Another approach is to look at verbal similarities between the entries. Most of the entries are very terse. The Arthur entries leap out because of their sentence structure and detail. If their particular linguistic features occur in other entries, this may give a pattern pointing to a common origin. The diagnostic features on which we can make the comparisons are: explanations of events as being ones in quo/in qua (in which) something happened; use of Gueith to mean battle; and the results of battle being corruit/corruerunt (he/they were slain) or victor fuit/victores fuerunt (he was the victor/they were the victors).
Those features occur, with their approximate dates, in these entries:
516
Bellum Badonis, in quo Arthur . . . et Britones victores fuerunt
Unlocated
537
Gueith Camlann, in qua Arthur et Medraut corruerunt
Unlocated
547
Mortalitas magna in qua pausat Mailcun
North Wales
613
Gueith Cair Legion . . .
North Wales
630
Gueith Meicen. . . . Catguollaun autem victor fuit
North Wales
631
Bellum Cantscaul in quo Catguollaun corruit
North Wales
644
Bellum Cocboy in quo Oswald . . . et Eoba . . . corruerunt
North Britain
682
Mortalitas magna . . . in qua Catgualart . . . obiit
North Wales
722
Bellum Hehil apud Cornuenses, Geuith Gartmailauc . . . et Brittones victores fuerunt
Cornish
750
Id est Gueith Mocetauc
North Britain
760
Id est Gueith Hirford
South Wales
813
Bellum . . . Higuel victor fuit
North Wales
844
Gueith Cetill
Unlocated?
873
Gueith Bannguolou
Unlocated?
The pattern is strikingly clear. It refutes the charge that the Arthur entries are late additions based on their complexity and structure. The majority of those of similar length and construction refer to the seventh/eighth centuries and the style is not continued beyond 873. Four of the last battle entries are similar only in their use of the word Gueith, twice as a gloss for entries which probably originally did not include it.
As is readily apparent they are features of a North Welsh phase of writing up to 813. All except one are about the North Welsh or (once) the Mercian and Northumbrian participants in the North Welsh wars. The only exception is the detailed description of the wars in Cornwall in 722.
Badon we have assumed to be a southern location, but for Camlann, with no other information to guide us, we should consider that it fits the rest of the pattern. Although a Cornish location is not out of the question, the balance of probability is that Camlann is a North Welsh battle. As there is a perfectly plausibly North Welsh Camlan, on the Dyfi, we should assume that is the battle-site the writer meant. Camlann continued to be a focus of Welsh tradition long after memories of the other battles, including Badon, faded, adding to the likelihood that it was a known location in Wales..
A later scribe did not expand a terse entry in the Annales reading ‘Bellum Badonis’ to incorporate the now famous figure of Arthur. This new analysis demonstrates the exact reverse. The description of Arthur’s victory is perfectly in accordance with the seventh- and eighth-century North Welsh entries. It is the location, Bellum Badonis, which is the anomaly. If any doctoring has taken place, it would be a replacement of an original obscure (North Welsh?) battle name with that of the more famous Badon, possibly derived from the Historia battle-list. If the battle at which Arthur carried the cross thirty years before the death of Mailcun was not originally equated with Badon, then a major discrepancy between the Annales and Gildas on their dating would be removed.
Bridging the Gap
We have focused on the Historia and Annales as the major battleground in the fight to prove the existence of a historical Arthur. If these accounts derive from the fifth and sixth centuries in any way, there is good reason for believing in him. If they are later fabrications, then the concept of Arthur as a real leader falls to pieces.
These sources impress because they are consistent, both with each other and well-established facts; they are plausible and they derive from sources which are independent and which existed before the works in which they appear. Information about the historical Arthur could only have been passed down between the sixth and the ninth centuries orally or in writing.
Oral Transmission?
There are two main categories of oral evidence. The first is local hearsay and legend, subject to the vagaries of memory and changing perception of the past. Only two pieces of the early Arthur material fall into this category: the wonders of Carn Cabal and Licat Anir. The author gives no indication that the wonders derive from anything but current observation and local tradition. The very fact of their existence is important for us. If Arthur was known to be a fighter only in the north-east and Kent, it would be odd for his name to be attached to wonders in South Wales. Only he and St Illtud have wonders attributed to them and no one would argue that the saint was just a generally famous cleric capriciously attached to a South Welsh site. The idea that Arthur was a soldier is another believable piece of information. Whether Arthur had actually killed his son Anir or hunted a boar called Troynt with his dog Cabal is another matter. It is unlikely that either of these pieces of information were supported by reliable sources.
The second is deliberate oral preservation. The British maintained an oral tradition, based on bardic poetry and genealogies, which was professional and systematic. This lasted deep into the Middle Ages and certainly was to be found between AD 500 and 800.
Genealogy is easy to dismiss. No genealogical links are provided for Arthur, nor is he placed in a sequential list of kings. There is no evidence for oral chronicle-style material in Wales. That leaves elegies and praise-poems as potential sources.
The Historia places the career of Arthur in the generation preceding the era of famous Welsh poets. It may be that Nennius implicitly claims these poets as his sources. Talhearn Tataguen stands out, with the others included more as an afterthought. Only Neirin and Taliessin have left their names attached to surviving works of poetry. In the ‘Book of Aneirin’ (the later Welsh version of his name) we find ‘Neirin’s’ poem Y Gododdin. We will look at some of the poetry attributed to Taliessin later.
Arthur’s mention in the Gododdin shows poetry is indeed a possible source. Arthurian poems would be composed in South Wales or the north-east, but are unlikely to survive in Kent. Arthur’s battle-list is no elegy for a fallen hero. There is no indication that he has died at Badon and the Annales flatly contradict this. We are then, perforce, thrown back on the concept of a praise-poem.
The arguments for a praise-poem are briefly summed up. The battle-list either comes from before Arthur’s death or derives from a poet writing in the persona of someone from before Arthur’s death. Welsh poets were quite capable of such imaginative writing. Two words, ostium and humeros, are said to point to Welsh originals meaning confluence or shield, respectively. Welsh poets were, however, equally able to refer to river mouths and men’s shoulders, the literal meanings of the Latin, and the list is not difficult to understand without altering these words. Finally, some of the battle names would rhyme in a putative Welsh source.
The Arthur material does not read like any surviving Welsh poetry. It is light on poetic imagery, has little repetition or other indication of verse structure. Most importantly, why are so few Welsh names preserved? Every battle in the source would have its Welsh name. Nennius translated Silva Celidonis for his British readers, but did not think to offer glosses for Castellum Guinnion, Urbs Legionis or Linnuis Regio, none of which is self-explanatory.
In truth, the reign of Arthur lies at the extreme limit of the period from which Welsh poetry could derive. In the fifth century, the British language had yet to become recognisably ‘Welsh’. Words still maintained their changeable endings, which altered depending on their role in the sentence. These forms still occur in Gildas’s time, but would shortly disappear in favour of words whose function, as in modern Welsh and English, was determined by their position in a sentence.
This change would have rendered early poetry unpoetic and even unintelligible. Rhymes dependent on case endings would disappear in updated versions. This means that the mid-sixth-century poetry of Neirin and Taliessin is probably the earliest that could have survived to be appreciated by medieval Welsh audiences.
None of the comparable Welsh poems provide the all-important element of context. We hear of the battles of Catraeth and Meigen, for instance, but not the era when they were fought. For this we have to turn to written sources, as surely the author of Historia Brittonum must have done.
Written Sources?
Though the Nennian prologue condemns the Britons as preliterate fools, the Historia proper contradicts this. For instance, the descent of Brutus from Noah is said to be preserved in ‘old books of our elders’ (HB 17). One written source, a Life of St Germanus, is specifically referred to in the text. This linked Germanus to the Powys dynasty, and was written before 820 when Powys was devastated by the English. A real fifth-century Life of St Germanus survives, and was used by Bede. It included victories over the Saxons with Germanus in the position of elected warleader (Dux Belli, as Bede phrases it). If the version of this used by Nennius included Arthurian material, it would have been as a continuation of the main story. Germanus had visited Britain in the 420s, and although both Bede and Nennius stretch the chronology of the story as far as possible, it still does not reach the next generation after Vortigern, where Arthur is placed.
Nennius had access to British material, dealing with the wars against the Northumbrians, down to Cadwallader in about 682. This is used as a source after the Arthur battle-list. It covers much the same ground as Bede’s Ecclesiastical History and may be a commentary on it. A more intriguing possibility is that it might pre-date Bede, since it does not continue into the early eighth century as Bede does. This source seems likely to be the principal ‘north-eastern’ source for the Arthurian material.
The northern material is interlinked with English genealogies, extending down to about 796 with Offa’s son in the genealogy of the Mercians. Most of the other genealogies end in the seventh century. Dumville rejects a more convoluted theory, that the northern British and English materials had already been merged by an early eighth-century historian, perhaps the ‘son of Urbagen’ given as the author in the Chartres Recension. It is safest to conclude that there is only one author responsible for linking together the sources of Historia Brittonum, the early ninth-century ‘Nennius’.
The last source, which seems to have been responsible for the framework in which the battle-list fits, is an English chronicle relating to the settlement of Kent. The material in this goes no further than the late sixth century, but it would be wrong to deduce it was written then. When the terms ‘late sixth century’ and ‘Kent’ are linked, the subject matter becomes obvious: the chronicle covered the origins of the Kingdom of Kent down to its conversion to Christianity in 597. If all the Saxon-fighters are linked within this frame, then Outigirn would be no later than this date, which accords with the other synchronisms. In its current form, the romanticised Kentish Chronicle seems to post-date Bede’s similar version of the story, and is therefore mid- to late eighth century, but its sources may be earlier.
A shared feature of the Arthur and Vortimer battle-lists is that they appear to have been composed in a language other than Welsh. Episford is glossed as ‘in nostra lingua Rithergabail’ (in our [Welsh] language Rithergabail) and ‘bellum in silva Celidonis’ as ‘id est Cat Coit Celidon’ (that is [in Welsh] battle of the wood of Celidon). If the lists were composed in Welsh, then translated into Latin, we would expect to find the opposite, e.g. ‘the battle of Cat Coit Celidon, that is [in Latin] the Celidonian Wood’. This points us towards a potential English source.
It is conceivable that the framing for the Arthurian battle-list and Outigirn entry, which are identical in style to the Vortimer/Hengist passages, might derive from the same ‘Kentish Chronicle’. Arthur’s battles are specifically said to be happening in the same area, against the same people. An English history would have its own limitations. The Anglo-Saxons had no written records of the fifth century, neither did they have a well-developed oral tradition.
The Anglo-Saxon Chronicle shows how Saxon historians worked. It is written to celebrate the West Saxon kings. The compilers have Bede, place-names, probably the kings’ lists and genealogies, and ‘traditions’. These are worked rather crudely into an analytic frame using Bede’s AD dating system, crudely, we can say, because some of the methods remain obvious. For instance, the chronological frame is almost certainly one in which leap years were marked out, probably for liturgical reasons. Many of the events of early Saxon history are set down as happening in leap years. The events of different dynasties are not interwoven. The entries of Kent are followed by those of the South Saxons, then the West Saxons. It is surely more likely that periods of Saxon expansion saw activities by all groups, rather than baton-passing between them. Bede’s story of Hengist and Horsa has resulted in the foundations of other kingdoms being presented as small groups, often under pairs of leaders, arriving on the south coast and then fighting their way to dominance over the local Britons. Finally, nothing is said of any peoples who were not ‘kingdoms’ in the ninth century.
In spite of these limitations, the Chronicle gives a useful illustration of how ninth-century writers imagined the sixth century. Of great importance are entries like those of Aelle which do not reinforce the West Saxon case. Saxon sources agree that this period saw reversals in their fortunes, though their adversaries’ names were rarely preserved.
We have an archaeological check on the Chronicle. Myres (1969) noted a break in the archaeological pottery sequence between the early and mid-sixth centuries in Kent, Essex, Hertfordshire, East Suffolk and Buckinghamshire. This would confirm the impression from the written sources that there was a real period of Saxon reversal around this time, and that the wars of Arthur had been synchronised to a very plausible period.
How does this compare with possible English sources for Historia Brittonum? One indication that their origins, at least, pre-date Bede, is that they lack Bede’s most important innovation, the AD dating system. Other accounts of the conquest, like the Anglo-Saxon Chronicle, eagerly latched on to this system to give legitimacy to the origin legends. The difference can be seen in, for example, the assigning of an AD date to the death of Horsa in the Chronicle, where none is even hinted at in the Historia. After Bede, AD dates became ubiquitous and this gives us the strongest hint that the Kentish Chronicle and the north-eastern kings’ lists use sources older than the early years of the eighth century.
The sources for the Arthurian section seem to be something like this: Nennius provides contemporary early ninth-century folk material on local wonders and activities of Arthur in regions adjoining South Wales. Some historical framing is given to the start of the period by the Book of the Blessed Germanus, a work from Powys before 820. Material relating to the north comes from British and English sources, combined by Nennius. Both these sources extend to the end of the seventh century, but have been revised, to bring some of the genealogical material to the late eighth century and to update British material. The sources are unknown, but include written English and British material, and possibly poetry from the named poets. Some of the Kentish material comes from an essentially pre-Bede (late seventh-century) source. Are these sources inadmissible?
Dark uses a generation count of thirty years to estimate the extent of time over which oral traditions might have been preserve
d (Dark 2000). Working on the supposition of a historian writing down the words of the oldest person available, recording what they had been told by, for example, a grandparent when they were a child, he deduces that 200 years is the maximum time one can reasonably expect oral tradition to survive without serious distortion. On this assumption, all we need is for the Arthurian material to be written down c. 740, less than 100 years before its incorporation into the Historia. As we can see, there is every reason to suppose that the author drew on both British and Saxon written sources of at least this age. This, combined with the Gododdin reference, suggests that, far from being inadmissible, the first historical sources to mention Arthur deserve serious consideration.
Taken together, the sources we have studied so far have produced a plausible and consistent picture of Arthur the Warleader. For those historians who affirm that this gives us no reason to accept it, we can reply that it gives us no reason to reject it either. To make that decision, we must look directly at the evidence from the turn of the sixth century. Only if we can find no trace of the reign of Arthur here can we be confident in dismissing it.
FOUR
Gildas was a man of God who believed he was watching the imminent destruction of Britain. Although he thought he could discern both the pattern leading to its destruction and the only way to avert it, he waited for ten years, uncertain that he was worthy to pronounce on the matter. Britain, after all, had leaders enough to deal with the situation.
At last, Gildas could wait no longer. He wrote the book which we call de Excidio Britanniae, ‘On the Destruction of Britain’. In it, he denounces his contemporaries, both religious and secular, named and unnamed, for not just being oblivious to the destruction of Britain, but actively bringing it about. Gildas’s view of history, based on his close study of the prophetic books of the Bible, was that there were obvious cycles. When patterns – either in the time of the Israelites or in the recent history of the British, God’s ‘latter-day Israel’ – could be seen repeating themselves, then it was only a small step to deduce what the future was likely to hold.
The Reign of Arthur Page 9
