56 This is not to say that versus populum celebration never took place, and we must allow for the possibility of local variation.57 Also, it is important to emphasise that in versus ad orientem celebration the priest faced the congregation for the liturgy of the Word and only turned east for the liturgy of the Eucharist itself. There is evidence that the other clerics seated in the presbytery area behind the altar may also have turned east at this point: for example, the 9th-century Gallican version of the Ordo Romanus Primus indicates that the bishop, even while standing at his cathedra, must turn eastward towards the wall to pray.58 We should not be dogmatic about the situation in Ireland but a number of 8th- and 9th-century documentary sources, and what little archaeological evidence we have, suggest that, as elsewhere in Europe, versus ad orientem celebration predominated. In the Old Irish tract on the Mass in the Stowe Missal of c 700 it is specified that before breaking the host the priest must first 50 Parsons 1996, 63. 51 Lara 1994, 215. Lara argues that celebration at a western altar is what is being depicted in the 9th-century ivory sacramentary cover from Trier (Germany) that shows versus populum celebration. 52 Vogel 1964, 7–10; Lara 1994, 216. 53 Vogel 1962; 1964. 54 Vogel 1964, 21. 55 Vogel 1964, 29; for 4th- and 5th-century examples see Gamber 1976, 67–71. 56 Jungmann 1954, 56–60. 57 In particular there is evidence for some versus populum celebration in Italy where the weight of early-Christian Roman tradition was particularly influential despite the fact that later churches had their high altars at the east. Caillet 2005, 140–3; also Duval 2005, 15. 58 Lara 1994, 215–17. 130 tomás ó carragáin remove a piece from its lower left-hand quadrant in order to recall the wounding of Christ’s side with a lance on Calvary. The tract goes on to state that ‘west- wards was Christ’s face on the cross, to wit, contra civitatem (ie against the city of Jerusalem), and eastwards was the face of Longinus; what to him was the left to Christ was the right’. This implies that the priest, who was re-enacting Longinus’ action, also faced east.59 A passage from the 8th-century Monastery of Tallaght also implies that both the celebrant and congregation faced east towards ‘the light of the sun’.60 Clearer still is the passage from the 9th-century Rule of Tallaght quoted above which specifies that ‘the body of Christ and his blood’ is sacrificed in the space immediately west of the altar, ‘between the altar and the chancel screen which is in front of the altar’. Further evidence comes from a remarkable annalistic entry of 755 about the murder of a bishop by a priest in the church of Kildare, presumably the same one that had been described so vividly by Cogitosus about 80 years earlier: Eutighern, a bishop, was killed by a priest at the altar of St Brigit, at Kildare, between the chancel screen and the altar; whence it arose that ever since a priest does not celebrate mass in the presence of a bishop at Kildare (AFM 755). When he was killed this unfortunate bishop was celebrating mass while standing on the W side of the altar, and so he must have been facing east. The only Irish archaeological evidence that might illuminate early-Christian practice is the 5th- or early 6th-century church at Caherlehillan (Co Kerry) (Sheehan forthcoming) (Fig 5). Its table altar was apparently freestanding but the fig 5 Plans of small organic churches with altars. 1. Sod-built church rebuilt on a number of occasions excavated by Liam de Paor on Inishcealtra. Associated with one phase was a stain interpreted as evidence for a freestanding altar. The line of small posts along its E side may have supported a proto-retable (after de Paor 1997, fig 31). 2. The tiny 5th- or early 6th-century church at Caherlehillan also had a freestanding table altar. The grey rectangle south of the church indicates the location of a corner-post shrine overlying a group of graves (after Sheehan forthcoming, fig 1). 3. A pair of postholes inside the possible church at Dunmisk might possibly represent a two-prop altar leaning against the E wall (after Ivens 1989, fig 22). Drawing by H Kavanagh and T Ó Carragáin. 59 Ó Carragáin 1988, 8–9. For the text see Stokes and Strachen 1903, 254. On its date see Ó Néill 2000, 204. 60 Gwynn and Purton 1911, 156. In the Leabhar Breac it is stated that the consecration of a church should begin at the east because ‘from the east rises the sun which is an appellation of Christ’ (Stokes 1901, 380–1).  architectural setting of the mass 131 mensa would have had to be very small indeed for a priest to celebrate mass from behind it facing west. The altar may have been freestanding in emulation of the layout of larger, more prestigious churches with presbyteries at the east. We are perhaps as likely to find a missionary period church with the altar at the west and the door at the east, like the Romano-British example at Silchester (Hampshire), as we are to find clear evidence for westward-facing celebrants. As discussed in the next section, there are a few stone altars against the E walls of some small Irish churches that were obviously designed for eastward celebra- tion. However, it seems that in most churches altars remained freestanding and so other archaeological indicators of the position of the celebrant must be sought. In Francia the presence of reliquaries and inhumations privilegées immediately east of the altar are helpful in this regard, but in Ireland principal relics were usually kept outside the principal church.61 In the case of one of the undated churches of organic materials excavated on Inishcealtra (Co Clare), a rectilinear stain indicating the position of the freestanding altar was delimited at the east by three postholes (Fig 5). One possible interpretation is that these were for supports for something like a retable, in which case the celebrant must have faced east. The classic retable — a large board usually decorated with panels rising from behind the altar — did not develop fully until the 12th century, but various forms of proto-retable occur from at least the 9th century.62 The presence of large altar crosses ‘through’ which the whole community would direct their prayers would also indicate eastward celebration. In mainland Europe there is documentary evidence for placing a cross on the altar from the 6th century and the practice becomes increasingly common in the 8th and 9th centuries: for example, crosses are present on all the subsidiary altars in the St Gall plan, where eastward celebration is also indicated by the fact that the square sanctuary enclosures around these altars leave no room for priests to stand behind them.63 An altar cross would have been appropriate in the context of the Stowe Missal tract in which the Mass is interpreted allegorically as a figure for the Passion to the extent that, as O’Loughlin puts it, the portion received by each communicant is conceived of as ‘a relic of the true cross’.64 A small bronze cross-arm depicting St Paul from Shanmullagh (Co Armagh) may have adorned a reliquary shrine in nearby Armagh, but the much larger 8th-century cross recently discovered at Tully Lough (Co Roscommon) was apparently not designed as a reliquary. It may have stood on or more likely behind an altar, possibly in the nearby Patrician church of Kilmore (Co Roscom- mon).65 At Drom West (Co Kerry) a sculpture of possibly of 8th- or 9th-century date apparently depicts an altar complete with five mensal crosslets, surmounted by a large ringed altar cross.66 Thus there is some archaeological evidence to 61 Ó Carragáin 2003b. 62 Caillet 2005, 143–5. 63 Caillet 2005, 144. 64 O’Loughlin 2003, 17. The presence of a cross on the altar is taken for granted in Gille of Limerick’s De Statu Ecclesiae (c 1111) (Fleming 2001, 160–1). 65 On the Shanmullagh cross-arm see Bourke 1993. On the Tully Lough cross see Kelly 2003. I am grateful to Griffin Murray for discussion on this point. 66 Ó Carragáin 2003a, 133 n 8. 132 tomás ó carragáin support the documentary evidence that the celebrant usually stood west of the altar and faced east. LOCATING THE ALTAR In early-Christian basilicas the altar was usually near the E end of the nave, while the apse was where the bishop and other clergy sat.67 Variations on this arrangement remained common throughout Europe until well into the 11th century. In Anglo-Saxon areas evidence for it exists at sites as diverse as Winchester (Hampshire), Whithorn and Raunds (Northamptonshire). The latter was an 11th-century local church in which the eastern cell probably functioned as a presbytery and/or secretarium rather than a chancel.68 On the Continent, the only exceptions to this rule are some small unicameral churches in which the altar stands against the E wall for convenience sake. In Ireland there are 11 extant stone altars in early-medieval churches, usually small ones, and all of them are positioned against the E wall; however, most of these are undatable and at least some are probably post-1100 additions. There is a possible wooden altar abutting the E wall of the putative church at Dunmisk (Fig 5), and another possible example was excavated at the E end of the little church at Ardwall Isle (Dumfries and Galloway), a site possibly established in the context of 6th- to 8th-century Irish influence in the area.69 In contrast excavations at Caherlehillan and Inishcealtra have shown that even very small churches in Ireland sometimes had freestanding altars (Fig 5).70 To date we lack excavation evidence for the position of the altar in a single important Irish church, and the documentary sources are also generally vague on this issue. The odds are against discovering the remains of an altar within a large church, which will inevitably be riddled with burials, for a range of sources not considered in detail here suggest that wooden altars were the norm in early-medieval Ireland.71 A handful of minor churches have produced other features, such as aumbries and ablution drains, which might help in determining the position of the altar, but in these small buildings there is usually not much doubt about its position. Unfortunately, no aumbries or ablution drains are known from larger churches in which the position of the altar is more difficult to determine. The only quantifiable variable that may be relevant to the issue is window position. As noted above, there is usually just one window in the E wall and one in the south. Artificial illumination was probably used but even so these churches must have remained quite dark. At the beginning of midnight office on Iona the church was illuminated only by lamps that individual monks had brought with them, but we do not learn whether these were then supplemented by light from permanent fixtures.72 Stone lamps from Labbamolaga (Co Cork), Inishvickillane (Co Kerry) and Temple Brecan on Aran (Co Galway) may have 67 Krautheimer 1986, 43. 68 Taylor 1973; Parsons 1986, 106; 1996, 63; Fernie 1983, 41; Blair 1996, 15; Hill 1997, fig 4.13. 69 Thomas 1967, 136–8, fig 26; 1971, 72–3, 177–80. 70 Sheehan forthcoming; de Paor 1997, fig 31. 71 Ó Carragáin forthcoming. 72 Adomnán, Life of Columba, Book III, chapter 23 in Sharpe 1995.  architectural setting of the mass 133 illuminated the early churches at these sites and, according to Griffin Murray, decorated metal strips found in excavations on Inishcealtra may be from an 11th- or 12th-century hanging candelabrum.73 Navigatio Brendani describes a church illuminated by seven lamps hung before its three altars,74 and at Kildare the shrines of Brigit and Conlaed were illuminated by ‘gold and silver chande- liers hanging from above’.75 While this latter church also had ‘many windows’, those building the stone churches decided to use natural light much more sparingly. There was no practical reason why Irish masons could not have built larger windows and more of them. In this regard they diverged significantly from their early-Christian models, for writers of that period tend to emphasise the quality of light within churches.76 It seems, however, that some of the northern successors of these buildings were not so well lit: the wooden church built by Clovis in c 504 in Strasbourg (France) apparently had just one window.77 In the case of the Irish churches it is almost as if those commissioning them were aiming to create an atmosphere that was as different as possible from the outdoors, or indeed the relatively flimsy domestic buildings of the day, perhaps to heighten the sense that one was entering a sacred, almost otherworldly space. The coldness of the stone walls and the dimly seen but massive beams of the roof above would have contributed to this, as would carefully positioned lighting, through which certain focal points were illuminated in the gloom. In good weather, the thin shafts of sunlight projected by the unnecessarily small windows were undoubtedly the strongest sources of light in the church; and it would make sense if this light fell on the most important focal point: the altar or perhaps the priest as he consecrated the host just west of the altar. While I cannot fully substantiate this hypothesis, Romanesque evidence strongly sup- ports it. As far as I am aware, we can determine the original positions of the windows and altar of just two Irish Romanesque churches: Cormac’s Chapel, Cashel (Co Tipperary) and St Caimin’s, Inishcealtra (Fig 6). Cormac’s Chapel dates to 1134 and was built largely by English masons, but in some respects it harks back to earlier architecture in Ireland so individuals who were conscious of this tradition must have had a direct input into its design.78 In plan the church proper comprises a nave, chancel and eastern altar niche. This space was sparingly lit; there was one window in the W wall above the level of the side wall doorways and a smaller one at an even higher level at the midpoint of each of the side walls of the chancel. The only two windows at the level of the people who used the church were positioned in each of the side walls of the altar niche. They are considerably larger than those in the chancel and were clearly positioned to illuminate the altar. Footings still evident within the niche suggest that it was a four-prop altar of stone. 73 The stone lamps from Labbamolaga and Inishvickillane belong to a group of cresset lamps for which a 10th- to 12th-century date range has been suggested (Moore 1984). It seems clear that they were not designed specifically as church furniture, however, for only a minority are associated with ecclesiastical sites. On the Inishcealtra strips see Murray (2007, I, 296). 74 O’Donoghue 1893, 138. 75 Connolly and Picard 1987, 25. The term used is coronae, but Connolly and Picard are almost certainly correct to translate it as chandeliers rather than crowns (Cormac Bourke pers comm). 76 McClendon 2005, 22. 77 Ahrens 2001, I, 78. 78 O’Keeffe 2003, 125. 134 tomás ó carragáin fig 6 Cormac’s Chapel (top) and St Caimin’s, Inishcealtra suggest that the easternmost sidewall windows in Irish Romanesque churches were usually positioned to illuminate the altar. Drawing by H Kavanagh and T Ó Carragáin.  architectural setting of the mass 135 St Caimin’s, Inishcealtra, comprises a pre-Romanesque nave and Roman- esque chancel. The pre-Romanesque church originally had just one S window located roughly midway along the wall, but a second one was added further east during the 12th century (further below). The mid-12th-century chancel was illuminated by one window in the S wall and almost certainly another at the midpoint of the E wall, though this does not survive. According to a report on the 1879 restoration of the church by the Board of Works, ‘the chancel walls had near disappeared . . . the stones of the altar were found and re-erected, as well as portions of the chancel walls and windows of same’.79 One might conclude from this that nothing of the original S window had survived in situ, in which case we would not be able to draw any conclusions from its present position. Most of the earlier accounts of the church do not mention or illustrate it,80 but fortunately Brash states that ‘the sill of a south window exists’ and his plan shows it aligned exactly on the altar as it still is after the restoration. We know that the altar is also in its original position for Brash states that a portion of it ‘exists in situ . . . having a passage of 15 inch [0.38 m] between it and the east wall’.81 It may be slightly later than the chancel itself but it is certainly Romanesque. Like most Romanesque churches, these two had more than one S window, but in both cases the easternmost one was positioned to illuminate the altar. Even where the original altar does not survive it is significant that Romanesque churches usually have a S window near the E end of the chancel for, as discussed below, we know that this is where altars were usually positioned in the 12th century.82 Here I will proceed on the assumption that there was already a close relationship between the position of the altar and that of the S window in pre-Romanesque churches. If this principal did not apply, then it is unlikely that we will ever get a clear understanding of the position of the altar in the major churches. Interesting patterns emerge from an analysis of the positions of the 40 or so pre-Romanesque S windows whose position we know. None are sufficiently close to the E wall to strongly suggest that the altar was placed against it; except for some in very small churches, altars seem to have been freestanding. However, the majority were about three-quarters of the way along the wall, leaving little room for a priests’ bench behind the altar and certainly not enough room for a proper presbytery. This arrangement was most common in modest churches that probably functioned primarily as pastoral rather than monastic centres: sites such as Agha (Co Carlow), Confey (Co Kildare), Agharra (Co Longfor) and Kilcummin (Co Mayo) (Figs 7–8).83 By contrast, in the principal churches at major complexes such as Inishcealtra, Kilmacduagh (Co Galway), 79 Quoted in Macalister 1916, 128. 80 For example Petrie 1845, 282 and Dunraven 1875–7, vol 2, 7. 81 Brash 1866, 12–15. 82 See Graves 2000, fig 16 for a similar analysis of the relationship between window position and altar position in a 13th-century English parish church. 83 The churches with S windows about three-quarters of the way from the west, or a little less, are Agha Phase 1, St John’s Kilmacduagh (two S windows), Kiltiernan (Co Galway), Church Island (Co Kerry), Confey, Kilfinny (Co Limerick), Kilrush (Co Limerick), Mungret 1 (Co Limerick) (two S windows), Agharra, Kilcum- min, St Declan’s Ardmore (Co Waterford) and Kilbarrymeaden (Co Waterford). The S window of St John’s Point (Co Down) is between two-thirds and three-quarters of the way from the west. 136 tomás ó carragáin fig 7 Kilcummin. The easterly position of its S window, and by implication its altar, is typical of relatively small churches that may have had a pastoral function. Photograph by T Ó Carragáin. Inishfallen (Co Kerry), Kilree (Co Kildare), Lorrha and Glendalough Cathedral (Co Wicklow), the S windows tended to be two-thirds or just half-way along the wall (Figs 3 and 8).84 This is despite the fact that, given their size, there would have been ample space for an eastern presbytery area if, as in the smaller churches, the altar had been placed three-quarters of the way from the W wall. How are we to interpret this striking pattern? It is possible that it reflects liturgical change, for I have shown elsewhere that large monastic churches tend to be slightly earlier than the smaller churches at less important sites.85 The case of Mungret (Co Limerick) supports this possibility. Its masonry style and lack of antae (pilaster-like projections of the side walls beyond the end walls) suggest it is later than most of the other large monastic churches, and its easternmost S window (unusually it has two) is further east than usual for such churches: about three-quarters of the way along the wall. However, other examples tell against the idea that this pattern is primarily chronological. The drystone church on 84 The churches with S windows about two thirds of the way from the west are Kilmacduagh 1, Tullaherin (Co Kilkenny), Inishfallen (Co Kerry), Clonkeen (Co Limerick), Lorrha and the Cathedral (two S windows) and St Mary’s at Glendalough. Churches with windows about half way along the S wall are Inishcealtra, Derry 2 (Co Down), St MacDara’s, Templemacduagh, Kildreelig (Co Kerry), Killabuonia (Co Kerry) and Kilree (Co Kilkenny). 85 Ó Carragáin 2005.  architectural setting of the mass 137 fig 8 Comparative plans of 14 churches restored to their likely early-medieval form, selected here because their original proportions and the positions of their original windows can be determined. These plans illustrate the contrast in window position between relatively minor churches (upper two rows) and the principal churches of relatively important sites (bottom three rows with the exception of Templemacduagh, which is quite a minor site). The plans omit later features and restore original features where necessary. For example, compare the multi-phase plan of St Caimin’s, Inishcealtra (Fig 6) with that here, which omits the Romanesque chancel, door and inserted S window and restores the original E window and door. The second S window in St John the Baptist’s, Kilmacduagh, may have illuminated a baptismal font. Drawing by H Kavanagh and T Ó Carragáin. Church Island and the shrine chapel of St Declan’s, Ardmore, may be 8th- and 9th-century respectively, and yet their S windows are very near their E walls. The only two late churches with windows midway along the S wall are St MacDara’s (Co Galway) and Templemacduagh (Co Galway). It may be significant that both are at island establishments: possibly they were monastic in character unlike many of their contemporaries on the mainland with windows near the east. Most of the later churches were built in the latter half of the 11th century and the early 12th century, a time when, elsewhere in Europe, altars were increasingly positioned at the E end of churches, where the presbytery area had once been (below). It is possible that the position of their altars is a reflection of this trend. However, most of them probably date to an early stage in this development and are at relatively modest sites possibly unaffected by it. Furthermore, the exceptions cited above, such as Church Island, hint that the position of their altars may be as much a product of longstanding functional 138 tomás ó carragáin differences as they are of chronology; primarily pastoral churches served by just one or two priests did not require a large presbytery area. The main concern of those designing these churches may have been to make as much of the interior as possible available to the congregation. THE SPACE EAST OF THE ALTAR In contrast, the more westerly position of S windows in the principal churches at major episcopal-monastic centres suggests that a smaller proportion of these buildings was set aside for the congregation. These large churches also tend to have slightly longer proportions than average. This may have helped to separate out the relatively diverse religious and lay groups that these churches were probably designed to accommodate. The range and character of these groups seems to have varied a good deal from site to site. In 7th-century Iona the principal church was primarily for monks so there may have been relatively few internal partitions, but at Armagh it accommodated ‘bishops, priests, ancho- rites and other religious’, while at Kildare it accommodated laity, nuns, monks, priests and bishops.86 Thus at many sites the principal church was not solely, or even primarily, monastic. It has been argued that in multiple church-groups in early-medieval Provence (France) there is often a dichotomy between the cathedral church on the one hand and the monastic church on the other.87 Aidan MacDonald has made the intriguing suggestion that there was a similar dichotomy in 10th- to 12th-century Clonmacnoise (Co Offaly) between the stone cathedral associated with bishops, and the dairthech (later known as Temple Kelly) just north of it, which he argues was associated with the monastic community.88 If he is correct, this seems to have been a rare if not unique arrangement in the Irish context. There is evidence that some major sites, possibly including Clonmacnoise, had separate small ‘parish’ churches for the lay com- munity; but even these sites may have allowed a lay congregation into the prin- cipal church on important feasts, as was common even in purely monastic churches on the Continent.89 Whatever their monastic component, these principal churches must have been designed to accommodate many more priests than the churches at less important pastoral sites. For example, the wooden church described in Hisperica Famina had ‘a holy altar in the centre, on which the assembled priests celebrate the Mass’.90 In early-medieval texts ‘centre’ can simply mean on the axis of a building so does not necessarily indicate a large area behind the altar,91 but this possibility is supported in this case by the fact that the likely sanctuary (porticum), that presumably accommodated the altar and the priests, is described as ‘extensive’ (above). The presence of a number of concelebrating priests performing relatively complex eucharistic liturgies is surely one of the reasons 86 Sharpe 1995; Bieler 1979, 187; Connolly and Picard 1987, 26–7. 87 Codou and Fixot 1996, 209. 88 MacDonald 2003, 129–30. This would be reminiscent of an église double complex but on the Continent these are pre-Carolingian in date and there is less of a size discrepancy between the two churches (Carre 1996). 89 For example Rabe 1995, 122–32. 90 Herren 1974, 109; also O’Donoghue 1893, 138. 91 Brady 1997, 329–30.  architectural setting of the mass 139 for the large space that we have apparently identified east of the altars of important churches. The question arises does this adequately account for the space or could it also have accommodated other groups and served other functions? We know from documentary sources that in early-medieval Rome large numbers of clerics — sometimes more than 50 — were accommodated behind the altars of the stational churches on important feast days. In addition, from the 5th century most churches in the city had a ‘lower choir’ west of the altar at the E end of the nave, which was defined by low stone screens. According to the Ordo Romanus Primus (§126), this was for the less important religious, including monks and some priests. In most cases, the laity was located to the west and to either side of this lower choir (above). It is argued that choir spaces west of the altar also became increasingly common in Francia and England from around 700 onwards.92 While it was usual for some clerics to be seated east of the altar until the later 11th century, in many churches these more westerly choir spaces accommodated the majority of religious (further below). The 11th-century text, Fís Adomnáin, provides us with strong evidence that some Viking-Age churches in Ireland had properly defined spaces for assemblies or choirs of religious. In it is a description of heaven in which choirs of angels are arrayed around God and ‘between every two choirs is a partition (crand caingil) of crystal, with splendid ornamentation of gold and silver on it’.93 Unfor- tunately, there is little textual evidence from the Viking Age to indicate where such choir spaces were located in relation to the altar. To explore this issue we have to return to the 7th-century description by Cogitosus of the church at Kildare. In this church the nuns were with the male and female laity in the nave until communion, but the archbishop’s ‘monastic chapter and those appointed to the sacred mysteries’ were in the sanctuary with him for the whole of the mass.94 The most convincing reconstruction of this church published to date is that by Neuman de Vegvar; however, she is probably incorrect to place the altar against the E wall.95 In this regard, Ralegh Radford’s reconstruction is better. Citing evidence from a range of 4th- to 6th-century churches abroad, he suggested a significant number of individuals were seated behind the altar with the bishop.96 The mid-8th-century annalistic reference to Cogitosus’ church quoted above would seem to support this. It states that the bishop offered mass ‘between the chancel screen and the altar’. This suggests a small space reserved for the celebrant, as the writer is unlikely to have chosen this phrase if the monastic chapter sat between the bishop and the chancel screen. Indeed, as noted above, the 9th-century Rule of Tallaght explicitly states that religious are not allowed to occupy the space where the Eucharist is consecrated between the altar and the chancel screen ‘which is in front of the altar’. These sources suggest 92 Gem 2005, 282. 93 Herbert and MacNamara 1989, §14. ‘Crand caingil dig lain eter cach dá claiss co cumtach derscaigtech dergóir 7 argit fair’ (Best and Bergin 1929, 70). See also DIL on clas from the Latin classis: an assembly or choir. 94 Connolly and Picard 1987, 25–6. 95 Neuman de Vegvar 2003, fig 1. 96 Radford 1977, 7. Note that some other aspects of Radford’s reconstruction of Cogitosus’ church are doubtful (Radford 1977, fig 1). 140 tomás ó carragáin that the only N/S-oriented partition in the church described by Cogitosus was a short distance west of the altar. If so, then the significant group of religious that was located east of this partition must also have been east of the altar. This is where we would expect to find ‘those appointed to the sacred mysteries’, but the possible presence of a ‘monastic chapter’ in this area requires more consideration. The phrase in Cogitosus is schola regularis. As Jean-Michel Picard has pointed out to me, this does not necessarily mean a choir of monks under an abbot; in fact, in this context it clearly denotes a chapter of clerics living according to a rule (regularis) under the authority of the archbishop. We must remember that Benedict of Aniane (France) was the first person to make a formal distinction between monks and canons in c 816–20.97 We do not know what proportion of the schola at Kildare were in priestly orders, but it is safe to assume that they were of lesser rank than the clerics seated behind the altar in the stational churches of Rome, many of whom were bishops. The Roman evidence would lead us to expect most canons to be located in a lower choir west of the altar. We can perhaps best explain this apparent difference in the rank of those located behind the altar as a by-product of the process of transposing a model for the internal layout of a basilica from major Continental centres to a modest establishment like Kildare. Cogitosus gives us little information about the people located on the right side of the nave, except that they were male. Assuming that some of them were religious and that these were positioned east of the laymen, then they occupied a position compa- rable to those in Continental lower choirs, though we do not know whether their portion of the nave was formally defined with a partition (further below). Excavations at Whithorn uncovered a wealth of evidence for the internal layout of a mid- to late-8th-century church that, in some respects, seems to have been remarkably similar to that described by Cogitosus (Fig 9). In overall form and scale, this church is also broadly similar to the later Irish mortared churches and therefore merits detailed discussion. There was apparently just one altar positioned slightly east of the midpoint of the church’s axis. A N/S screen of unknown height originally bisected this so that it was accessible from both chambers (altar 1).98 About 20 years later the altar was moved very slightly east (altar 2) and a new screen was inserted immediately east of it so that now it stood between two screens at the E end of the nave with a large space to the east of it. I have already mentioned the two partitions dividing at least part of the space west of the altar longitudinally, possibly to segregate the laity. This space was also divided from north to south by a third traverse screen. The area east of this screen, nearest the altar, could have been a lower choir for monks and lesser clerics, like the one postulated above for Cogitosus’ church. Another possibility is that it was for the more important members of the laity. In the case of Kildare, though Cogitosus only describes one longitudinal and one traverse ‘wall’ in detail, it could be argued that his final sentence leaves open the pos- sibility that there were other less substantial partitions separating sub-groups of the laity and religious: ‘And so, in one vast basilica, a large congregation of 97 Picard 2000, 153; pers comm 2008. 98 Hill 1997, 148.  architectural setting of the mass 141 fig 9 The church excavated at Whithorn as it had evolved by c 800 (after Hill 1997, figs 4.7, 4.9, 4.10, 4.13 and 4.15). Drawing by H Kavanagh and T Ó Carragáin. people of varying status, rank, sex and local origin, with partitions placed between them, prays to the omnipotent Master, differing in status, but one in spirit’.99 At Whithorn the chamber east of the second altar was about the same size as that immediately west of it but it was not divided longitudinally. The excavator suggested it was ‘for the clergy’, a conclusion supported by this comparison with Kildare. Another smaller chamber was later added to the E end of the church, interpreted as a sacristy complete with a soak-away pit for washing the altar plate and the celebrant’s hands.100 While there was only one altar in 7th-century Kildare and 8th-century Whithorn, we should consider the possibility that some of the Irish Viking-Age stone churches had more than one. As we have seen, in most major Viking-Age churches on the European mainland and in England the lower choir was west of the high altar, but there was at least one other axial altar west of this that was used for masses for the laity.101 Might the S windows of large Irish churches have illuminated secondary altars like these? Perhaps, but we might then expect a second S window to illuminate the side as well as the back of the high altar at the east. The only large churches with two S windows are the cathedrals of Glendalough and Mungret and we cannot rule out the possibility that these were simply a concession to better illumination. Given the short proportions of Irish churches, lateral altars are perhaps more likely. The only Irish source that explicitly mentions subsidiary altars is the Navigatio Brendani (c 800) where there is one to either side of the main altar, but other details of 99 Connolly and Picard 1987, 25–6. 100 Hill 1997, 157–8. 101 In the sanctuary area there was a high altar near the east for the chief ceremonies, including the capitular mass of the day, and a lesser altar where the morning mass was said (Conant 1959, 46; Fernie 2000, 88). For the monastic office, the community could be divided into separate choirs who sang before the different axial altars (Rabe 1995, 123). 142 tomás ó carragáin that narrative suggest that the author was not restricting himself to the realities of Irish ecclesiastical architecture. A somewhat more convincing case is the lateral exedra of the church on Iona: though Adomnán does not mention an altar within it, he does tell us that a monk called Fergnae went there specifically to pray and, though the term exedra usually denotes an alcove or recess, in de Locis Sanctis Adomnán uses it for the Chapel of the Chalice in the Holy Sepulchre complex at Jerusalem.102 Neither of these sources describes a typical Irish church and it seems safe to conclude that most of them had just one altar. Let us assume for a moment that the space behind the altar was primarily to accommodate substantial groups of religious. Allowing for two people stand- ing per metre squared, c 100 people could be comfortably accommodated east of the altar at Lorrha, c 95 at Glendalough, c 60 at Inishcealtra, c 50 at Kilree, c 40 at Inishfallen and c 25 at Templemacduagh. Thus, in theory, there would have been room for a large number of religious in addition to the priests assisting the celebrant. However, for a number of reasons this scenario seems unlikely. First, by analogy with churches abroad, most of the individuals in this part of the church probably sat. There is no Irish archaeological evidence for the positioning, or even the presence, of seats or benches,103 and most of our principal documentary sources fail to mention seating.104 The only exception is the Navigatio Brendani: ‘Around the church were ranged twenty four benches [one each for the 24 brothers of the monastery], with the abbot’s seat between the two choirs of monks in rows on either side’.105 Notwithstanding the fact that we must treat this source with caution (above), this excerpt is useful because it suggests that monks as well as abbots might sit. However, it does not give us a clear indication of where they sat. On the one hand their benches were in two groups of 12 to either side of the abbot’s chair that one might expect to be behind the altar, but on the other hand it states that they were ranged ‘around the church’. While Cogitosus refers to Kildare as a ‘cathedra episcopalis’,106 there is no mention of the cathedra itself in his description of the church. The term cathedra is also used for Armagh,107 and among the things destroyed in a fire there in 1020 (AU) was ‘the old preaching chair’ (sen-chathair preciupta). We are not told where exactly it had stood but, based on parallels abroad, we can be fairly confident that it was behind the altar of one of the three churches that were damaged in that fire, most likely that of the principal church known as the damliac mór.108 Assuming that all of the clerics east of the altar were seated, then the estimates above should be at least halved, and reduced further if we assume that the seats are well spaced. Second, it is possible, indeed likely, that some of the space behind the altar had other uses. The probable sacristy at the E end of the church excavated at Whithorn (above) suggests one possible additional purpose for this space; 102 Adomnán, Life of Columba, Book III, chapter 19 in Sharpe 1995; Meehan 1958, 51. 103 A clergy bench was added along the N wall of the 12th-century chancel of St Peter’s, Waterford (Hurley and McCutcheon 1997, 199–200). 104 For example Sharpe 1995; Herren 1974; Bieler 1979. 105 O’Donoghue 1893, 138. 106 Connolly and Picard 1987, 11–12. 107 Bieler 1979, 170–2. Iona (AU 716, 724; Etchingham 1999, 92) is the only other site for which the term cathedra is used. 108 The damliac and two other churches were also burned so it was probably in one of these.  architectural setting of the mass 143 grander Anglo-Saxon and Carolingian churches suggest others. A recurring feature of important churches in Carolingian Francia and in England throughout the Anglo-Saxon period is a pair of lateral chambers or porticus to either side of the high altar and accessible only from the sanctuary. This arrangement is found, for example, in 7th-century churches such as Reculver and SS Peter and Paul, Canterbury (Kent), the 9th-century minster dedicated to St Oswald, Gloucester, and in the late 10th-century extension to the Old Minster, Winchester.109 In some cases, including this latter example, there is also a large space of uncer- tain function east of the high altar. Unlike other porticus, the pairs of lateral chambers do not seem to have been primarily for burial. Rarely can we iden- tify their particular functions through architectural analysis or excavation, but on the St Gall plan it is specified that the northern one was a library for liturgical books while the southern one was a vestry and sacristy where altar plate was stored.110 Perhaps in the absence of externally defined lateral chambers, some of the space behind the altars of Irish churches was sometimes divided off to serve these functions. However, if so it seems unlikely that there were high partitions directly behind the altar that would prevent light from the E window illuminating it. Taking these factors into account, the number of religious behind the altars of the largest surviving churches like Lorrha could easily be reduced from a maximum of around 100 to less than 20. Thus, it seems likely that, as elsewhere in Europe, the majority of religious were usually accommodated west of the altar. Doubtless there was a good deal of variation over time and from site to site, but at least we have found evidence to suggest that, compared to minor churches, the principal ones at major sites were characterised by relatively complex spatial arrangements, and by extension it seems safe to conclude that they were designed for relatively complex liturgies. LATE 11TH- AND 12TH-CENTURY CHANGES While the liturgy is never static, it underwent particularly significant changes in the period between c 1050 and 1200 that find clear expression in Irish church architecture. Before then, it was already accepted that the bread and wine was the body and blood of Christ. For example, in the mid-8th century the Irish poet Blathmac son of Cú Brettan wrote: ‘It is your son’s body that comes to us when one goes to the Sacrament; the pure wine has been transmuted for us into the blood of the Son of the King’.111 But this mystery was not interrogated, except among small circles of intellectuals.112 Most believers took their cue from Augustine: ‘If you ask how this can be so, I shall briefly tell you. A mystery of faith can be profitably believed; it cannot be profitably examined’.113 But the Church began to develop a more rigorous doctrine on the matter in the latter half of the 11th century: a doctrine that the 109 Gem 2005. 110 Ibid, 272. 111 Carney 1964, 68–9, § 203. 112 On the Carolingian debate between Ratramnus and Rabbertus of Corbie see McCormick 1994 and Mazza 1999, 183–7. 113 Quoted in Rubin 1991, 21. 144 tomás ó carragáin Fourth Lateran Council (1215) would later refer to as Transubstantiation.114 During this period, therefore, the ‘real presence’ of the body and blood of Christ became far more important than those understandings of the mass that had predominated in the early Middle Ages: as a commemorative, communal meal or as an allegory for the Passion.115 Elsewhere in Europe the eastern cell or apse, which had previously functioned as a presbytery area, now became a chancel in which the altar was first placed at the W end but, before long, commonly stood against the E wall.116 This not only heightened the sense of mystery surrounding the sacrament but also the sense of separation between clergy and laity that the Church was also cultivating at this time.117 The writings of Berengarius sparked this development. He argued that the Eucharist was merely a figure of Christ’s body and blood.118 In his criticisms of Berengarius, the man who articulated the orthodox position most forcefully and who therefore had a formative influence on the development of the doctrine was Lanfranc of Bec.119 The strong position he took on this matter was one of the reasons he was appointed as Archbishop of Canterbury (c 1071–89). While there, he corresponded regularly with Irish kings and clerics about theological and pastoral matters, as did his successor, Anselm.120 For example, in 1074 Lanfranc wrote to high king Toirdealbach Ua Briain about the proper conduct of baptism, and arising from a synod on this issue in Dublin in 1080 Irish clerics sought further advice from him about it.121 While no similar correspondence relating directly to the character and architectural context of the mass survives, these contacts nevertheless provide a likely context for the construction of the first churches with eastern cells in Ireland in the last decade or so of the 11th century. These churches are found at just a handful of sites, all of which were in direct contact with Canterbury, namely the Hiberno-Norse ports and two earlier establishments: Killaloe (Co Clare) and Glendalough (Figs 10 and 11). The Hiberno-Norse ports had particularly close links with Canterbury, but we know the eastern end of just one of the port churches, St Peter’s in Waterford. Built in the late 11th or beginning of the 12th century, it originally comprised a stone chancel and wooden nave.122 The nave was later rebuilt in stone and an apse was added sometime in the first half of the 12th century: the only 12th-century apse known from Ireland. Killaloe was the chief church of the high king, Muirchertach Ua Briain, the principal secular champion of ecclesiastical reform in Ireland, and episcopal centre of his closest ecclesiastical ally, Máel Muire Ua Dúnáin. Richard Gem has shown that St Flannan’s, a barrel-vaulted nave-and-chancel church, was built there around 1098, and a stone-roofed chancel was probably added to an earlier church on nearby Friar’s Island (Co Clare) by the same masons. Finally, most if not all of the stone 114 de Lubac 1944, 23, 35. 115 Mazza 1999, 182–3. 116 Davidson 1999, 76; Barnwell 2004, 55. 117 For example Rubin 1991, 51; Tellenbach 1993, 167. 118 Mazza 1999, 190–1. 119 Rubin 1991, 18–19. 120 On the building projects of these two archbishops see Fernie 2000, 104–6, 140–4. 121 Holland 2002, 91; Cowdrey 2003, 114, 145. 122 Hurley and McCutcheon 1997, 198.  architectural setting of the mass 145 fig 10 Nave-and-chancel churches around 1100. Drawing by H Kavanagh and T Ó Carragáin. buildings of Glendalough were rebuilt in the period c 1096–1111 under the patronage of Muirchertach Ua Briain, probably in collaboration with Máel Muire Ua Dúnáin, the local Uí Muiredaig king and Glendalough clerics.123 The new buildings included at least four churches of nave-and-chancel type: St Kevin’s House, Reefert, Trinity and St Kieran’s. Up to that point, adherence to tradition had been more important than providing clergy with externally defined presbyteries,124 but as soon as eastern cells acquired this new liturgical and theological significance clerics and kings at these particular sites began to commission them. In addition to this architec- tural evidence, some of the texts produced at these sites also suggest an interest in keeping abreast of changes to the eucharistic liturgy. For example, the 123 Ó Carragáin forthcoming. 124 Ó Carragáin 2007. 146 tomás ó carragáin fig 11 The nave-and-chancel church of Trinity, Glendalough, which dates to around 1100. Photograph by T Ó Carragáin. mid-12th-century Latin Life of Flannan of Killaloe portrays the saint bringing an up-to-date version of the mass back with him from Rome: ‘Then he recounted the additions (incrementa) of the Holy Church through the four corners of the earth and the rites and the missarium solemnia of John the supreme pontiff’.125 More interesting still is a vernacular ‘Treatise on the Eucharist’ of around this date that focuses entirely on the issue of the Real Presence.126 It begins: ‘O you who do not have true belief regarding the feast you enjoy at the altar will be subject to a severe and painful judgement’. Among the various sources of evidence cited to demonstrate that Christ is fully present in the host is a miracle believed to have taken place at Whithorn in which the host temporarily takes the form of the Christ child (§ 56–66). A passage earlier in the treatise is worth quoting: For reasons of conscience in the sight of the Lord, do not share [your church] with devils. O man, believe that you partake of the body of the Son of the living God, that your wickedness may be forgiven . . . Those who believe are the body, Christ is the beautiful strong head; from these two — and this is not a matter of little moment — the perfection of the Church has grown (§ 30, 34). 125 Translated from Heist 1965 Vita Flannani § 13. 126 Murphy 1961, 21–8; Ó Maidín 1966, 147–54.  architectural setting of the mass 147 By this time the image of the church building as a metaphor for the institu- tion of the Church already had a long pedigree in Ireland; good Christians were likened to the individual timbers or building blocks and, collectively, they constituted the body of the church, while Christ was its foundation, cornerstone/ finial or head.127 However, this particular instance is distinct because it focuses on a bipartite division between the body of believers on the one hand and the head one the other. In the context of this treatise, the head is clearly the sanctuary area around the altar where Christ is made present. This gives the metaphor special resonance at a time when bicameral churches were beginning to appear at Irish sites. Linguistically this treatise could be as early as c 1090, but it seems to have been written by Isaac Ua Cuanáin, bishop of Roscrea (Co Tipperary), who died in 1161 (AFM), and so is more likely to date to the first half of the 12th century. It may therefore be contemporary with the relatively early Romanesque cathedral at Roscrea that was probably built in the 1120s or 1130s.128 Though only the W façade of this building now survives, by analogy with its closest comparanda at Ardfert (Co Kerry) and Cashel it must have had an externally defined eastern cell. In theory, existing churches could be adapted to take account of these changes simply by moving the altar eastwards and by adding more substantial screens between the altar and the congregation, like those Barnwell proposes in the case of the mid-12th-century unicameral church at Upton, Northampton.129 This is probably what happened to churches that never acquired an externally defined chancel, such as Clonmacnoise Cathedral. Another possible instance is St Caimin’s, Inishcealtra (Fig 6, phase 2). As mentioned above, this was originally a single-cell church, probably of late 10th- or early 11th-century date and with just one S window located roughly midway along the wall. A second S window was added further east during the 12th century. This could be contemporary with the Romanesque chancel, but with its broad fillet moulding and an absence of beaded moulding on its upper edge, it is quite different in style from the ex situ stones from the windows of the chancel. It may therefore have been inserted before the chancel was added, perhaps because the altar had been repositioned less than a metre from the E wall of the church. If so, it seems this rearrangement was not deemed adequate because the chancel was added soon afterwards and, as shown above, the altar was repositioned again, this time about 0.4 m from its E wall. Similarly, at Kilmalkedar (Co Kerry) the original 1130s church was unicameral with a niche for the altar projecting slightly from the E wall, but a proper chancel replaced this probably within a decade or two. Of the early nave-and-chancel churches the only one that has been excavated to modern standards is St Peter’s in Waterford. It had a chancel from the outset and it is possible that its altar was originally at the W end of the chancel, though the evidence is not conclusive.130 Within a few decades this putative altar was replaced by one near the E end, at which point the clergy 127 Ó Carragáin forthcoming. 128 For discussion of its date see O’Keeffe 2003, 180. 129 Barnwell 2004, 52. 130 Hurley and McCutcheon 1997, 199–200. 148 tomás ó carragáin bench was displaced to one side, as became the norm in the later Middle Ages (Fig 10). In the case of some of the other bicameral churches, the altar seems to have been at or near the E wall from the beginning. The stone altar-frontal of Reefert, Glendalough, is not in situ, but it is likely to have stood against the E wall for, because of its large size (over 1.5 m wide), any space behind it would have been inaccessible. As noted above, Cormac’s Chapel (c 1134) was designed with an altar against the E wall of the chancel in an externally defined altar niche from the outset (Fig 6). Elsewhere in Europe at this time large altars of stone were becoming more common and it has been argued that this represents a valorisation of the altar because of the increased emphasis on the idea that Christ’s body is made present upon it during the mass.131 It is therefore interest- ing that the earliest closely datable Irish stone altars are associated with some of these early bicameral churches, namely Reefert, St Kieran’s, St Peter’s and Cormac’s Chapel. These are the exceptions rather than the rule, however, and it seems likely that wooden altars remained common in high-medieval Ireland, as indeed they did in England.132 One other feature of the early nave-and-chancel churches remains to discuss. The smallest of them — Friar’s Island and St Kieran’s Glendalough — had a second doorway, in the S wall of the chancel (Fig 10). The S doorway of St Kieran’s hung from the outside, probably so that it could be opened outwards because of the tiny size of the chancel. Leask suggested that it led to a wooden sacristy but this is doubtful.133 Even in Britain, sacristies are unknown at minor sites in this period,134 and the excavation of Friar’s Island produced no evidence for one outside its S door.135 Furthermore, the early chancels that are better preserved than St Kieran’s (Friar’s Island, Reefert and Trinity) have one or more aumbries that probably provided adequate storage space for the sacred vessels associated with these buildings. Instead, these were probably priests’ doorways that further accentuated the division between congregation and clergy. CONCLUSION The principal conclusions of this research are as follows. First, there is no evidence to support the common assumption that Irish churches were not congregational. The church could accommodate the congregation, including lay people in many cases, who possibly entered the sanctuary for communion. Secondly, there is no evidence that the celebrant faced west during the Eucharist. Probably he usually faced east, though he would have faced the congregation for the liturgy of the Word. Thirdly, there are significant variations in altar position as indicated by the position of the S window. At relatively modest, pastoral churches it was probably usually freestanding but near the E wall. However, in important monastic and/or episcopal churches it was further west. 131 For discussion see Rauwel 2005, 177–8 who, however, is not fully convinced by this argument. 132 Cox and Harvey 1907, 1–2. 133 Leask nd 25–6. 134 Barnwell 2004, 53. 135 Macalister 1929.  architectural setting of the mass 149 This probably facilitated relatively complex liturgies with many concelebrating priests. In some cases this space may also have served as a sacristy, vestry and/ or library. Finally, from around 1100, in response to an increased emphasis on the real presence of Christ in the Eucharist and on separation between laity and clergy, altars tended to be positioned further east, often in externally defined chancels. The most important of these conclusions is the first one, for it has implica- tions beyond liturgy or theology. If we accept that Irish churches were gener- ally congregational, we can in future proceed on the assumption that there was a positive, though by no means exact, correlation between their size and the community that they were designed to serve, as has been shown in the case of later parish churches.136 Early-medieval churches were more densely distributed than parish churches and were generally smaller, but proportionally speaking they varied much more in size. Along with other archaeological and historical evidence, this expresses the fact that early ecclesiastical sites were established for smaller and much more varied groups than their high-medieval successors, including individual families, larger lay communities and/or a variety of clerical and monastic communities. Beyond that, we can now recognise that the size and form of churches are important sources of evidence for anyone attempting to distinguish between these diverse groups. This is, of course, crucial if we are to develop better understandings of early ecclesiastical organisation in Ireland and of the relationships between ecclesiastical and secular power structures.137 acknowledgements I am very grateful to Cormac Bourke, Dr Richard Gem and Prof Éamonn Ó Carragáin for reading the paper and improving it considerably. Thanks also to Prof Jean-Michel Picard for discussing aspects of Cogitosus’ Life of St Brigit with me and to Fr Neil O’Donoghue for a copy of his PhD thesis. Any errors that remain are my own. BIBLIOGRAPHY Ahrens, C 2001, Die frühen Holzkirchen Bhreathnach, E 2001, ‘Abbesses, minor Europas, Stuttgart: Theiss. dynasties and kings in clericatu: perspec- Barnwell, P 2004, ‘The laity, the clergy tives of Ireland, 700–850’, in M P Brown and the divine presence: the use of space and C A Farr (eds), Mercia: An Anglo-Saxon in smaller churches of the eleventh and Kingdom in Europe, Leicester: Leicester twelfth centuries’, J Brit Archaeol Ass 157, University Press, 113–25. 41–60. Blair, J 1996, ‘Churches in the early English Best, R and Bergin, O 1929, Lebor na hUidre, landscape: social and cultural contexts’, Dublin: Royal Irish Academy. in J Blair and C Pyrah (eds), Church Biddle, M 1970, ‘Excavations at Winchester Archaeology: Research Directions for the Future, 1970. Ninth interim report’, Antiq J 52, York: Council for British Archaeology, 104–7. 6–18. Bieler, L 1963, The Irish Penitentials, Dublin: Bolman, E 2006, ‘Veiling sanctity in Chris- Dublin Institute for Advanced Studies. tian Egypt: visual and spatial solutions’, Bieler, L (ed) 1979, The Patrician Texts in the in Gerstel (ed), 73–106. Book of Armagh, Dublin: Dublin Institute Bond, C 1988, ‘Church and parish in for Advanced Studies. Norman Worcestershire’, in J Blair (ed), 136 Ní Ghabhláin 1995 187–236; Bond 1988, 119, 141–4. 137 See further Ó Carragáin 2006, 109–23. 150 tomás ó carragáin Minster and Parish Churches, Oxford: Oxford Dodds, J 1990, Architecture and Ideology in University Press, 119–58. Early Medieval Spain, Pennsylvania: Penn- Bourke, C 1993, Patrick: The Archaeology of sylvania State Press. a Saint, Belfast: HMSO. Dunraven, E 1875–7, Notes on Irish Architecture, Brady, N 1997, ‘De Oratorio: Hisperica 2 vols, London: G Bell. Famina and church building’, Peritia 11, Duval, N 2005, ‘L’autel paléochrétien: les 327–35. progress depuis le livre de Braun (1924) Brash, R 1866, ‘Inishcealtra and its rem- et les questions à résoudre’, Hortus Artium ains’, The Gentleman’s Magazine January Medievalium 11, 7–18. 1866, 7–22. Etchingham, C 1999, Church Organisation in Caillet J-P 2005, ‘L’image cultuelle sur Ireland AD 650 to 1000, Maynooth: NUI l’autel et le positonnement du celebrant Maynooth. (IXe–XIVe siècles)’, Hortus Artium Medie- Fanning, T 1981, ‘Excavation of an early valium 11, 139–48. Christian cemetery and settlement at Carney, J 1964, The Poems of Blathmac Son of Reask, County Kerry’, Proc Roy Ir Acad Cú Brettan, Dublin: Irish Texts Society. 81C, 67–172. Carre, J-M (ed) 1996, Les Églises Doubles et Fernie, E 1983, The Architecture of the les Familles d’Églises, Tournhout: Brepols. Anglo-Saxons, London: Batsford. Carver M (ed) 2003, The Cross Goes North: Fernie, E 2000, The Architecture of Norman Processes of Conversion in Northern Europe, England, Oxford: Oxford University Press. AD300–1300, Woodbridge: Boydell and Fleming, J (ed) 2001, Gille of Limerick Brewer. (C 1070–1145): Architect of a Medieval Church, Codou, Y and Fixot, M 1996, ‘Églises Dublin: Four Courts. doubles et familles d’églises en provence’, Gamber, K 1976, Liturgie und Kirchenbau: in Carre (ed), 196–210. Studien zur Geschichte der Messfeier und des Conant, J 1959, Carolingian and Romanesque Gotteshauses in der Frühzeit, Regensburg: Architecture, 800 to 1200, Yale: Yale Pustet. University Press. Gem R 2005, ‘How much can Anglo-Saxon Connolly, S and Picard, J-M 1987, ‘Cogito- buildings tell us about liturgy’, in H sus: Life of Saint Brigit’, J Roy Soc Antiq Ir Gittos and M Bradford Bedingfield (eds), The Liturgy of the Late Anglo-Saxon Church, 117, 11–27. Woodbridge: Boydell and Brewer, 271– Cowdrey, H 2003, Lanfranc. Scholar, Monk 90. and Archbishop, Oxford: Oxford University Gerstel, S (ed) 2006, Thresholds of the Sacred. Press. Architectural, Art Historical, Liturgical and Cox, J and Harvey, A 1907, English Church Theological Perspectives on Religious Screens, East Furniture, London: Methuen. and West, Washington DC: Dumbarton Davidson, C F 1999, ‘Change and change Oaks Research Library. back: the development of English parish Gillespie, R, Duffy S and FitzPatrick E church chancels’, in R N Swanson (ed), (eds) 2006, The Parish in Medieval and Post- Continuity and Change in Christian Worship, Medieval Ireland, Dublin: Four Courts. Woodbridge: Boydell and Brewer, 65–77. Graves, P 2000, The Form and Fabric of Doig, A 2008, Liturgy and Architecture from the Belief. An Archaeology of the Lay Experience Early Church to the Middle Ages, Aldershot: of Religion in Medieval Norfolk and Devon, Ashgate. Brit Archaeol Rep Brit Ser 311. De Blaauw, S 1994, Cultus et Decor. Liturgia Gwynn E (ed) 1927, The Rule of Tallaght, e Architettura nella Roman Tardoantica e Dublin: Hodges and Figgis. Medievale, Città del Vaticano: Biblioteca Gwynn E and Purton, W (eds) 1911–12, Apostolica Vaticana. ‘The Monastery of Tallaght’, Proc Roy Ir de Lubac, H 1944, Corpus mysticum. Acad 29C, 115–80. L’Eucharistie et l’Église au Moyen Age, Paris: Heist, W (ed) 1965, Vitae sanctorum Hiberniae: Aubier. ex codice olim salmanticensi nunc Bruxellensi, de Paor, L 1997, Inis Cealtra. Report on Brussels: Société Des Bollandistes. archaeological and other investigations Henry, F 1940, Irish Art in the Early Christian of the monuments on the island (unpubl Period, Ithaca: Cornell University Press. report to the National Monuments Service, Herbert, M and McNamara, M 1989, Department of the Environment Heritage Irish Biblical Apocrypha: Selected Texts in and Local Government, Ireland). Translation, Edinburgh: T and T Clark.  architectural setting of the mass 151 Herren, M W (ed) 1974, The Hisperica McClendon, C 2005, The Origins of Early Famina 1. The A-Text, Toronto: Pontifical Medieval Architecture, Yale: Yale University Institute of Mediaeval Studies. Press. Herren, M and Brown, S 2002, Christ in McCormick Zirkel, P 1994, ‘The ninth- Celtic Christianity Britain and Ireland from century eucharistic controversy: a context the Fifth to the Tenth Century, Woodbridge: for the beginnings of eucharistic doctrine Boydell and Brewer. in the West’, Worship 68:1, 2–23. Hill, P 1997, Whithorn and St Ninian The MacDonald, A 2003, ‘The “cathedral”, Excavation of a Monastic Town, 1984–91, Temple Kelly and Temple Ciarán: Stroud: Sutton. notes from the annals’, in H King (ed), Holland, M 2002, ‘The synod of Dublin in Clonmacnoise Studies Volume 2 Seminar Papers 1080’, in S Duffy, Medieval Dublin III, 1998, Dublin: Wordwell, 125–35. Dublin: Four Courts, 81–94. Mazza, E 1999, The Celebration of the Eucha- Hunwicke, J W 2002, ‘Kerry and Stowe rist: The Origin of the Rite and the Development revisited’, Proc Roy Ir Acad 102C, 1–19. of its Interpretation, Collegeville: Liturgical Hurley, M and McCutcheon, S 1997, Press. ‘St Peter’s church and graveyard’, in Marshall, J W and Walsh, C 2005, Illaun- M F Hurley, O M B Scully and W J loughan Island. An Early Medieval Monastery McCutcheon (eds), Late Viking Age and in County Kerry, Dublin: Wordwell. Medieval Waterford: Excavations 1986–1992, Meeder, S 2005, ‘The early Irish Stowe Dublin: Institute of Public Administration, Missal’s destination and function’, Early 190–227. Medieval Europe 13:2, 179–94. Ivens, R 1989, ‘Dunmisk fort, Carrickmore Meehan, D (ed) 1958, Adamnán’s De Co. Tyrone excavations 1984-86’, Ulster Locis Sanctis, Dublin: Dublin Institute of J Archaeol 52, 17–110. Advanced Studies. Jungmann, J A 1954, Das Eucharistische Meyer, K (ed) 1912, Sanas Chormaic, an Old- Hochgebet: Grundgedanken des Canon Missae, Irish Glossary compiled by Cormac mac Wÿuurzburg: Werkbund. Cuilennáin king-bishop of Cashel in the Kelly, E 2003, ‘The Tully Lough Cross’, 10th century edited from the copy in the Archaeol Ir 64, 9–10. Yellow Book of Lecan. Anecdota from Irish Manuscripts, Halle: Niemeyer. Klauser, T 1979, A Short History of the Western Moore, M J 1984, ‘Irish cresset-stones’, Liturgy, Oxford: Oxford University Press. J Roy Soc Antiq Ir 114, 98–116. Krautheimer, R 1986, Early Christian and Morris, R 1989, Churches in the Landscape, Byzantine Architecture, Yale: Yale University London: Dent. Press. Murphy, G 1961, ‘Eleventh- or twelfth- Laing, L 2006, The Archaeology of Celtic Britain century Irish doctrine concerning the and Ireland c AD 400–1200, Cambridge: Real Presence’, in J Watt, J Morrall and Cambridge University Press. F X Martin (eds), Medieval Studies Present Lara, J 1994, ‘“Versus populum” revisited’, to Aubrey Gwynn, S J, Dublin: Colin Worship 68.3, 210–21. O’Lochlainn, 19–28. Leask, H 1955, Irish Churches and Monastic Murray, G 2007, The Cross of Cong and Buildings, Dundalk: Dundalgan. Church Metalwork from Romanesque Leask, H nd, Glendalough, Co. Wicklow. Ireland (unpubl PhD thesis, University National Monuments vested in the Com- College Cork). missioners of Public Works. Official His- Neuman de Vegvar, C 2003, ‘Romanitas torical and Descriptive Guide, Dublin: and Realpolitik in Cogitosus’s description Government of Ireland. of the Church of St Brigit, Kildare’, in Macalister, R A S 1916, ‘The history and Carver (ed) 2003, 153–70. antiquities of Inis Cealtra’, J Roy Soc Antiq Ní Ghabhláin, S 1995, Church, parish and Ir 33, 93–174. polity: the medieval diocese of Kilfenora, Macalister, R A S 1929, ‘On some excava- Ireland (unpubl PhD thesis, University of tions recently conducted on Friar’s Island’, California, Los Angeles). J Roy Soc Antiq Ir 59, 16–24. Nussbaum, O 1965, Der Standort des Liturgen Macalister, R A S 1935, Ancient Ireland: A am Christlichen Altar vor den Jahre 1000. Study in the Lessons of Archaeology and History, Eine Archäologische und Liturgiegeschichtliche London: Methuen. Unterscuchung, Bonn: Hanstein. 152 tomás ó carragáin Ó Carragáin, É 1988, ‘The meeting of O’Loughlin, T 2000, Celtic Theology: Human- Saint Paul and Saint Anthony: visual and ity, World and God in Early Irish Writings, literary uses of a eucharistic motif’, in London: Continuum. G Mac Niocaill and P F Wallace (eds), O’Loughlin, T 2003, ‘The praxis and Keimelia: Studies in Medieval Archaeology and explanations of eucharistic fraction in History in Memory of Tom Delaney, Galway: the ninth century: the Insular evidence’, Galway University Press, 1–58. Archiv für Liturgiewissenschaft 45, 1–20. Ó Carragáin, É 1999, ‘The term porticus O’Loughlin, T 2004, ‘Translating Panis and imitatio Romae in Early Anglo-Saxon in a eucharistic Context: A Problem England’, in H C O’Briain, A M d’Arcy of Language and Theology’, Worship 78, and J Scattergood (eds), Text and Gloss: 226–35. Studies in Insular Learning and Literature Ó Maidín, U 1966, The Celtic Monk. Rules Presented to Joseph Donovan Pheifer, Dublin: and Writings of Early Irish Monks, Kalamazoo: Four Courts, 13–34. Cistercian Publications. Ó Carragáin, T 2003a, ‘A Landscape Con- verted: Archaeology and Early Church Ó Riain, P 1991, ‘The shrine of the Stowe Organisation on Iveragh and Dingle, missal, redated’, Proc Roy Ir Acad 91C, Ireland’, in Carver (ed), 127–52. 285–95. Ó Carragáin, T 2003b, ‘The architectural O’Sullivan, J and Ó Carragáin, T 2008, setting of the cult of relics in early Inishmurray. Monks and Pilgrims in an Atlantic medieval Ireland’, J Roy Soc Antiq Ir 133, Landscape, Cork: Collins. 130–76. Ó Néill, P 2000, ‘The Old-Irish tract on Ó Carragáin, T 2005, ‘Habitual masonry the Mass in the Stowe Missal: some styles and the local organisation of church observations on its origins and textual building in early medieval Ireland’, Proc history’, in Smyth (ed), 199–204. Roy Ir Acad 105C, 99–149. Parsons, D 1986, ‘Sacrarium: ablution Ó Carragáin, T 2006, ‘Church buildings drains in EM churches’, in L A S Butler and pastoral care in early medieval and R K Morris (eds), The Anglo-Saxon Ireland’, in Gillespie et al (eds), 91–123. Church. Papers on History, Architecture and Ó Carragáin, T 2007, ‘Skeuomorphs and Archaeology in Honour of Dr H Taylor, spolia. The presence of the past in Irish London: Council for British Archaeology, pre-Romanesque architecture’, in R Moss 105–20. (ed), Making and Meaning in Insular Art, Parsons, D 1996, ‘Liturgical and social Dublin: Four Courts, 95–109. aspects’, in A Boddington, Raunds Furnells. Ó Carragáin, T forthcoming, Churches in The Anglo-Saxon Church and Cemetery, Early Medieval Ireland. Architecture, Ritual London: English Heritage, 58–66. and Memory, Yale: Yale University Press. Petrie, G 1845, The Ecclesiastical Architecture of O’Donoghue, D (ed) 1893, Brendaniana: Ireland, Dublin: Hodges and Smith. St Brendan the Voyager in Story and Legend, Picard, J-M 2000, ‘Princeps and principatus in Dublin: Browne and Nolan. the early Irish Church: a reassessment’, O’Donoghue, N 2006, The eucharist in in Smyth (ed), 146–60. pre-Norman Ireland: liturgy, practice and society (unpubl PhD thesis, St Rabe, S 1995, Faith, Art and Politics at Patrick’s College, Maynooth). Saint-Riquier. The Symbolic Vision of Angilbert, O’Donovan J 1868, Sanas Chormaic: Cormac’s Philadelphia: University of Pennsylvania. Glossary, Calcutta: Irish Archaeological Radford R 1977, ‘The earliest Irish and Celtic Society. churches’, Ulster J Archaeol 40, 1–10. O’Keeffe, T 2003, Romanesque Ireland. Archi- Rauwel, A 2005, ‘Théologie de L’Eucharistie tecture and Ideology in the Twelfth Century, et Valorisation de L’Autel à L’Âge Dublin: Four Courts. Roman’, Hortus Artium Medievalium 11, O’Keeffe, T 2006, ’The built environment 177–83. of local community worship between Repsher, B V (ed) 1998, The Rite of the late eleventh and early thirteenth Church Dedication in the Early Medieval Era, centuries’, in Gillespie et al (eds), 124–47. Lewiston: Edwin Mellen Press. O’Kelly, M J 1958, ‘Church Island near Ryan, M 2007, ‘Sacred Cities?’, in A Valencia, Co. Kerry’, Proc Roy Ir Acad Minnis and J Roberts (eds), Text, Image, 59C, 57–136. Interpretation. Studies in Anglo-Saxon Literature  architectural setting of the mass 153 and its Insular Content in Honour of Éamonn cause, effect or neither?’, in Gerstel (ed), Ó Carragáin, Turnhout: Brepols, 515–28. 27–52. Rubin, M 1991, Corpus Christi. The Eucharist Taylor, H M 1973, ‘The position of the in Late Medieval Culture, Cambridge: Cam- altar in early Anglo-Saxon churches’, bridge University. Antiq J 53, 52–8. Schneiders, M 1996, ‘The origins of early Taylor, H W 1979, From Temple to Meeting Irish liturgy’, in P Ní Chatháin and M House. The Phenomenology and Theology of Richter (eds) 1996, Irland und Europa im Places of Worship, The Hague: Mouton. früheren Mittelalter, Stuttgart: Klett-Cotta, Tellenbach, G 1993, The Church in Western 76–98. Europe from the Tenth to the Early Twelfth Sharpe, R (ed) 1995, Adomnán of Iona Life of Century, Cambridge: Cambridge University. St Columba, London: Penguin. Thomas, C 1967, ‘An Early Christian Sheehan, J forthcoming, ‘A peacock’s tale: cemetery and chapel on Ardwall Isle, excavations at Caherlehillan, Iveragh, Kircudbright’, Medieval Archaeol 11, 127– Ireland’, in N Edwards (ed), The Archaeology 88. of the Early Medieval Celtic Churches, Leeds: Thomas, C 1971, The Early Christian Archae- Maney. ology of North Britain, London: Oxford Smyth, A P (ed) 2000, Seanchas: Studies in University Press. Early and Medieval Irish Archaeology, History Vogel, C 1962, ‘Sol aequinoctialis: Problemes and Literature in Honour of Francis J. Byrne, et technique de l’orientation dans le culte Dublin: Four Courts. Chretien’, Revue des Sciences Religieuses 36, Smyth, M 2003 La Liturgie Oubliée. La 175–211. Prière Eucharistique en Gaule Antique et dans Vogel C 1964, ‘L’orientation vers l’ Est l’Occident non Romain, Paris: Les Éditions du celebrant et des fidèles pendant le du Cerf. celebration Eucharistique’, L’Orient Syrien Stokes, W (ed) 1901, ‘The Lebar Brecc 9, 3–37. tractate on the consecration of a church’, Warren, F E 1881, The Liturgy and Ritual of in S Timpanaro (ed), Miscellanea Linguis- the Celtic Church, Oxford: Clarendon. tica in Onore di Graziadio Ascol, Turin (repr Génève: Slatkine Reprints), 363–87. Abbreviations Stokes, W and Strachan, J (eds) 1903, Thesaurus Palaeohibernicus A Collection of AFM O’Donovan, J (ed) 1851, Annals of the Old-Irish Glosses Scholia Prose and Verse, Kingdom of Ireland by the Four Masters, Cambridge: Cambridge University. Dublin: Royal Irish Academy. Taft, R 1978, The Great Entrance. A History of AU Mac Airt, S and MacNiocaill, G the Transfer of Gifts and other Pre-Anaphoral (eds) 1983, The Annals of Ulster (to Rites, Rome: Pontificium Institutum AD 1131), Dublin: Dublin Institute Studiorum Orientalium. of Advanced Studies. Taft, R 2006, ‘The decline of communion DIL 1913–1976 Contributions to a Dictionary in Byzantium and the Distancing of the of the Irish Language, Dublin: Royal Congregation from the liturgical action: Irish Academy. Résumé Le cadre architectural de la messe dans l’Irlande du très haut Moyen Âge par Tomás Ó Carragáin Les églises et les documents qui ont survécu jusqu’à nos jours contiennent des indications sur le contexte architectural de la messe dans l’Irlande du très haut Moyen Âge. Après examen, il apparaît qu’aucune preuve ne vient étayer l’opinion largement répandue selon laquelle les fidèles se tenaient à l’extérieur de l’église. La taille variable, mais relativement petite, de ces églises témoigne plutôt du fait qu’elles servaient des paroisses moins nom- breuses et plus diverses que leurs remplaçantes du haut Moyen Âge. Les autels des grandes églises épiscopales et/ou monastiques semblent être placés plus à l’ouest que ceux des églises champêtres d’assez petite taille. Il s’agissait sans doute en partie de faciliter des liturgies eucharistiques relativement complexes. Des chœurs distincts apparaissent pour la première fois à la fin du XIe siècle alors que se renforce l’idée de la réelle présence du Christ 154 tomás ó carragáin dans l’Eucharistie. Il est significatif de noter qu’on les trouve dans quelques sites importants dont les clercs et les patrons étaient en contact direct avec Lanfranc de Canterbury, l’un des principaux représentants de cette doctrine. Zusammenfassung Der architektonische Rahmen für die Messe im frühmittelalterlichen Irland von Tomás Ó Carragáin Erhaltene Kirchen und Dokumente enthalten Beweisstücke für den architektonischen Rahmen der Messe im frühmittelalterlichen Irland. Die Untersuchung ergibt, dass es keine Anhaltspunkte gibt, die die weit verbreitete Annahme stützen, dass die Gemeinde draußen vor der Kirche stand. Stattdessen drückt sich in der variablen, aber relativ kleinen Größe dieser Kirchen die Tatsache aus, dass sie kleineren und sehr viel unterschiedlicheren Gemeinden dienten als ihre hochmittelalterlichen Nachfolger. Die Altäre in großen Bischofs- und/oder Klosterkirchen scheinen weiter westlich zu stehen als diejenigen in relativ kleinen Gemeindekirchen. Teilweise sollte damit wahrscheinlich Raum für die relativ komplexe eucharistische Liturgie geschaffen werden. Kirchen mit klar definiertem Schiff und Altar- raum tauchen zum ersten Mal im späten 11. Jahrhundert n. Chr. auf und waren ein Ausdruck der zunehmenden Betonung der wahren Gegenwart Christi in der Eucharistie. Bezeichnenderweise erscheinen sie in einer Hand voll wichtiger Kirchen, deren Priester und Gönner direkt mit Lanfranc von Canterbury in Kontakt waren, der als einer der wichtigsten Vertreter dieser Doktrin gilt. Riassunto L’architettura degli edifici in cui si celebrava la messa nell’Irlanda del primo Medioevo di Tomás Ó Carragáin Le chiese e i documenti giunti fino a noi contengono indicazioni riguardo al contesto architettonico degli edifici in cui si celebrava la messa nell’Irlanda del primo Medioevo. L’esame rivela che non esiste alcuna prova a supporto dell’opinione largamente diffusa che i fedeli restassero all’esterno della chiesa. Anzi, le dimensioni variabili, ma relativamente piccole, di queste chiese sono una dimostrazione del fatto che esse servivano comunità meno numerose e più variabili rispetto alle successive chiese dell’alto Medioevo. Nelle grandi chiese episcopali e/o monasteriali gli altari sembrano essere collocati più a ovest rispetto a quelli delle chiese campestri di dimensioni piuttosto piccole. In parte questo era dovuto probabilmente all’esigenza di facilitare le liturgie eucaristiche, relativamente complesse. Il corpo del presbiterio è distinguibile all’esterno per la prima volta nel tardo XI secolo d.C., come conseguenza della cresciuta importanza data alla reale presenza di Cristo nell’Eucaristia. Non a caso esso compare in diversi siti importanti, i cui ecclesiastici e finanziatori erano in contatto diretto con Lanfranco di Canterbury, uno dei principali esponenti di questa dottrina.