05 February, 2010

Wroxeter: the sixth-century rebuilding

In earlier posts on Cynddylan and a possible Bishop of Chester in the post-Roman period, I mentioned the Roman town of Wroxeter in modern Shropshire (Roman name Viroconium). Archaeological excavation on the site of the baths basilica has shown that large-scale building was undertaken there at some time in the mid- to late sixth century (White and Barker 2002). What form did this take?

The baths basilica

Roman Viroconium (Wroxeter) was the civitas capital of the Cornovii, whose territory covered roughly the area of modern Cheshire and Shropshire. Like any self-respecting, prosperous and important Roman town, Wroxeter had a large public baths complex, approximately analogous to something like a modern leisure centre. In Wroxeter the baths complex was in the middle of the town, and attached to it was a large exercise hall, the baths basilica. The baths basilica was 74 m long by about 20 m wide, with its long axis oriented roughly east-west. The west side (one of the short sides) opened onto the main road, Watling Street, and a side-street ran along the full north side of the basilica. To the south of the basilica was the baths complex itself and a market hall. The complex was built around 150 AD.

Fast-forward to the late fourth century. The entire baths basilica was refloored some time after 375, dated by a coin of Gratian (367-375) found under the new floor (White and Barker 2002, p. 115). This new floor had been repaired and renewed twice more, indicating substantial usage and maintenance over an unknown, but considerable, period of time. The baths complex may have been the focus of the town during this period, replacing the city forum which had been destroyed by fire at around the early fourth century and not rebuilt (White and Barker 2002, p. 112-116).

The baths basilica continued in use until around the end of the fifth century, getting steadily more dilapidated, as deduced from the successive refloorings (White and Barker 2002, p. 119). Then the building was apparently used as a builders’ yard (perhaps by a maintenance team keeping the adjacent baths building in repair), and a large bread oven was built outside the west door. A magnetic date derived from the remains of the bread oven gave a date range for its last firing of 490-550 (White and Barker 2002, p. 121).

At some point after this, the roof of the baths basilica was taken down and its internal columns removed. A new floor surface made of re-used roofing slates was laid in the interior of the building, which was now presumably something like a walled open-air courtyard, and the bread oven outside the west entrance went out of use (White and Barker 2002, p. 120-121). The new slate floor showed considerable wear. The interpretation is that the shell of the baths basilica was in use by large numbers of people, perhaps as an open-air market (White and Barker 2002, p.121).

The great rebuilding

The next activity on the baths basilica site was to demolish most of the remaining north wall of the baths basilica to ground level. Part of the adjacent Roman street that had run along the north wall was dug out and relaid as a new gravel street.

A series of building platforms was constructed in what had been the interior of the baths basilica, made from carefully sorted and laid demolition rubble. The largest (Building 10) occupied the middle of the north side of the basilica, exactly opposite the doorway in the middle of the south wall. Parts of this building platform were composed of painted plaster from a curved ceiling, probably from one of the vaulted heated rooms in the adjacent baths complex, indicating that the baths had gone out of use and been at least partially demolished. The plaster was broken into pieces of similar size, which is consistent with the room it came from having been demolished at or shortly before the time of the rebuilding (rather than the plaster having been scavenged from a long-derelict site). The rubble platform was 33.5 m by 15.6 m, with two short square projections on the corners on the south side. Between these projections, exactly opposite the doorway in the middle of the south wall of the baths basilica, a column drum and column capital had been placed 3.2 m apart. These stones probably supported a columned porch with a veranda behind, since a narrow rectangular area immediately behind the stones had an untrampled surface, consistent with it having been planked over. The solidity of the building platform suggested that it had supported a very substantial structure, interpreted as a two-storey timber building with towers flanking a columned entrance porch (White and Barker 2002 p. 123-125).

It’s worth noting that this impressive building has its back to the newly-laid gravel street north of the old baths basilica, and that its monumental columned-and-towered entrance way looks out onto what had been the interior face of the south wall of the old basilica. The positioning of the column bases exactly opposite the position of the doorway in the middle of the south wall suggests that the south wall still stood to some height at the time of construction. It is also worth noting that the cold room (frigidarium) of the baths complex stood immediately behind the basilica south wall to the east of the doorway. Given that some of the frigidarium walls are still standing to this day (known as the ‘Old Work’), they must still have been standing at the time of the rebuilding, and the frigidarium may well still have been intact and roofed.

Another, somewhat smaller building platform (Building 11) was laid between the big building and the west wall of the baths basilica, and another was laid at the east end of the former baths basilica interior, more or less opposite the frigidarium. Six more building platforms were laid out along what had been the interior face of the baths basilica south wall, between the west wall and the doorway in the middle of the south wall (so these face across the remains of the interior towards the columned entrance fa├žade of the big building and the building to its west). Four of these platforms were 28 Roman feet long, and the two others were 29 and 27 Roman feet long. This is consistent with them having been laid out by someone using a Roman system of measurements (White and Barker 2002 p. 124-125).

On the north side of the new gravel street, north of what had been the baths basilica, there were six or seven buildings of 20 or 24 Roman feet in length. These were later replaced by a set of five buildings, also built to Roman measurements (White and Barker 2002 p. 127).

Along the south side of the new gravel street there were several small lightweight structures, too small to be houses and interpreted as market stalls or similar (White and Barker 2002 p. 127). If the shell of the baths basilica had indeed been used as an open-air market before the rebuilding, these structures may have been continuing its function

Outside what had been the west wall of the baths basilica there were two building platforms, one blocking the former west doorway and one on the corner of the new gravel street. Three successive pairs of these buildings were erected on more or less the same sites (White and Barker 2002 p. 127-128). In all, 33 separate buildings constructed during this phase were identified in and around the baths basilica site (White and Barker 2002 p. 128).

Dating the ‘great rebuilding’

The exact dates and duration of this ‘great rebuilding’ phase are not known, but it is book-ended by two pieces of dating evidence:

  • it happened after the bread oven at the west entrance went out of use (last firing dated to 490-550)

  • it happened before a grave was cut through the platform of Building 11, radiocarbon dated to 600-790 (White and Barker 2002, p. 136)

Taking the mid-points of these two book-end dates, the ‘great rebuilding’ falls into the period between around 520 and around 700 AD, in other words some time in the later sixth and seventh centuries. Since there was time to replace the buildings along the north side of the new gravel street at least once, and to replace the buildings outside the former west entrance at least twice, the phase clearly lasted quite a long time. White and Barker suggest about 75 years (p. 136). It may have been longer, since timber buildings can have lives considerably longer than 25 years; it’s hard to see that it can have been much less, though of course parts of the site may have gone out of use earlier or later than others.

Who was responsible?

Alas, no helpful inscriptions along the lines of “Built by X, who held the position of Y, in the year Z” came to light in the excavations, so the identity of the individual or authority who was responsible for this massive redevelopment project in sixth/seventh century Wroxeter is unknown. Some inferences can be drawn:

  • whoever was responsible had sufficient authority to take over and redevelop the whole area of the baths basilica, and presumably also had the authority to terminate whatever activity had previously been going on in the shell of the buildings (i.e. the postulated open-air market);

  • if the demolition of the heated rooms in the baths complex happened at the same time, this authority extended to the baths complex as well;

  • whoever was responsible controlled substantial resources, since digging up and relaying the street and constructing the building platforms and the buildings they supported (especially Building 10) would have required considerable amounts of labour. If the buildings were massive timber-framed constructions, as seems likely from the platforms, access to large quantities of building timber would also be required;

  • whoever was responsible had some organisational ability, since managing a construction project of this size is not a trivial task, and the arrangement of the buildings is consistent with the site having been planned in advance;

  • whoever did the site surveying and layout used Roman measurements (though not Roman building techniques);

  • whoever ordered the construction of Building 10 wanted an imposing and important-looking building, and so was presumably important (or at least wished to convey the impression of being important);

  • whoever ordered the construction of Building 10 was focussed not on the new gravel street at the back of the building and any commercial activity taking place on it, but on the former interior of the baths basilica, since this is the orientation of the impressive columned entrance;

  • the deliberate placing of a building to block the west entrance to the former basilica is consistent with the former interior being used as a restricted or private space, presumably for the use of the occupant of Building 10 and its immediate neighbours, like a sort of private close or courtyard.

All these are consistent with the rebuilding having been ordered and controlled by the ruler of the local area. The most obvious candidate for a local ruler is a king, and it is possible that Wroxeter was the seat of a king or sub-king of Powys or a territory based on the territory of the Cornovii. Cynddylan, who according to Welsh poetry was a powerful ruler in the first half of the seventh century, would be contemporary with the later part of the date range for the ‘great rebuilding’, and he or his predecessor(s) must be a candidate for the ruler of Wroxeter. If Cynddylan died in 655 at the Battle of Winwaed (see earlier post on Cynddylan for the rationale), his death could also coincide with the approximate end-date of the baths basilica site. It should be noted that there is no evidence at all of destruction at any of the buildings on the site, so if “The court of Pengwern is a raging fire” as the Canu Heledd poetry has it, either the court wasn’t at the baths basilica site in Wroxeter, or the evidence for its destruction hasn’t survived, or a certain amount of poetic licence was being applied.

A local secular ruler need not have been a king, although the distinction may be largely one of semantics. Gildas says:
“Britain has kings, but they are tyrants; she has judges, but unrighteous ones”
--Gildas, De Excidio (On the Ruin and Conquest of Britain) III, 27, available online

Ken Dark argues that some sort of bureaucratic government by officials, rather than royal government by kings, existed in parts of west-central Britain, including the Wroxeter area, in the fifth and sixth centuries (Dark 2000, p.147-149). The ‘judges’ mentioned by Gildas could be rulers of this type, perhaps derived from the remnants of Roman civitas councils. Such a ruler may have been responsible for rebuilding Wroxeter’s baths basilica site (though how much he differed from a king, except in title, is open to question).

Another possible candidate for a local ruler may be a bishop. Cities run by bishops existed in fifth-century Gaul, where Sidonius Apollinaris, bishop of what is now Clermont-Ferrand in the Auvergne, governed his city and negotiated with the local king (of the Goths) in 470-480. The writing styli found on the site (Dark 2000, p.142), may be a pointer in favour of ecclesiastical occupation, although I personally wouldn’t assume automatically that secular aristocratic society was not literate.

White and Barker suggest that the frigidarium of the baths complex, which was diagonally opposite the grand front entrance of the new Building 10, may have been in use as a church (p. 125-126). It was in more or less the correct east-west orientation and would have had a plunge pool for use in baptisms. Built in stone in the Roman fashion, and perhaps still retaining painted wall plaster and decorations, it would have been an obviously ‘Roman’-looking building and perhaps therefore suitable for use by a religion that had been strongly associated with Rome and Roman government in Britain. A dozen burials found in the surrounding hypocaust in the nineteenth century would be consistent with a graveyard around a church*.

The use of Roman measurements for the building layout also suggests an engineer or surveyor familiar with Roman techniques. While such skills could have been handed down locally in a secular context, they could also indicate access to Roman knowledge handed down in written form through the church, or clerical contact with craftsmen from continental Europe.

These possibilities may not be mutually exclusive. There seems no obvious reason why a bishop might not also have been one of Gildas’ ‘judges’, especially if such non-royal government was indeed a distant descendant of the Roman civitas council.

A bishop could also have had close links with a royal dynasty. Several early medieval churchmen had aristocratic or royal connections. For example, Sidonius Apollinaris and Gregory of Tours (late sixth century) in Gaul were both members of the aristocracy, and Sidonius was married to a relative of the emperor. Saint Columba (late sixth century) was descended from Irish royalty. Gildas says that his contemporary King Maglocunus (usually identified with Maelgwn of Gwynedd, died mid-sixth century) became a monk for a while before becoming king (De Excidio III, 34-37). A hypothetical Bishop of Wroxeter could have been equally well connected, and if so may have been the brother or nephew or cousin of a local king.

Dark K. Britain and the end of the Roman Empire. Tempus, 2000, ISBN 0-7524-2532-3.
Gildas, De Excidio (On the Ruin and Conquest of Britain), available online
White R, Barker P. Wroxeter: Life and death of a Roman city. Tempus, 2002, ISBN 978-0-7524-1409-7.

Map links
Wroxeter at 1:25,000 scale, showing the location of the modern village in relation to the Roman town wall and the surviving remains of the Roman baths complex

*These burials were luridly interpreted by the Victorian excavator as the wretched remnants of the city’s population who had taken refuge from barbarian attack by hiding in the hypocaust and perished there miserably.


Kathryn said...

What a fantastically detailed post, Carla. I'd love to respond with some intelligent and thought-provoking comment, but unfortunately I'll just have to settle for saying how much I enjoyed reading the post. :-)

tenthmedieval said...

As usual a masterful summary of quite a lot of disparate information, Carla! I think I prefer the episcopal idea, which is not to say that kings wouldn't have visited of course. All the same, I think a king's court in that region would more usually be based on a fort, no matter how Romanised he might be (and the poetry you quote shows exactly why!). I quite agree, though, that a bishop might also be a judge of almost Old Testament significance.

Carla said...

Kathryn - thank you! I'm glad you enjoyed it, and thanks for taking the trouble to say so.

Tenthmedieval - thank you, and credit to Barker and White for their very readable Tempus book on the subject. I like the episcopal idea as well. It seems to fit well with a town location, especially if the frigidarium had been pressed into service as a church. The former interior of the baths basilica may have been the first ever cathedral close :-)

Wroxeter had a town wall, so it could be considered a fort in that sense. If the complex lasted 75 years or more it would have had more than one occupant so a king could have succeeded a bishop or vice versa. Or a bishop might have been the local boss but subject in some way to a sort of high king, and a high king may well have travelled round his territory eating the food-rent and reminding everyone of his presence, so the bishop's residence could have also been the king's court from time to time. I rather suspect that ecclesiastical and royal power may have been closely connected.

Gabriele C. said...

Several Roman buildings in Trier survived because they had been used as churches, so it's not impossible the Wroxeter frigidarium was used the same way. Maybe the rest of the baths was either considered too un-Christian, or in bad repair and thus has been rebuilt.

Carla said...

Gabriele - something similar also happened in Leicester, where St Nicholas' church is on the site of the frigidarium of the Roman baths complex. The proximity of water supplies and/or plunge pools is handy for baptisms, so baths buildings would be among the logical candidates for conversion. The rest of the baths complex in Wroxeter was presumably demolished at or before the time of the redevelopment, given the freshly broken curved ceiling plaster forming the building platform for Building 10. Perhaps the vaulted roof had become too difficult to maintain, or had developed a structural fault that couldn't be repaired with the existing technology.

Rick said...

I only now checked and found this post! As an aside, the coin of Gratian once again shows how much of our archeological dating information is due to people who lost their money.

As I recall, earlier phases at Wroxeter have architecture of distinctly classical style, hinting at conscious romanitas. That earlier work had presumably fallen into a shambles by this time, but presumably with some kind of continuity.

I'd also lean toward episcopal administration, but I'm curious about Ken Dark's theory of 'bureaucratic officials.' That implies a remarkable degree of Roman survival (or revival?), or at any rate a sort of institutional complexity we don't usually associate with heroic ages.

Carla said...

Rick - indeed, though that's partly because coins are marked with the name of the issuer and their manufacture can thus be dated more closely than many other artefacts. Samian ware is nearly as precise in some periods, but with most other artefacts it's lucky to get within the right quarter-century or so. Usual caveat that a coin (or anything else) could be lost or deposited long after its date of manufacture. If people had been in the habit of losing Post-It notes with today's date jotted down on them, we'd use those for dating evidence instead :-)

Roman Wroxeter was a standard Roman town with all the expected trappings - forum, basilica, baths complex, posh houses, market hall, town wall etc. It was clearly built and maintained by people who wanted their town to look 'Roman'. I favour the episcopal theory too. As far as I can see it isn't incompatible with Dark's theory of administration by people who weren't called kings; the bishop could have been governor of the town and its hinterland, even if he wasn't called a king. Dark's suggestion is based on Gildas, partly the quote I cited in the post, which may imply that there were authorities other than kings, and partly because all the kings Gildas names can be located to the far west of Britain (roughly what's now Wales and the south-western peninsula), which may imply that there was some other form of government in the rest of the country. His thesis is broadly that post-Roman Britain has a lot in common with Late Antiquity in the rest of Europe, suggesting much more continuity from Roman customs than the destruction-by-invading-barbarians scenario usually allows for. Wroxeter is strong evidence in support, at least at that location (usual caveat that regions may have varied).

Rick said...

Dark's argument is intriguing, but then the question is when, how, and why did post Roman Britain finally evaporate?

And I gather there's a lively scholarly debate about all of this. A quick google turned up this contrary argument, by Neil Faulkner, that Roman Britain had fallen into a shambles and was pretty much gone before the legions left.

Stepping back, these people can't both be right, but they both presumably have plenty of evidence to point to. Faulkner's survey didn't make up all those towns and villas fallen into ruin before 400, but Wroxeter isn't made up, either.

Perhaps there was decline, then spotty revival - the 'great rebuilding' at Wroxeter could have been roughly contemporary with Justinian. (And might have looked no more 'Roman' to our eyes than Haghia Sophia does.)

Carla said...

Rick - the fact that respected scholars can disagree so fundamentally says something about the relative paucity of evidence :-) You're right that they can't both be right in all respects and for the whole country. One of the reasons I favour regional variation as a model is that it allows them both to be right in different places (and perhaps also at different times, if there were patchy local revivals as you suggest). Looking at Wroxeter, for example, the forum-basilica had burnt down in the late third/early fourth century and not been replaced, so if you focus on that the picture looks like Faulkner's view of a shambles full of ruins. Then you look at the Wroxeter baths basilica site a couple of streets away, which was maintained more or less intact and then replaced with this big redevelopment in timber in the sixth century, and the picture looks more like continuity or revival, albeit with a sharp change of architecture and no doubt of other cultural markers as well.

How you define continuity or collapse may also influence interpretation. If you define it in terms of Roman masonry buildings, then Wroxeter fits Faulkner's picture - the town evidently switched from maintaining two public building complexes (the forum-basilica and the baths complex) to only one (the baths complex) in the early fourth century, and then gave up on masonry altogether and switched to timber in the sixth. On the other hand, there was evidently still some sort of organising authority with access to skills and resources, so society didn't disintegrate into total chaos. Similarly at Birdoswald (coming up in the next post), where the buildings change radically but there was evidently still some sort of organising power. You can regard that as evidence for continuity, in the sense that there evidently wasn't total anarchy, or as evidence for collapse, in the sense that people evidently stopped building and maintaining Roman masonry architecture. Pays your money, takes your choice....

The rebuilding at Wroxeter is roughly contemporary with Justinian, give or take half a century or so. Also roughly contemporary with the Byzantine luxury goods that turn up at high-status sites along the west coast of Britain. For what it's worth, my feeling is that society became much more localised and fragmented after the end of Roman administration, and that the effects differed depending on who you were, where you were and who happened to be running the show locally.

Rick said...

We've discussed this before, that 'localized and fragmented' gets to the nitty gritty of revenue, taxes or tithes. A monk in a waddle hut can read and write books, but monumental architecture takes money, and maintaining troops takes even more.

Carla said...

Indeed. Building in stone needs a sufficiently complex economy to support specialist trades and transport over long distances, which would get more problematic as political units got smaller.

Rick said...

Whereas impressive work can be done in wood on a fairly modest budget. In this region of California there were some rather palatial hotels built in the 19th century, in towns of a few thousand people.

A regional chieftain might have a comparable resource base, and be able to put up a decently palatial palace. Unfortunately wooden buildings don't leave much evidence of their above ground appearance.

Most of the hotels I mentioned burned down within a few decades, and I'm surprised there aren't more indications of accidental fire destruction on wood building sites. (Or were Victorian hotels more at fire risk because of so many oil lamps?)

Carla said...

There are some; Yeavering was burned down twice, for example. I guess it depends on whether the archaeological traces of a fire (a) survive, (b) are excavated and (c) are recognised and understood. If a timber building sitting on top of a rubble platform burned down, it might leave little more than a heap of charred timbers and wood ash lying on top of the platform, which might be lost if the site was subsequently disturbed and/or redeveloped. Wroxeter is unusual in that the site isn't underneath a modern town and is relatively undisturbed, so for once there's a reasonable chance of telling the difference between evidence of absence and absence of evidence. Has anybody excavated any of your Californian hotels to see what traces survive?

Rick said...

Good question about whether the hotels have been excavated; I have no idea. Probably most of the sites have been rebuilt on, and modern foundation building must be very destructive of whatever was there before.

But this raises all sorts of interesting technical questions about archeology, such as under what conditions a fire leaves archeological traces.

Carla said...

Anything that involves digging foundations is bad news for any archaeology that isn't buried deeper than the foundations. Victorian cellars can be as destructive as modern buildings with deep groundworks.

Good question about the conditions that preserve traces of fire. You need an expert field archaeologist to answer that properly. The ones I know of are: deep-set timber posts or beams that leave charred remains in the bottom of the settings, which go down deep enough to escape subsequent ploughing or redevelopment etc; clay walls that get 'baked' in a fire and leave the baked remains in the bottom of wall foundation trenches; fire debris that's left on the surface, either as loose rubble or the stumps of walls, and either gets covered over by natural accumulation or where redevelopment amounts to just levelling the debris and then building on top without disturbing it; fire marks on the ground surface underneath a building that burned down (some clay soils will sort of bake to a red colour if the fire is hot enough) and then subsequent rebuilding just sweeps away any loose remains and builds on top of the burned ground surface without disturbing it. All of these depend on somebody not coming along with spades or a bulldozer, or a deep plough, before the archaeologists get there X centuries later.

A really bad fire of the 'destruction by fire and sword' variety certainly leaves an archaeological trail. Boudica's revolt regularly makes an appearance as a thick layer of burned clay and charred debris deep underneath London and Colchester.

Conversely, fire can presumably be ruled out if you have any surviving foundations where the timber is either still there (unusual except on waterlogged sites) or can be seen from the soil stains to have rotted out without burning. Stumps of unburned clay walls presumably also indicate absence of fire, or at least absence of fire serious enough to affect the walls. Logically, a surviving ground surface or floor with no apparent signs of burning or fire debris would indicate that there hadn't been a fire sufficiently fierce to affect the ground surface, although I don't know if it's possible for a fire to destroy the superstructure of a building without leaving any trace on the floor or ground surface.