King Arthur of Somerset. King Arthur of Lyonesse. Was Somerset the land that King Arthur ruled? Was Somerset the true Lyonesse, the legendary lost kingdom beyond the western sunset? A striking, yet un-noticed, fact is that all the events of King Arthur’s life in sources whose roots are contemporary to his time, or nearly so, took place definitely, or at any rate perhaps, in or adjacent to Somerset.
Over time I will explore here the real King Arthur, Somerset, Lyonesse, and other aspects of fifth and sixth century Celtic Britain. I will write of the life of Taliesin, the great sixth-century bard, and decode his spiritual masterpiece Preiddeu Annwn. I will identify the antiquity of Glastonbury Abbey, and the rather lesser antiquity of the semi-mysterious Glastonbury Zodiac. I will feature the Somerset birth and death of St. Patrick, and the emergence of Saxon, down-to-earth, Somerset from the ancient Celtic sacred otherworldly summerland.
I begin though, almost as King Arthur did, with a little skirmish on a Somerset border; or, to be more accurate, on the border of Lindinis, the Roman civitas that included nearly all of what we now call Somerset. I say ‘almost’ as King Arthur did, because my focus is the “second, third, fourth and fifth battles” which Arthur fought according to the battle list in the Historia Brittonum – and not his first battle, which I will look at in due course.
King Arthur’s battles: in regione linnuis
The Historia Brittonum (HB) says Arthur fought his second, third, fourth, and fifth battles “in regione linnuis”. In language resembling that of arrogant climate-change believers and dogmatic environmentalist ‘we have to change our ways’ fundamentalists, the historian Keith Fitzpatrick-Matthews has recently [“The Arthurian Battle-List of the Historia Britonnum”: 2010] asserted that the often-suggested derivation of linnuis from Lindinis is “untenable”. He says that Lindinis “would have produced the form *lindinenses” – itself undisturbing, for it is pretty compatible with the genitive singular lendiniensis attested from inscriptions on Hadrian’s Wall – “which could not have developed into the attested linnuis”. Linnuis he says requires a prior *lindenses.
Perhaps this assertion would be sustainable if there had been a smooth transmission by precise scribes between the c600CE Welsh poem which Fitzpatrick-Matthews detects as the origin of the battle list and the ninth-century Latin of the HB. However, scholars regularly discover copying errors in Latin manuscripts. Indeed, Fitzpatrick-Matthews identifies some himself in a version of the HB, including a mistranscription of glein, the site of Arthur’s first battle, as glem. The obvious mistake was that the scribe mislinked the i and n into an m. Letters comprised only of loops and vertical lines are particularly vulnerable to copying errors of this kind, leading to confusions among the letters n, m, u, v, and i. There is therefore no solid assurance that linnuis is a true rendition of the word in the original composition.
Further, in the sixth and seventh centuries, Welsh poems were invariably transmitted orally. The culture rejoiced in bardic singing of poems from memory. It is not until the eighth century, at the earliest, that they come to be written down. It is easily possible that an actual fifth-century lindinenses became shortened to lindenses during the three centuries between Arthur fighting the battle and the record of it being written in the HB, especially as the HB has a north Welsh provenance and the city and civitas/kingdom of Lindinis were not only far away geographically but also had long since disappeared, metamorphosing into a Saxon-ruled Somerset.
Besides, even the putative modest elision from lindinenses to lindenses is unnecessary. There is a more direct route by which Lindinis may have elided by error into linnuis. Lindinenses means: the citizens of Lindinis. Grammatically, a battle site would more sensibly have been described as in regione lindinis than as in regione lindinenses. All that is required for a genitive lindinis to have become linnuis over centuries is its direct miscopying; and it is easy to see that i n i s in handwriting could have been misread by a copyist as n u i s. The evaporation of the d is less obvious, but as Fitzpatrick-Matthews himself indicates that philologically linnuis can or should be derived from a prior lindo– word, presumably the loss of the d is not in fact surprising.
Neither Fitzpatrick-Matthews nor any other scholar I am aware of has closely examined the immediate previous words in regione. They indicate, or implicitly assume, the meaning “in the region or district of”. Latin dictionaries, however, give another meaning for regio: boundary, or boundary-line. In regione can therefore equally validly mean “on the boundary of”. Culturally, this meaning makes excellent sense for the location of the early battles of a future warrior leader. An aspirant warrior king needed to prove his credentials by attacking a neighbouring kingdom and proving his war skills in a raid for cattle or other booty. A clear example of this practice is the raid by King Cynan of Powys on Gwent eulogised by the sixth-century bard Taliesin [Trawsganu Cynan Garwyn – see J E Caerwyn Williams: 1987. The Poems of Taliesin, p 1, and p xviii ff].
A linguistically and culturally coherent interpretation of in regione linnuis, therefore, is that Arthur earned his spurs as a warrior leader at the beginning of his period in power by fighting and winning four skirmishes on the boundary of the civitas of Lindinis.
The Boundaries of Lindinis
Where, then, were the fifth-century boundaries of Lindinis? These are not known exactly – Roman sources are not that good – but can be pretty closely surmised from knowledge of the Roman political structure of Britain at the start of the century and application of military geographical common sense. The civitates in the southwest peninsula were: Dumnonia, with its capital city at Exeter; Durotriges, capital Dorchester; Lindinis, capital Ilchester; and Corinium, capital Cirencester. Roger White [Britannia Prima: 2007] shows that after Emperor Diocletian’s reforms in 296CE, these civitates were all part of the Roman Province of Britannia Prima, as also were the special-status cities of Gloucester and Bath [map 38, p98]. To the east, in the Province of Maxima Caesariensis, lay the civitates whose capitals were Winchester and Silchester.
The eastern boundary of Lindinis must have been part of the eastern boundary of Britannia Prima. White’s map shows this running through Wiltshire in an arc which places Mere and its White Sheet Hill, and the modern towns Warminster, Westbury, Trowbridge and Melksham, on the Lindinis side of the line; but not Salisbury Plain or Market Lavington. The name Devizes is known to mean ‘at the division’. It is wholly sensible to assume Devizes was the triple border point between Lindinis, the Corinium civitas to the north, and the Silchester civitas to the east.
Military geographical common sense says that civitas boundaries will where possible have been along defensible frontiers. In the first instance, the dividing lines between tribal territories before the Roman conquest will have usually been decided by warfare, and inherited by the Romans. Secondly, it is likely that if there were any indefensible civitas boundaries when the Britons overthrew their Roman imperial governors in 409CE, such boundaries would have been shifted by warfare during the turbulent fifth century.
For Lindinis, therefore, the western boundary could have been substantially similar to the present Devon/Somerset border, along the Blackdown Hills and across Exmoor. However, the places named for Cador in north Devon, Cadbury Castle and Cadbury Barton, imply that the north-western boundary of King Arthur’s Lindinis was further west, along the River Taw. The boundary below CAdbury Castle could have run across the hills to the north of Crediton and Silverton and up the western escarpment of the Culm Valley to follow the present Somerset-Devon border south-estwards from Hockworthy. The southern border will have been further south than the arbitrary line across lowland that currently marks the austral limit of Somerset. The most defensible boundary runs along the North Dorset Downs from approximately Broadwindsor to Hod Hill and thence along the Wiltshire/Dorset border. Lindinis’s most likely northern boundary will have been along the River Avon from Avonmouth to Melksham, and from there across just six miles to the Britannia Prima border at Devizes – except for an excision to provide for the spa town of Bath. The size of similar excisions for Lincoln and Buxton, attested by Roman mileposts, suggests the Bath district boundary might have run south away from the Avon somewhere in the Keynsham area to the Cam Brook, perhaps near Dunkerton, and then along the Cam back to the Avon at Bradford-on-Avon.
The HB says Arthur’s four battles on the borders of Lindinis were “above the river which is called Dubglas.” Linguists inform us that the name Divelish can have derived from Dubglas. The River Divelish rises on the N Dorset Downs south of Ibberton and flows north until it joins the Stour close to the rich, and therefore well worth defending, Roman Villa at Hinton St Mary, the villa well-known today for its early evidence of British Christian practices.
The area above the Divelish is an utterly credible location for a young Arthur to have battled “on the boundary line of Lindinis” between his Lindinis warriors and the men of the Dorchester-based Durotriges civitas. The HB‘s super aliud flumen quod dicitur dubglas in regione linnuis is one piece of evidence that the warrior leader who came to be known as King Arthur was, at least as a youth, a man of Lindinis/Somerset.