New book claims that Offa's Dyke is Roman!
Right: Stretch of Offa's Dyke on Llanfair Hill, Shropshire.
The remarkable linear earthwork known as Offa's Dyke has recently been in the news thanks to a radical re-interpretation arguing the Dyke has nothing to do with the 8th century Anglo-Saxon King Offa whose name it carries, but was really built nearly 600 years earlier by the Roman Emperor Septimius Severus.
Researchers Steve Blake and Scott Lloyd make the claim in a new book 'The Keys to Avalon'. Quoting references to a wall of Severus in later classical and other texts, including the 4th century 'Scriptores Historia Augusta' which mentions the Severan wall in addition to the archaeologically known northern British frontiers of Hadrian and Antoninus Pius, the authors suggest that Offa's Dyke is this hitherto unrecognised structure. Noting accounts such as that of Eutropius, who says that the wall of Severus was built of turf and 132 miles long, Blake and Lloyd point to the similarity with the apparent length and construction of Offa's Dyke. They also draw attention to the absence of any concrete archaeological material to support the conventional Anglo-Saxon association for the Dyke, and indeed to excavation evidence from Ffrith where Roman pottery sealed beneath the Dyke bank dated to no later than the 2nd century AD.
'This is an interesting and, on the face of it, an attractive argument' says Ian Bapty, Offa's Dyke Archaeological Management Officer with CPAT 'and the press and media coverage it has attracted is certainly good news for raising public awareness of what is, in any terms, one of the most important and impressive archaeological monuments in western Europe. Moreover, Scott and Steve are quite right to indicate how little detailed understanding we have both of Offa's Dyke and the wider historical period to which it belongs. However, that should not cloud the fact that there are really very good archaeological reasons for thinking Offa's Dyke to be Anglo-Saxon rather than Roman. The ancient written sources are potentially insightful but also present a minefield of interpretative and palaeographic problems for modern historians, and it is crucial to look rigorously at the evidence on the ground as a balance in this process.'
Left: Milecastle on Hadrian's Wall, Northumberland
Roman building practice has been well studied by archaeologists, and the series of military installations in northern Britain which includes Hadrian's Wall and the Antonine Wall represents the most sophisticated and complete frontier complex in the Roman world. Detailed comparison with Offa's Dyke shows how different the latter monument is from these Roman works. There is no evidence of any palisade or wall associated with the Dyke; the style of the bank and ditch is not only very variable but also architecturally distinct from Roman turf wall construction; there are no forts, turrets, or any known additional infrastructure connected with the Dyke; there is no military road associated with the Dyke; there are no Roman (or other) building inscriptions from the Dyke; there are very few finds at all from the Dyke in marked contrast to the range of diagnostic artefacts typically recovered from Roman contexts.
As difficult to reconcile with plausible Roman procedure is the evident dislocation of Offa's Dyke from the currently recognised pattern of early 3rd century military sites in the Welsh borders. This includes the legionary fortresses at Chester in the north and Caerleon in the south, other forts such as those at Leintwardine, Caersws and Forden Gaer, and a road system of which some elements are still in use today as parts of the modern, A5, A39 and A41 routes. The alignment of Offa's Dyke shows no tangible geographical association or functional integration with this network. Indeed, it is in any case very hard to see what possible purpose such an undertaking could have served in Roman occupied western Britain, especially when the surviving Dyke is actually not a 130 mile complete frontier but is only spread discontinuously over that approximate length with extensive unexplained gaps (80 miles of earthwork are known). In short, the archaeological and historical anomalies become simply intractable in relation to any attempt to interpret the Dyke as a Roman structure. The stratigraphic sequence from Ffrith tells us no more than that the Dyke is later than the 2nd century in date; what the broader picture indicates is that it must be much later and almost certainly post-Roman.
This conclusion is reinforced by the direct evidence for connecting the Dyke with Offa and the ancient Anglo-Saxon kingdom of Mercia. At a general level, the form and nature of Offa's Dyke can be compared most closely with other linear earthworks of probable Anglo-Saxon date. These include not only the nearby Wat's Dyke (possibly later 5th or early 6th century AD on recent radiocarbon dating), but other monuments such as the Devil's Dyke in Cambridgeshire and the Wansdyke in Wiltshire. More specific corroboration includes Frank Noble's work showing how the line of Offa's Dyke runs across Anglo-Saxon townships. This strongly suggests that the Dyke was built after the Anglo-Saxon land boundaries first come into being - or else those boundaries somehow ignored a massive and ready made existing land division, which would seem very unlikely. Perhaps most telling is that it is really only in the late 8th century context of Offa's Mercia, and the fractious process of English expansion on the fringes of the independent British kingdoms then occupying present day Wales, that we have some kind of plausible historical situation in which to develop an understanding of why the Dyke was built - even if that understanding is still difficult to elaborate in detail.
Right: Section of Offa's Dyke in Radnorshire, Powys.
It is for all these reasons that - despite the lack of absolute archaeological dating evidence for the monument - the attribution of the Dyke to Offa by Asser in his late 9th century 'Life of Alfred', echoed by the tradition of the 'Offa's Dyke' name itself which can be documented back as far as the 13th century, has been accepted as correct by Anglo-Saxon scholars. 'Offa's Dyke is an extraordinary survival from our Anglo-Saxon past' says Ian Bapty 'and extraordinary exactly because it is Anglo-Saxon and as such sheds crucial light on a key period of our history when the modern political geography of Britain was beginning to appear. While we can perhaps associate descriptions of the 'missing' wall of Severus with somewhat confused and secondarily derived later accounts of Hadrian's Wall - which was much rebuilt in the time of Severus - we surely cannot backdate Offa's Dyke to Roman times, and to do so would be to miss the real significance and historical impact of this amazing earthwork'.
'Ultimately I'd be ready to wager my granny on the fact that Offa's Dyke is Anglo-Saxon and not Roman!' says Ian 'although I'd also have to be say that I'd be keeping granny firmly out of the stakes when it comes to betting on most other aspects of our understanding of the Dyke, including key issues such as exactly why it was built, how it was built, and what it's original appearance and total extent was. I think it is the process of trying to answer these questions which may throw up some real and lasting revelations concerning not just Offa's Dyke itself, but the very origins of Welsh and English culture and society'.
For more information about Offa's Dyke and the English Heritage/Cadw supported Offa's Dyke Initiative aiming to improve the conservation management of the earthwork, check out our Offa's Dyke web pages.
Ian Bapty, April 2000