Roman York to Anglian York: a speculative model

In this sequence of posts, I have summarised some documentary and archaeological evidence that may help to sketch out a picture of York in the post-Roman centuries. For a summary, see the preceding post in the series.

I have also discussed the Brittonic ruler Peredur, recorded in the late sixth century and associated with York in later medieval tradition.

I suggested that York continued to be inhabited, probably at a low density, and to be used at least on occasion by the local rulers during the fifth and sixth centuries. If correct, this could provide a mechanism for York to retain its status as a political, ecclesiastical and military centre, and possibly some of its cultural heritage, throughout the gap in the historical record.

However, the context in which such a status functioned clearly changed between the fourth and seventh centuries.  In the fourth century, York was under the control of Roman officials, part of the diocese of Britain and the Western Roman Empire, the base of a legion of regular army troops, and the seat of a bishop.  In the early seventh century, York was under the control of the early English (‘Anglo-Saxon’) king of Deira/ Northumbria, had no established bishopric (or at least not one that Pope Gregory recognised), and was not part of a political entity bigger than the kingdom of Northumbria (or possibly of whatever was represented by the title of ‘Bretwalda’, which on the most generous interpretation only extends to most, not all, of modern England and parts of southern Scotland).  How might this transition have come about?  There are several broad possibilities. 

Direct transfer of power from Rome to English kings

One possibility is that power was transferred directly from the last Roman officials to the first English king, some time after Emperor Honorius told the British civitates to ‘look to their own defences’ in the early fifth century. The Late Roman Army was in the habit of recruiting Germanic ‘barbarians’ as allies and mercenaries, some of whom reached positions of great power. Stilicho, the general and de facto Emperor in the early fifth century, had a Vandal father. In the 360s the Dux Britanniarum had a Germanic name, Fullofaudes, and in 300 a Germanic warrior-king called Crocus and his troops helped to elevate Constantine the Great to Emperor in York. If the Roman Army based at York in the early fifth century was either commanded by a Germanic general like Fullofaudes or Stilicho, or relied heavily on a Germanic mercenary ally like Crocus, it is not hard to see how such an individual could have become in effect the ruler of York and its surroundings, and effectively founded a kingdom with little more than a change of terminology. In this model one of the last Roman officials could also have been the first English king.

Something like this happened in parts of Continental Europe, where Germanic kings such as Clovis in Gaul (roughly modern France) effectively took over chunks of the former Western Roman Empire wholesale. However, such a direct transfer does not fit easily with some aspects of the situation in York. First, the Christian church hierarchy seems to have disappeared in York, or at least was no longer recognised by Rome, since Pope Gregory clearly expected to establish a new bishopric there. This contrasts with the situation in Gaul, where Christian bishops continued under Clovis and were recognised in Rome. Second, although Clovis was a Frank and Frankish was a Germanic language, the language that became dominant in his territory was a descendant of Latin (eventually evolving into modern French). By contrast, the language that emerged in York was English, a Germanic language. It is possible that a particular set of circumstances could explain both of these differences – e.g. if the leader who took over in York happened to be a committed pagan who chased out any Christian church hierarchy, or if Germanic languages were already widely spoken in and around York after generations of recruiting Germanic soldiers who retained connections with friends and relatives across the North Sea. Nevertheless, my interpretation is that these differences are consistent with a less direct transition in York.

Invasion and conquest

A second possibility is that English warriors invaded and conquered Roman York and its surrounding area, destroyed the Roman aristocracy and the Christian church, expelled or oppressed the Roman population, and established their own kingdom without reference to anything that had gone before.  Such a hypothetical conquest could have happened at any time between the early fifth century and the late sixth or early seventh century. This model explains the absence of a bishop in the early seventh century, and the presence of pagan English cremation cemeteries at The Mount and Heworth in York in the fifth-sixth century.

Given York’s importance in Roman Britain, one might have expected to find such a conquest recorded in Annales Cambriae alongside battles such as Arthuret, or as a comment on an early king of Deira in the king-lists in Historia Brittonum or the Anglo-Saxon Chronicle (‘X who conquered York’ like the enigmatic comment about Soemil), or to appear at least as a passing reference in poetry or the Triads. It may be possible that the reference to Soemil was supposed to mean that he had conquered York, expressed in an oblique way. It may also be possible that the un-located battle of Caer Greu mentioned in the Triads, where Peredur and his brother Gwrgi were killed (see post on Peredur), could be an oblique reference to a battle at York, called for some reason by the name of ‘Caer Greu’ instead of its more usual ‘Caer Ebrauc’ or ‘Caer Efrawg’.  It may also be possible that the enigmatic entry for the death of ‘Bishop Ebur’ in Annales Cambriae in 501 AD (see post on the documentary sources for post-Roman York) could be an oblique reference to an invasion and conquest that extinguished the bishopric. I am not convinced, because these are all very oblique; they could refer to a conquest of York, but that’s mostly because it isn’t clear what they refer to, so they could mean almost anything. It seems odd to me that there is no clear ‘X conquered Eboracum/ Caer Ebrauc’ or ‘X was killed in the battle of Eboracum /Caer Ebrauc’. However, the sources are so sparse that absence of evidence cannot be taken as evidence of absence.

Staged transition from Roman to a Brittonic kingdom to an English kingdom

A third possibility is that Roman York became an independent Brittonic kingdom that later became an English kingdom. Brittonic kingdoms are recorded in the late sixth and early seventh century in what is now northern England, such as the kingdom of Elmet in the area around modern Leeds, a few miles west of York (see post on Elmet), and the kingdom of Rheged somewhere in what is now north-western England and/or south-western Scotland (more on Rheged in a later post). If York was the centre of a similar kingdom, that would fit with the pattern.

Such a kingdom could have evolved from the military authority held by the late Roman army commander based at York (by a mechanism similar to the first possibility outlined above) and/or from the civilian political authority held by the leader(s) of the colonia. It could be similar to the situation at Birdoswald, where someone was building timber halls fit for a chieftain in the fifth to sixth centuries, presumably using the Roman fort for its defensive capability or its prestige or both. If some of the people in authority in Late Roman York were members of, or had close links with, the local British aristocracy, such a hypothetical post-Roman kingdom in York could have developed into a Brittonic tribal kingdom.

This model can accommodate the reference in Annales Cambriae to ‘Bishop Ebur’ in 501.  If the hypothetical Brittonic kingdom had developed out of a Late Roman Christian Brittonic aristocracy, it may also have retained the Late Roman Christian church hierarchy, at least for a while*. It can also accommodate the medieval tradition that the late sixth-century Brittonic king Peredur was associated with York (see post on Peredur). In this model, Peredur would have been the king of this hypothetical Brittonic kingdom that had developed in or around York. Peredur’s genealogy extends back to Coel Hen, the founder figure of most of the northern Brittonic dynasties.  Generation counting places Coel Hen somewhere in the early to mid fifth century, i.e. in the immediately post-Roman period, a plausible context for the emergence of a ruling dynasty (caveat, as always, that distant founder figures in genealogies may owe as much to imagination as to history).  Peredur’s father Eliffer had the epithet “of the Great Army”, which may imply that he had considerable military power in his time.  These scraps are consistent with a powerful Brittonic kingdom based in York from the immediate post-Roman period up to the later sixth century, though they do not prove it.

This staged transition model is not necessarily inconsistent with the presence of the pagan English cremation cemeteries at York; those only pose a problem if one assumes that populations must be ethnically, culturally and religiously homogeneous. The cremation cemeteries may represent one element of a mixed population living in post-Roman York, perhaps Germanic mercenaries hired by a Brittonic king or people descended from Germanic soldiers in the Late Roman Army, who happened to practice a particular funerary custom.

The transition from such a hypothetical post-Roman Brittonic kingdom to the English kingdom of Deira recorded by Bede could have occurred by conquest, alliance, intermarriage or inheritance, or any combination thereof. Peredur was killed in battle in 580 according to Annales Cambriae, and his son did not (re)claim his inheritance according to the Triads. Peredur’s death is a plausible context for a shift from Brittonic to Deiran control of York. Whether it represents direct military conquest by Deira, or an alliance hastily patched up after a military defeat by a common enemy, or an inheritance (disputed or otherwise) by the English-oriented children of a dynastic intermarriage in the previous generation, is open to interpretation.

A speculative model

I prefer the staged transition model, for several reasons. First, because it seems to me to fit with fragmentation of Roman Britain into a large number of small local polities, each ruled by whoever happened to be in the best position to grasp and maintain power in a given place at the time, which then changed and evolved over the following centuries.  Second, because a Brittonic kingdom based at York fits easily among the known post-Roman kingdoms ruled by kings with Brittonic names elsewhere in the region of what is now northern England/southern Scotland in the sixth century. Third, because it can accommodate the tradition associating Peredur with York (this is hardly strong evidence, since it comes from a medieval romance written centuries later, but it may reflect a genuine tradition).

So my speculative model for the development of post-Roman York sees a Brittonic tribal kingdom established in and around York, initially developing from the local Brittonic aristocracy and/or Roman officials based in York in the early to mid fifth century. By the mid to late sixth century it was a powerful kingdom capable of fighting a battle many miles away at Arthuret, whose rulers were important enough to be mentioned in the Annales Cambriae and were the subject of stories that survive in cryptic references in the Triads and later medieval romance. 

In this speculative model, the early English kingdom of Deira is postulated as separate from Brittonic York, based on the Yorkshire Wolds with an important centre near the extensive early cemetery at Sancton near Market Weighton and extending east to the coast and west towards the River Derwent.  If Soemil’s action that ‘first separated Deur from Berneich’ (see earlier post on Soemil) refers to gaining Deiran independence from a polity based in York, Deira may have initially been a sort of sub-region of Brittonic York, perhaps a land-grant to federate troops employed by the Late Roman Army based at York and/or the postulated Brittonic kingdom that succeeded it.  (More on the possible origins of Deira in a later post). I see Brittonic York and early English Deira as more or less independent neighbouring kingdoms for much of the fifth and sixth centuries, sometimes rivals and sometimes allies, depending on circumstances and the personalities of their respective leaders.  If Deira was initially founded by people who were formally granted land by Roman or post-Roman authorities at York, it is possible that both kingdoms may have shared a sense of Roman heritage (however hazy it may have become over time) and a tradition that they were supposed to co-operate militarily (whether they always did so in practice is a different matter).  Intermarriage could have reinforced such a (hypothetical) tradition, eventually leading to the effective merger of the two kingdoms under a Deiran king after the deaths of Peredur of York and his brother Gwrgi in 580 AD.  Whether this was voluntary, forcible or somewhere in between is open to interpretation. Since Peredur’s son Gwgaun is said in the Triads not to have (re)claimed his inheritance, implying that he was displaced, such a hypothetical merger may not have been entirely voluntary.  I lean to ‘somewhere in between’, with the aristocracy of Brittonic York accepting a Deiran king as the least-bad option available to them in a chaotic situation after their own kings had been killed in battle.  And thus this speculative model arrives at a situation in which York is a royal centre under the control of the early English kings of Deira in the late sixth century, ready to reappear in that guise in the documentary records in 627. 

I need hardly say that this is speculative.

 

*Even if the Annales Cambriae record means that there was a bishopric in York that came to an end in 501, it does not necessarily mean that Christianity disappeared along with the bishop. Monasticism was a powerful force in western Britain and Ireland in the early medieval period. If a monastery was established in Brittonic York during the fifth century, it may be possible that it had supplanted the local bishopric by 501.