Church Archaeology 410 to 597: the Problems of Continuity

CHURCH ARCHAEOLOGY 410 TO 597: THE PROBLEMS OF CONTINUITY ALAN WARD The Venerable Bede, writing c.730 in an oft-quoted passage referring to the beginnings of Christianity in Anglo-Saxon England, tells us that Bertha, the Frankish and Christian wife of Aethelberht King of Kent (c.589-c.616), worshipped at the church of St. Martin on the east side of Canterbury (Bede 1974, p.70). He also tells us that this church had been built by the Romans and was dedicated to St. Martin from its foundation.1 Bede tells us a similar story for the foundation of the cathedral within the city walls, where St Augustine founded a church consecrated in the name of Our Saviour, God and Lord Jesus Christ and a' dwelling for himself and his successors' on the site of an old church built by Roman Christians (Bede J 974, p. 91). Two themes are here intertwined, one the idea of continuity of religious structures and the other, perhaps not so obvious, the implied continuity of the Christian religion within Kent from the end of Roman Britain and through the fifth and sixth centuries. Let us take the second theme first. Bede's statements are telling us a tradition of Christian worship had been kept alive. The obvious question is, kept alive by whom? The equally obvious answer is, by a community which had retained Christianity as their religion ( or at least as part of their religion). Needless to say several objections can be made against this attractive theory. We have a tendency to imagine most of the population of Roman Britain in the late fourth and early fifth centuries as being Christian. Whilst this may have been true of the elite and perhaps even of the urban population, most people were agricultural labourers and we have no way of knowing how far 'down' the social scale Christianity had travelled in Kent by (say) AD 410. 2 The monotheistic, intolerant and hierarchical Christian Church of that date would have been at home in the police state of the late Roman Empire. The collapse of that centralised government within Britain may well have lessened the pressure on Romano-British pagans to convert. Whereas in the West and North there was enough stability and time during the course 375 ALAN WARD of the fifth century for Christianity (for whatever reason) to be accepted, the same may not have been true of the eastern part of the country. There is little evidence for Christianity being practised in the South-East in the first half of the fifth century and no definite evidence after about 450.3 The second objection concerns our most obvious source of evidence for the fifth and sixth centuries - burials. As far as the writer is aware not a single definite Christian burial of this two hundred year period has been found in Kent. It is regularly pointed out that many burials in supposed Anglo-Saxon pagan cemeteries are orientated west to east and have no grave goods. That is of course a perfectly true statement, but many burials with grave goods are also orientated west to east and others, both with and without grave goods, north to south. Burials without grave goods may merely reflect materials such as the choicest cuts of meat, carved wooden objects or rich imported textiles, of which no trace survives. The poverty of the deceased could be reflected in what is not appearing in the grave or indeed the sheer selfishness and greed of relations or neighbours and/or their dislike of the deceased. An increasing number of archaeologists have also realised that Germanic grave goods do not necessarily convey the ethnic origins of the deceased (e.g. Higham 1992, p. 179f, 225). Suggestions have also been put forward that some pagan period Anglo-Saxon graves may in fact have been Christian burials; in Kent for example, material from cemeteries at Strood, Horton Kirby and the Chatham Lines have been mentioned in this context (Dark 2000, pp.78-83). The writer takes the conservative (perhaps negative) view that these graves contained objects possibly of Christian origin but buried in a pagan context. The east of the country, including Kent, is not well endowed with stone, but there were abundant ruined masonry Roman structures and yet from all the Anglo-Saxon cemeteries excavated not a single stone marker with a chi-rho, let alone an inscription, has been recovered. Ultimately there is no way of knowing whether Christianity is represented in these pagan cemeteries but the evidence used to support such an assertion is at best ambiguous. How many of those buried were British rather than Germanic is another question entirely. Thirdly but perhaps of not such importance is that the intolerant, courageous, fundamentalist and yet charitable Martin, Bishop of Tours died in 397 and was (presumably) canonised very shortly after.4 The Roman administration had left Britain by 41 0 and this does not allow much time for a Roman church dedicated to St Martin to have been founded just outside Canterbury. The Taylors have suggested a sub-Roman or early Anglo-Saxon date for this structure (Taylor and Taylor 1965, p. 143), but whilst we can see archaeologically that 376 CHURCH ARCHAEOLOGY 410-597: THE PROBLEMS OF CONTINUITY stone buildings were still in use into the fifth, perhaps well into the fifth century, there is very little, if any, evidence for new mortared masonry structures being built in the South-East, and again, as far as the writer is aware, no evidence for such structures in Kent (with the exception of the baptistery at Richborough, itself almost certainly pre-410) until after 597. Continuing in this negative mode several points put forward by Nicholas Brooks which are used to support, or at least imply, the continuation of Christianity within Kent (Brooks 1984, pp. 17-21) are also open to alternative interpretations. The Roman villa of Eccles is, of course, famous in Kentish archaeology. As a place-name it is mentioned by all and sundry as being derived from the British ecles (from the Greek ecclesia, 'a church') and is then used to support the idea of Christian continuity from the Roman into the Anglo-Saxon period. A mid seventh-century Anglo-Saxon Christian cemetery immediately adjacent to the villa is used to support this view. A timber building possibly of the same phase as the cemetery is sometimes also quoted, and by implication is regarded as a church. This story is potentially one of the biggest 'red-herrings' in Kentish archaeology. The Christian phase of the cemetery appears to follow on from an earlier pagan or, perhaps more likely, transitional phase when grave goods were still deposited (Detsicas 1974, p. 130). The cemetery may have been established at this point for the very simple reason the land was so strewn with rubble that it was regarded as waste and hence no use for the growing of crops. The timber building appears not to have been excavated in its entirety (Detsicas 1975, fig. I). A greater criticism lies in the derivation of the place-name itself. Whilst the name appears quite early (975) and Wallenberg supports the traditional interpretation (I 931, p. 305), Judith Glover gives an Old English alternative, aec laes, meaning meadow of the oak (1976, p. 65) and Ekwall puts forward the view that some Eccles place-names may derive from a (presumably Germanic) personal-name, Ecca (1960, p. 159). Even if the religious interpretation is correct the obvious question to ask is why is a British derivation necessary? The study of place-names is fraught with hidden dangers for the unwary, but as far as this writer can discern, the place-name could just as likely derive (albeit in a corrupted form) directly from the Greek, perhaps in the seventh or eighth century, as from an earlier British source. The story of Sixtus a local Romano-British Christian martyr (Brooks 1984, p. 20) whilst interesting in itself may have little to do with Kent. A letter from Pope Gregory the Great shows that a local martyr was venerated within the area of Augustine' s missionary authority. That is of course the catch, no specific place is mentioned. Bede tells us 'AJl the bishops of Britain ... ' were committed to 377 ALANWARD Augustine's charge (Bede 1974, p. 76), i.e. Celtic as well as future Anglo-Saxon bishops, including any future bishop of York and we know he met British ecclesiastics in the west country (p. 101). There appears to be no good reason to associate Sixtus with Kent; he could have been venerated more or less anywhere in Britain. The story is just as likely to belong to western areas with their large British population recently conquered by an Anglo-Saxon aristocracy. The position of some of the churches of Canterbury have also been used to suggest religious continuity, namely the alignment of St Peter' s on the Roman street system, and St Dunstan' s, St Paul's, St Sepulchre's and St Augustine's occupying the sites of Roman cemeteries outside the city wall (Brooks 1984, pp. 20-21). For St Peter's the church is (interestingly) on two different alignments, the earliest (earlymid twelfth century) whilst it could be argued runs (approximately) parallel to the Roman road alignment is also at right angles to St Peter's Lane, the curved nature of which suggests it is not Roman. With all the many thousands of deaths that must have occurred during the life of the Roman town it is hardly surprising that a few churches end up on, or close to, Roman cemeteries. As a religion Christianity depends very largely on two factors, organisation (usually hierarchical) and Scripture. This is perhaps especially true of the early Church. The leaders of the fourth century Church were the bishops, and if we follow the Romano-Gaulish model it seems probable that most were drawn from the aristocracy. The bishoprics were also dependent on the towns. With the demise of both an urban society and the Romano-British aristocracy in Kent and East Anglia by the end of the fifth century, it seems probable that any episcopal organisation in these areas, and probably well beyond, would have disappeared. Without that literate organisation, however corrupt (judging by sixth century Gaulish standards) it may have been, the possibility of Christianity surviving in the South-East seems remote. There is no hint of travelling clergy attempting to keep Christianity alive. Even assuming that literate Romano-British peasants had at one time existed, it seems highly unlikely that any would have the ability to read Scripture in a meaningful way to a congregation. Also what appears to be a complete lack of physical evidence, in the form of writing equipment, inscribed stone markers and (perhaps tellingly) graffiti on the abundant pottery recovered from the fifth- and sixth-century cemeteries argues in favour of a totally non-literate society within the eastern part of the country. The survival of Scripture therefore seems highly unlikely. Based on the lack of evidence for Christianity surviving within Kent for most of the fifth and sixth centuries the writer can see no reason why a Germanic pagan kingdom should keep alive a tradition 378 CHURCH ARCHAEOLOGY 410-597: THE PROBLEMS OF CONTINUITY of St Martin's Church and the cathedral being founded by Roman Christians. This 'tradition' may have had more to do with ecclesiastical and perhaps secular politics of the early eighth century. Albinus Abbot of SS Peter and Paul at Canterbury (and Bede's probable informant) may have been attempting to emphasise the position of Canterbury over the other English bishoprics and perhaps by implication to enhance the, by now much weakened, Kentish kingdom. Alternatively, by the early eighth century the churches built of Roman material were over a hundred years old and may have had the look of being older than they actually were (Blockley et al 1997, p. 99) and hence Albin us may have been repeating a tradition that had built up during the course of the seventh century. It is also perhaps worth noting that St. Augustine and his party of monks made a detour to the shrine of St Martin at Tours on their way to Kent (Mayr-Harting 1991, p. 61). Such a detour is not on the most direct route across Francia from Marseilles to Kent and there is the temptation to think they wanted a church dedicated to St Martin to exist once they arrived. That dedication may have been given, of course, by Bertha and her chaplain Bishop Liudhard to a still roofed Roman building on their own arrival c.580; 5 it then only takes a slight twist to make that dedication datable to the Roman period. The Seventh-Century Canterbury Churches The Cathedral: at least two Roman masonry structures are known to lie below the medieval cathedral. One, quoted on no good evidence, as being a temple lies below St Gabriel's Chapel at the east end (Rady 1994, p. 90) and the other was found in the extensive excavation undertaken in 1993 within the nave. However, it is very noticeable that the earliest Anglo-Saxon cathedral identified, presumably that built by St Augustine in the early seventh century, cut through a black earth deposit containing a pottery sherd of c.450-550 and this layer in turn covered the Roman layers (Blockley et al 1997, pp. 5, 12, 99, 211). The Anglo-Saxon church was also on a different alignment to the Roman street system and hence to any buildings fitting into that system. There is therefore no indication that the earliest church was formed from an old Roman building. St Martin's: the church of St Martin stands on the east side of the city some 600m outside the city wall and has long been regarded as that at which Bertha prayed. It is probably true that many archaeologists regard the earliest structural phase of this church as representing a Roman building (Fig. 1). The late Frank Jenkins was circumspect in his dating, pointing out that there was no conclusive proof of the 379 v,) 00 0 Tower 14th cenlllry 10 Broken tiles observed (noted on Uvett's plan of 1897) Anglo.Saxon nave (Routledge believed It to be Roman) Broken tiles obaerved by Livett c. 1896 Westward exlent of foondations not known Foundation observed by Routledge 1891 and Jenkins 1954 Extended chancel 12th/13th oenwry (ear1ier apse below suggested by RouUedge) Fig. I St Martin's Church Canterbury: summary of archaeological information based on various plans and texts. CHURCH ARCHAEOLOGY 410-597: THE PROBLEMS OF CONTINUITY structure being Roman and that nothing of Roman date had been found by gravediggers (Jenkins 1965). Initially Tim Tatton-Brown favoured a fourth century date for the construction of the earliest phase and regarded this building as a possible late Roman Christian tomb or cella memoria (Tatton-Brown 1980, p. 14). Charles Thomas followed this reasoning but pointed out the cella or mausoleum might just as likely be pagan (Thomas 1985, p.170-172). In 1987 moving away from the view that the building was a religious or funerary structure, Jonathan Rady, supported by Tatton-Brown suggested that the remains were those of a Roman domestic building, possibly even a villa (Rady 1987, p.124, 129, 201 ), although its proximity to Canterbury perhaps argues against it being a villa in the usually accepted sense. In 1897 Canon Routledge mentioned that the floor in the small room on the south side of the building was opus signinum (Routledge 1897, 11, 25). Although he gives no detailed description this floor appears to be in the region of 0.20m thick (Routledge 1 897, fig. facing p. 6 and plate between pp. 14-15). The writer knows of no single mortar floor in an Anglo-Saxon context of this sort of thickness. This floor, if correctly identified as 'genuine' opus signinum concrete would tend to reinforce a Roman date for the earliest structural phase. 6 As it appears to be exceptionally thick and as such concrete is waterproof this may indicate the copious use of water (see below). On the other hand the square headed south doorway (regarded as contemporary with the small southern room) with its megalithic lintel is usually regarded as classic Anglo-Saxon architecture. 7 Examination by Canon Routledge indicated that the foundations of the nave and chancel were of different construction and by implication of different dates (Routledge 1 891, p. 137). This factor could be used to suggest that the 'chancel' was constructed as a free standing structure (unlike Routledge, no one today seems to believe the nave is earlier). Although admittedly a timber building may have been associated with the 'chancel', its interpretation as a cell a memoria or a small temple does not seem inappropriate (see below). A cella memoria at this site would be similar to that which has been proposed for Stone Chapel near Ospringe. However, the writer has suggested that the latter was initially used as the shrine of a water deity for the stream that at one time (probably) flowed along the Newnham Valley (Ward 1997, p. 203; in preparation). It is here suggested that the same was true of the conjectured Roman structure forming the earliest phase of St Martin's Church. The building is 50m away from the nearest known Roman pagan burials (very few), but less than 20m away from a spring that regularly bubbles up and threatens to flood the History Department of Christ Church University College. 8 This spring flows more or less permanently through the garden of the 381 ALAN WARD adjacent property. Hasted also mentions the springs around the church.9 On a hill slope, above and visible from the town, with the springs from these hills probably providing the main water source for the urban population, a shrine to a water deity would not be out of place. As opus signinum is waterproof the southern room may have formed the shrine itself, with abundant water cascading over a figurine. However, there is a tendency for those who suggest a Roman date to ignore the foundation observed by Routledge ( 1897, pp. 5-7) extending westward below the floor of the nave (most notably by Thomas I 985 in his fig. 28: 1, p. I 85). Routledge initially suggested this foundation might represent a medieval parclose (a screen between nave and aisle). However, Grevile Livett regarded the observed foundation as earlier than the standing nave walls (Routledge 1897, 5-7; Jenkins 1965, 13) and contemporary with the chancel. He suggested this was reinforced by a broken face of tiles on the internal west end of the south wall of the chancel lining up with this foundation. He also noted broken tiles at the west end of the north wall of the chancel (this is not stated in the text but a broken face is noted on Routledge's plan). Based on Livett's observation most plans of the nave therefore show a northern partner to the southern foundation. The broken tiles on each side of the chancel may merely represent pilaster buttresses cut back to receive, at a later date, the one certain and one possible wall, now below the nave floor. However, if, as seems probable, the foundation observed below the nave floor is contemporary with the south wall of the chancel, the change in shape, from a simple square structure, does not militate against the structure being Roman, nor it being a temple. The later use of the assumed Roman building at St Martin's by Bertha, whether it be domestic, mausoleum or temple, may have more to do with it perhaps still being roofed, or at least its walls still standing so they could easily then be thatched, than any identification with a religious structure. Much the same scenario, perhaps in the seventh century can be envisaged for Stone Chapel (Fletcher and Meates 1969, 1977; Taylor and Yonge 1981). St Pancras Chapel:as a further complication Charles Thomas has put forward the idea that the St Martin's Church of Bede was in fact St Pancras Chapel within the precinct of the Abbey of SS Peter and Paul (later St Augustine's Abbey; Thomas 1985, p. 172). Whilst the suggestion is not new the arguments in favour o f this interpretation are quite complex. It has been suggested that the earliest phase of this structure was a church of late fourth or early fifth century date and was the building used by Bertha and Augustine. That there are two early structural phases separated by an unknown length of time there 382 CHURCH ARCHAEOLOGY 410-597: THE PROBLEMS OF CONTINUITY is no doubt (Jenkins 1976a, l 976b; Taylor and Taylor 1965, pp. 146- 148; Hope J 902, 23 l -232), but the phasing postulated by Hope has much (e.g. simplicity) to commend it. Several points are worthy of note in regard to the stratigraphy of this interesting site. First there is no mention of demolition or construction deposits above or below the 0.22m of soil that had supposedly accumulated between the Phase 1 and Phase 2 church (Jenkins 1976a, 163; 1976b, 4). Jenkins argued that demolition was shown by the irregular height of the Phase 1 walls. However, such a destruction depends on where the irregularities were actually situated. If the corners were higher than the body of the walls this could merely indicate construction of the strongest parts of the building first. Alternatively the explanation could be as simple as variation within the mix of mortar, itself dependent on the materials to hand or even the mood of the labourers. The apparent absence of demolition material may indicate the Phase l structure was in fact never completed or, perhaps more likely, as Hope suggests, the break was of short duration (Hope 1902, 232). The first two structural phases of the church are equated by Jenkins with two distinct floor levels, Phase 1 with a clay floor and Phase 2 with a concrete floor, the two being separated by the soil deposit mentioned above. However, there is nothing within the interim reports to show why the concrete floor could not be associated with the Phase 3 (Norman) structure. If that is correct, again there appears to be nothing to refute Hope's view that the walls bonded by yellow mortar and white mortar followed on more or less directly from one another. It is noticeable that Jenkins does not mention a floor relating to the Norman church (nor Hope or Routledge for that matter, see below). It is debatable as to what is meant by 'concrete' (Hope uses the word cement). If it could be shown that this was 'genuine' opus signinum, then concrete is the correct term and a Roman date would be assured. If 'pseudo' op. sig. then a much later date can be envisaged. However, Hope irrefutably demonstrated that it was not the genuine material, for he tells us that the floor was of 'white cement 6 inches thick ... with a surface coat of pinkish colour .... so thin as to be readily scratched (away) ... ' (Hope 1902, 232; writer's emphasis) and hence perhaps similar to the c.1133 floor at St. Gregory's Priory (see endnote 6). He goes on to tell us that, upon this floor in medieval times tiles 4.5in. square were laid. Within the south porticus Routledge tells us the tiles were laid upon an 8in. layer of brickearth (which then explains the step up from the nave) and this in turn was laid upon the concrete which he believed was scorched (red?) by fire (Routledge 1882b, 106). 10 Routledge regarded the tiles as being late fourteenth or early fifteenth century in date (1882b, 104) and there appears to be no reason to disagree with that date range. 383 ALAN WARD Within the nave the tiles were laid directly on the cement/concrete floor of the Phase 2 church. If we follow Charles Thomas' reasoning and Frank Jenkins' phasing then the concrete floor must belong to the seventh century. The question then has to be asked where are the floors and bedding deposits of the eighth (let's be generous and say ninth), through to the fourteenth century? Where is the build-up and patching that would normally be expected for a five hundred year time span? An alternative interpretation is to accept there was a short break in construction of the Phase I building and that Phase 2 saw the completion of the church with a clay floor. The 0.22m of soil then belongs either to a period of abandonment during the course of the eighth to eleventh centuries, or alternatively itself formed a build-up deposit for a destroyed earth floor, or (less likely) the bedding for the concrete floor above. The concrete (or perhaps more correctly mortar) floor, whether having a red surface or not, can then be associated with the Norman Phase 3 structure. If this alternative interpretation were to be accepted, the stratigraphic problem and interesting, but unprovable philosophical debate concerning this structure, then disappears. Until any surviving section drawings and elevation drawings of the walls are published, which after this length of time seems unlikely, there seems little likelihood of solving the chronological problems (and perhaps not even then). Lulling stone Let us turn to a surprisingly neglected example at the other end of the county. Lullingstone outwardly appears to have the potential for Christian religious continuity from the Roman period into the Anglo-Saxon. The Christian motifs within the villa are well known but archaeologists have kept remarkably quiet about the late AngloSaxon church of St John the Baptist on the site of the TempleMausoleum (Fig. 2). Little was written about the material remains of the church within the excavation report (Meates 1979, pp. 19, 123-4), although admittedly this may be because little was found. Meates was wise enough not to specifically state that there had been continuity of Christian worship on the site, but that idea is perhaps (as so often in archaeological reports) implanted into the mind of the reader. He suggests that a folk memory of either the pagan temple, or Christian worship in the villa, may have persisted in the locality and that it might be an example of Pope Gregory's instruction to convert pagan temples into Christian churches. He implies this is supported by the church not being constructed on a west to east axis nor upon the point of sunrise on St John the Baptist day. 384 CHURCH ARCHAEOLOGY 410-597: THE PROBLEMS OF CONTINUITY Terrace Medieval bui􀀌􀀍 • / Roman building 0 10 20 Lulhngstane church 􀀆::􀀇;,um falq Ambulatory Circular shrine Cella 30 40 Terrace alls destroyed by land movement --o-Z Granary 50 metres Based on Figures 2 and 34 of ---­ The Lu/lingstone Roman Villa Volume 1: The Site River Daren! Fig. 2 Lullingstone Roman villa and medieval buildings: site plan. Let us look at the evidence as supplied by Meates rather than being influenced by conjecture. The Temple-Mausoleum had, according to Meates been partly demolished in the late fourth century and certainly the presence of chalk blocks and voussoirs (supposedly) of the eel/a structure within the backfill of the tomb robber pit implies such demolition. However, the presence of the pit also implies the roof was intact at the time of the robbing, for otherwise it would be difficult to find the grave below masses of rubble, and Meates himself 385 ALANWARD tells us the grave was noticeable due to the slumping of the floor (Meates 1979, p. 123). Overall, the stratified sequence within the robber pit implies a relatively orderly backfilling using different deposits of Roman material (see Meates Fig. 33, here much simplified as Fig. 3). This backfilling is unlikely to have been the work of tomb robbers. Why would they bother? These factors also reinforce the view that the roof was intact at the time of the tomb robbing. If the portrayal of chalk or tufa (more correctly travertine) blocks in the backfill of the robber pit is a true reflection of their size, they appear to be very small for regularly cut blocks that at one time formed roof vaulting or cella walls.11 The stone blocks may (we are back to conjecture) represent an internal structure within the cella. This would then explain the lack of large quantities of demolition material within the robber pit. It is here suggested the collapse of the vault may have taken place long after the robbing of the tomb and this collapsed material was what attracted the church builders to the site. The area on the section drawing shown as earth and rubble of medieval date may have been the area of the collapsed vault and formed an obvious attraction to later builders. The travertine and chalk would then be used in the church, and the hollow thus created subsequently levelled with soil. This seems a reasonable interpretation although the absence of any floor in this area, over the rubble and earth deposit and below the collapsed medieval church wall, does give some cause for concern. Presumably this absence represents robbing of the (tile?) floor of the church after its abandonment c.1412. Whilst there are other stratigraphic scenarios this appears to be the simplest. Perhaps significantly there appears to be little build-up within the interior of the church; Fig. 3 shows Med I, II and III layers. The earliest, Med I, appears to be a levelling deposit, possibly with a tile floor, Med II, forming its surface. Med III could be interpreted as yet more build-up or alternatively as a demolition deposit; certainly there appears to be no floor on its surface. Collapse of the walls occurred sometime after c.1412 when the parish was amalgamated with the Church of St Botolph some 300m to the south; ruins were still visible in the late eighteenth century (Macknelly 2000, p. 7). The impression gained from studying Meates' original section drawing is that the 'rubble' adjacent to the western wall of the eel/a is part of the sandy gravel and flint within the ambulatory. Whilst the fallen wall plaster could represent collapse in the Roman period, Med I appears to lie immediately above and then slumps down eastwards over the sandy gravel and perhaps the rubble. The latter is certainly earlier in date than the earth and rubble further east and hence could be Roman. The earth and rubble however, even without the skull, could reasonably be interpreted as a medieval levelling deposit. 386 w 00 -..J Geologlc:al natural 0 2 3 4 5 Fig. 3 Lulling􀁔tone Roman villa: Temple mau􀁕olcum 􀁕cction looking south. (Ba􀁕ed on fig. 33 in Mcarc,. G. W .. The l,11//i111.:.r1onc Noman Villa. volume I. 1979.) ALAN WARD Leaving aside for the moment the problem of where the skull came from, the presence of this levelling implies that the cella walls, or perhaps more likely a mass of rubble, were visible at the time of construction of the church. Meates contradicts himself on this point: 'little if anything is likely to have remained above ground when the Christian church came to be built' (Meates I 979, p. 124) but earlier (p. 19) we have 'by late Saxon times its foundations were probably visible.' From his section drawing alone it is not possible to say which is correct. However, the drawing does give the impression that the site was being terraced and levelled prior to the construction of the church (an activity which would hardly be considered surprising) for the soil level outside the church is considerably higher than the church interior. 12 The construction of the west wall of the church away from any of the Roman walls may indicate that the builders had started digging a trench prior to such terracing and hence did not know of the presence of the west wall of the ambulatory. This may give some credence to the view that no solid masonry was visible. Even the north and south walls merely overlap rather than being constructed directly upon the Roman masonry and these too may have been begun prior to finding the structure below. The first mention of the church is in 1115. Meates states this was 'suggestive of a Saxon origin'. The writer has pointed out (Ward 2000, 263) that a date in the early twelfth century allows a period of forty or fifty years since the events of 1066, surely enough time for a purely Norman foundation. Other than the (disarticulated?) human skull there is no hint of any earlier medieval use of the site before the stone church was constructed and the skull could easily have been in a feature dug after construction of the church. 13 This view is perhaps hinted at by the layer of stones above the skull. The Taylors visited the site in 1959 and subsequently pointed out that there was no structural evidence to support an Anglo-Saxon date, and that a Norman date was just as likely (Taylor and Taylor 1965, p. 402). For the three manors of Lullingstone mentioned within Domesday Book of 1086 no church is mentioned (Morgan 1983) nor within the Textus Roffensis of c. I 110. Whilst in no-way conclusive this perhaps hints that no church existed at those dates and that the date of 1115 represents the foundation of the church. Whilst there may have been a folk memory of Christian worship no evidence is supplied to support such an assertion, for the simple reason that there can be none. The oft-quoted statement of Gregory the Great about reusing pagan temples is often misunderstood. Gregory is writing in the context of Anglo-Saxon Germanic pagan temples, not temples constructed in the Roman period. The former, if they existed in the form of a building, would have been of timber, or, 388 CHURCH ARCHAEOLOGY 410-597: THE PROBLEMS OF CONTINUITY alternatively they may merely have been a sacred grove. Finally the church is no more off west-to-east alignment than many other churches. If a Norman date for the church is accepted (or even if the 'traditional' late Saxon date is retained) the letter of Gregory the Great has no bearing and folk memory survival for as much as 700 years seems unlikely. Those who would support the survival of such a folk memory have to explain how it could 'jump' at least one cultural change from the centralised, literate villa owning society of the late fourth/early fifth century to the pagan non-literate AngloSaxon kin based society of the late fifth and sixth century. This change probably involved a change in family ownership and over the following centuries perhaps several such changes. Perhaps most important there would also have to be a jump across the language barrier from Celtic/Latin to Germanic. Taking all of these negative factors into consideration it is here suggested that the building of the church took place after c. I 100 and had more to do with the availability of building material on a conspicuous position and alongside the existing manorial settlement of Lullingstane than any religious factor In this article several themes have been taken in an attempt to show the problems attached to the idea of church continuity, whether it be structural or spiritual. No doubt the temptation will be irresistible for archaeologists to identify fourth-century Roman structures as churches and to look for continuity of religion on, or adjacent to, those buildings. However, such identification will be difficult to show beyond reasonable doubt. An inscription can (unfortunately) hardly be expected. One or more chi-rho monograms inscribed or painted onto stone or plaster might show that Christians were present but do not in themselves identify a structure as a church. More convincing would be a baptistery, such as exists at Richborough, or a large number of burials aligned west to east without grave goods, which could be associated with a late Roman or sub-Roman structure (such as at Butts Road, Colchester). Anything else, be it orientation, shape of building, use into the fifth century, an Anglo-Saxon or early Norman church being constructed on or adjacent to a specific structure, or any combination are not sufficient evidence to adduce the presence of a late Roman church let alone continuity of religion. These factors merely represent the personal opinion of individual excavators. Such opinion may or may not be valid, but it is certainly not evidence. 389 ALAN WARD ACKNOWLEDGEMENTS The writer's thanks go to Christine Hodge for providing information from various internet sources, Pauline Sieben, Tim Allen and Sean Greenwood, former Head of History at Christ Church University College, Canterbury (all of whom have an imposing grasp of the English language), for reading and commenting on earlier drafts of this work. Thanks also go to Andrew Linklater and Christopher Sparey Green for putting forward many of the ideas and discussion points that appear in both the endnotes and the appendix. ENDNOTES 1 Whilst the most readily available translation of Bede states, 'On the east side of the city stood an old church, built in honour of St. Martin during the Roman occupation ... .' there have been others who have read the text in a slightly different manner, 'There was a church (now) dedicated to the honour of St. Martin, built of old when the Romans inhabited Britain' (Routledge 1891. p. 120). The word 'now' is Canon Routledge's entry and emphasis and he suggests this is what is meant. To the writer of this article the statement implies the opposite, i.e. that the church was built in honour of St Martin from the outset. Routledge also argued for a church foundation date in the mid-fourth century which no other archaeologist has been brave enough to support. Such a date would of course be perfectly acceptable for a late Roman pagan temple. 2 Paganism survived amongst some members of the aristocracy within Rome itself well into the fifth. if not the sixth, century. By implication some individuals or groups within the 'lower classes· would also have remained pagan. 3 The visit of St Germanus (c.429) is perhaps the most obvious piece of evidence for Christianity surviving (Bede 1974, pp. 58-65). lf we take the story literally, that 'most of the army sought Holy Baptism' (writer's emphasis) this suggests that up to that date much (if not most) of the population were still pagan. The writer finds it doubtful that Pelagian heretics were regarded as pagans (their presence being the outward reason for the visit of St Germanus). Heretics may have been at 'fault' and have to admit their 'sin' of having gone outside the universal Church but they would already have received Christian baptism. 4 He must have been canonised by the time that Gregory Bishop of Tours was writing c.580, but the nearest to a date from any dictionary on saints, is the ambiguous statement that after his death in 397 churches were ·soon' consecrated in his honour (Farmer 1987, p. 288). lf the tradition that Ninian dedicated his church at Whithorn to his mentor St Martin is true then sainthood must have been bestowed by 432 (the date of St Ninian's death). Apparently many early saints were never officially canonised by the Vatican, but became saints because of local factors and were rendered religious honour soon after death. Again tradition has it thot St Brice, successor to Martin, had a chapel built over the latter's tomb. This in itself may have been enough lo bestow sainthood. (Thanks go Lo Christine Hodge for nil of this information). Canon Routledge complicates matters by slatlng that St Martin had not been cnnonised in the last ten years of the Roman occupation (Routledge 1891, p. 121 ). Such o statement needs a reference; none is provided. There appears to be no good reason as to why St Martin should not be regarded as a 'local' saint und probably obtained sainthood more or less 390 CHURCH ARCHAEOLOGY 4I0-597: THE PROBLEMS OF CONTINUITY as soon as he was buried. A process similar in form to the deification of the 'good' pagan emperors within a matter of days or weeks of their death. 5 There is at least one internet reference which states that Luidhard and Bertha rededicated the 'church' to St Martin. In 1882 Canon Routledge also used the term 're-dedicated' (Routledge 1882a, 108), but there appears to be no early documentary evidence for this assertion. 6 We enter into a difficult area in regard to 'concrete' floors of seventh-century and later churches. If we ignore what may be the ·genuine' opus sig11i11um floor found at St Martin's. some of the seventh-century churches, whilst they have floors constructed in a Roman manner appear to be. what is here called, 'pseudo' opus sig11inum. Reculver probably provides the best local example. The floors appear to have had only a surface which was pink in colour, 'a red polished surface of cement and pounded brick'; the underlying material is merely described as 'mortar' (Taylor and Taylor I 965, pp. 507-8). This was true for both the first phase floor of 669 and the (?eighth-century) second structural phase. At Monkwearmouth and Jarrow 674 and 682 respectively the early floors are stated as being opus sig11inum (Cramp I 969, p.22-3). However. the overall description of the material at both monasteries appears to indicate it is not the same as that of the Roman period. For Monkwearmouth, ' ... mortar about I in. thick with a very finely powdered brick surface. It seems rather fragile as a flooring material ... '(Cramp 1969, p.36). For Jarrow, 'The floor in the opus sig11ium technique, was of concrete with large pebble agglomerates laid on a bed of small stones and faced with powdered brick, and had an average thickness of 2in. a markedly stronger construction than al Monkwearmouth' (Cramp 1969, p.45). Both these statements imply it is the surface alone which is red. due to the inclusion of powdered brick. Although at Jarrow • ... the brick admixture goes right through ... • (Cramp 1969, p. 37) this does not negate the observation, for brick rubble as such does not make the concrete pink, it is the powdered material which creates this effect. At SS Peter and Paul, Canterbury three floors were identified in the excavations of 1955-57 (Saunders 1978. 46). The lowest was a thin concreted brick-chip floor and although not stated presumably dates form the seventh century. A dark concreted red brick-chip floor lay in the middle, with a pink floor 2in. thick and containing chalk and brick 'resembling opus signinum' forming the latest surface. This appears to be the closest to' genuine' opus sig11i11um and (assuming truncation had not taken place prior to the construction of the Norman church with its floors c.0.90m above) presumably dates from the tenth or eleventh century. A late floor of this type, red tinged in colour with crushed brick in its surface dating as late as c.1133 was found in excavations at St Gregory's Priory, Canterbury in 1990 (Hicks and Hicks 200 I, p. 21 and personal knowledge). At the time of completing this article the writer had not seen the latest KARU volume The Discovery and Excavation of Anglo-Saxon Dover by Brian Philp. The advert for this volume states that for the ninth century phase of St. Martin's Church. Dover a stone base had a 'marbled screed floor of opus sig11i11um', a description which seems to con form to the above. 7 We can see however, that early Anglo-Saxon churches copied Roman detail. and within Roman buildings timber and presumably stone lintels would have been used. Therefore, as such, the lintel does not confirm an Anglo-Saxon date. If Roman then this is probably the only surviving complete Roman doorway in the country. The difference in wall construction between the neatly coursed ragstone. travertine (tufa) and tile at Stone Chapel, nelll" Ospringe, definitely dated to the Roman period, may be a more telling point in favour of an Anglo-Saxon date for the south chancel wall at St Martin's which is of a shoddier build. However, not all Roman buildings are going to look the same and of course many may have at one time been rendered. Indeed 391 ALAN WARD it could be argued that support for a Roman date is provided by the same building. The threshold of the eel/a at Stone Chapel is a large Greensand block, the same as at St Martin's and the excavators state this would have 'supported the monolithic jambs and lintel' (Fletcher and Meates 1969, 278). The jambs have probably been found incorporated into part of the medieval church (Taylor and Yonge 1981, 129-130). 8 The eel/a at Stone Chapel, often referred to as a mausoleum is, with one exception, 375m distant from the nearest known Roman burials. Despite statements which imply the contrary, 110 evidence for burials was found within this room during the 1967-68 excavations. The Roman opus signinum floor had been cut by a feature described as 'perhaps originally for a burial', although none was found (Fletcher and Meates I 969, 279). Although in both text and photograph caption they say a clay floor overlay the pit, their fig. 4 shows this feature as also culfing the (assumed Anglo-Saxon) clay floor. The fact that this, impossibly shaped {as drawn), feature cuts the Roman opus signinum floor shows that, at least, initially the structure was not a mausoleum. Furthermore the shape of the feature and apparent lack of any human remains suggest it is not a grave and the confusion regarding its position in the stratigraphic sequence indicates it may not even be Roman. Whether there was any evidence for burials found during the Kent Archaeological Society excavation of 1872 or that of 1926 will never be known. With the one exception of a child burial to the south of the structure there appears to be a lack of Roman burials in the immediate area and the evidence (albeit slight) for a stream along the valley adjacent to this structure suggest to the writer that a small temple to a water deity is a distinct possibility. Admittedly more extensive excavation of the area around may revise the number of late Roman or sub-Roman burials. 9 'It is remarkable that though this church is situated on an eminence some way up the hill, yet that part of it on the north side of the alley, which leads through the midst of it, is flowed with water from the springs, almost as high up as the floors of the pews, whilst that part on the south side is remarkably dry for several feet deep'. (Hasted J 801, p. 282). 10 Routledge later states it was op. sig. (Routledge 1897, 25). However, he also men· tions there were traces of burnt earth and other materials throughout the excavation (Routledge 1882b, 106). Unfortunately he omits to give the stratigraphic position of this burning which is not mentioned by any other writer. However, whether scorched or with a surface coat of crushed brick this floor could easily be as late as the first half of the twelfth century. 11 The blocks appear to measure c. 120 x c.40mm (5in. x 2in.). The latter measurement seems to the writer to be far too small. The smallest squared wall blocks at Stone Chapel measure 190 x 130mm for the travertine and 130 x 100mm for the ragstone. 12 Alternatively of course the west wall could have acted as a dam for soil creeping down the hill slope. 13 On page 123 of the report (Meates 1979) an inhumation, implying a complete burial, is mentioned within the east end of the church. However, on the plan (his fig. 31) and section drawing (his fig. 33) only a skull is mentioned, with no sign of a grave cut on either. The present writer has therefore assumed the skull was disarticulated; even if this assumption is incorrect it does not alter the point being made. REFERENCES Bede, 1974, A History of the English Church and People, ed. L. Sherley-Price. Blockley. K., Sparks, M. and Tatton-Brown, T., 1997, Canterbury Cathedral Nave Archaeology, History and Architecture. Brooks, N., 1984, The Early History of the Church of Canterbury. 392 CHURCH ARCHAEOLOGY 410-597: THE PROBLEMS OF CONTINUITY Cramp, R., 1969, 'Excavations at the Saxon Monastic Sites of Wearmouth and Jarrow, Co. Durham: an Interim Report, Medieval Archaeology, vol. xiii, 21-66. Dark, K., 2000, Britain and the end of the Roman Empire. Detsicas, A., 1974, 'Excavations at Eccles, 1973', Archaeologia Cantiana, LXXXIX, 119-134. Detsicas, A., 1975, 'Excavations at Eccles, 1974', Archaeologia Cantiana, XCI, 41-45. Ekwall, E., I 960, The Concise Oxford Dictionary of English Place-Names. Farmer, D.H., 1987, The Oxford Dictionary of Saints. Fletcher, E. and Meates, G.W., 1969, 'The Ruined Church of Stone-byFaversham', The Antiquaries Journal, vol. xlix, 273-294. Fletcher, E. and Meates, G.W., 1977, 'The Ruined Church of Stone-byFaversham: Second Report', The Antiquaries Journal, vol. /vii, 67-72. Glover, J., I 976, The Place Names of Kent. Hasted, E., I 801, The History of the Ancient and Metropolitan City of Canterbury, vol. 1. Hicks, M. and Hicks A., 2001, St Gregory's Priory, Northgate, Canterbury. Higham, N., l 992, Britain, Rome and the Anglo-Saxons. Hope, W.H.St. J., 1902, 'Excavations at St Austin's Abbey Canterbury - I: The Chapel of St Pancras', Archaeologia Cantiana, xxv, 222-237. Jenkins, F., 1965, 'St. Martin's Church at Canterbury: A Survey of the Earliest Structural Features', Medieval Archaeology, vol. 9, 12-15. Jenkins, F., I 976a, 'Medieval Britain in I 975; Canterbury: St Pancras Church, Medieval Archaeology, vol. 20, 163-164. Jenkins, F., 1976b, 'The Church of St. Pancras at Canterbury', Canterbury Archaeology 1975/76, 4-5. Macknelly, G., 2000, A Short History of St. Botolph's Church, lullingstone. Mayr-Harting, H., 199 I, The Coming of Christianity to Anglo-Saxon England. Meates, G.W., 1979, The lullingstone Roman Villa. Morgan, P. (ed.), 1983, Domesday: Kent. Rady, J., 1987, 'Excavations at St. Martin's Hill, Canterbury, 1984-85', Archaeologia Cantiana, CIV, 123-218. Rady, J., 1990, 'Excavations around St Gabriel's Chapel, Canterbury Cathedral', in Driver, J.C., Rady, J. and Sparks, M., Excavations in the Cathedral Precincts, 2 linacre Garden, Meister Omers' and St Gabriel's Chapel. Routledge, C.F., 1882a, 'St. Martin's Church, Canterbury', Archaeologia Cantiana, XIV, 108-112. Routledge, C.F., I 882b, 'Roman Foundations at St. Pancras, Canterbury', Archaeologia Cantiana, XIV, 103-107. Routledge, C.F., I 89 I, St. Martin's Church, Canterbury. Routledge, C.F., 1897, 'St. Martin's Church, Canterbury', Archaeologia Cantiana, XXII, 1-28. Saunders, A.D., 1978, 'Excavations in the Church of St Augustine's Abbey, Canterbury 1955-58', Medieval Archaeology, vol. xxii, 25-63. 393 ALAN WARD Tatton-Brown, T., 1980, 'St. Martin's Church in the 6th and 7th Centuries', in The Parish of St. Martin and St. Paul Canterbury (ed.) Margaret Sparks, pp. 12-18. Taylor, H.M. and Taylor J., 1965, Anglo-Saxon Architecture, Volumes 1 and 2. Taylor, H.M. and Yonge, D.D., 1981, 'The Ruined Church at Stone-by- Faversham', The Archaeological Journal, vol. 138, 118-45. Thomas, C., 1985, Christianity in Roman Britain to AD 500. Wallenberg. J.K., 1931, Kentish Place-Names. Ward, A., I 997, 'The Roman Site at Ospringe'. Kent Archaeological Review, 129, 199-205. Ward, A., 2000, 'St Mary Magdalene's Chapel, Lidsing', Archaeologia Cantiana, CXX, 261-276. Ward, A. (in preparation), Fieldwalking at Hartlip Roman Villa. APPENDIX: THE DATING OF ST MARTIN'S CHURCH For an earlier summary of the dating the reader is referred to Archaeologia Cantiana 1897. Evidence for a Roman date: a. The tradition that Bertha and Luidhard used an old Roman church for worship. That they would have constructed a mortared stone building from scratch whilst possible is perhaps unlikely, for two reasons. First, it would mean bringing over masons from Francia. Secondly and more important (as identified by Routledge in 1897, p.18) any Kentish Christian tradition about Bertha building the first church since the Romans should (in theory) have survived and been passed on to Bede. For such an act she may even have obtained sainthood. However, it needs to be stated that Bede was notoriously reluctant to mention any churches which were the result of royal foundation. To the writer (in this instance) tradition provides a strong, but not conclusive, argument in favour for the earliest structure (whatever its function) on the site being Roman. b. The apparent 'genuine' opus signium floor within the small south room. c. The apparent similarity of ground plan of the 'chancel' to a Roman cella, or mausoleum or small temple. However, the foundation extending westward below the nave floor needs to be taken into consideration. 394 CHURCH ARCHAEOLOGY 410-597: THE PROBLEMS OF CONTINUITY d. The possibility that the south 'room' was in fact a free standing baptistery with an associated Roman timber church such as at Richborough Roman fort. Bertha and Luidhard could then easily have built a new timber church in association with the stone structure. However, Routledge, quoting Canon Grevile Livett, makes a convincing case for the contemporaneity of south room and chancel (Routledge 1897, p.11-12). It seems unlikely therefore that the south room, internally just 1 .45m east to west and perhaps much the same north to south, stood as a low-walled, open, free standing structure. Christopher Sparey-Green has also pointed out that all the known or suspected Roman baptisteries in this country are considerably smaller in size. Evidence for an Anglo-Saxon date: e. The apparent total absence of Roman archaeology in the immediate area but the abundant presence of Anglo-Saxon material (admittedly of the eighth and ninth centuries; Rady 1987). However, a tessellated pavement was found on an 'adjacent' part of St. Martin's Hill in the seventeenth century (Routledge I 891 , p. 127). Unfortunately we do not know how close this adjacent part of the hill was to the church. f. The standing Anglo-Saxon fabric (although Routledge even argued for a Roman date for the nave; Routledge 1897, p.25-26). g. The shoddier build of the south chancel wall when compared with the Roman walls of Stone Chapel. However, Routledge points out that not all Roman walls will be the same and some may have been rendered. h. The possibility that the floor in the smalJ southern room is in fact 'pseudo' op. sig. i. The occurrence of a structural feature found in the cathedral excavations of 1993 (Blockley et al 1997, p. 18, fig. I 0) standing in much the same position to the cathedral as does the small southern room to St. Martin's Church. This feature was lined with a skim of opus signinum just J0-15mm thick, which we can see the AngloSaxons were capable of creating. The suggested date for this feature was the early ninth century and that its function was that of a mausoleum; the occurrence of a black pigment and lime-scale however, may argue against this interpretation and perhaps shows the presence of some form of liquid (water for a baptistery?). 395