The Twelfth Battle: Mount Badon (from my book, THE ARTHUR OF HISTORY)

St. Ann's Well, Buxton, Derbyshire

Badon is a difficult place-name for an unexpected reason. As Kenneth Jackson proclaimed:

"No such British name is known, nor any such stem." [To be briefly mentioned in the context of Badon is the Middle Welsh word bad, 'plague, pestilence, death' (GPC; first attested in the 14th century), from Proto-Celtic *bato-, cf. Old Irish bath. Some have asked me whether this word could be the root of Badon - to which Dr. Graham I. Isaac, of the National University of Ire-land, Galway, responds emphatically, "No, absolutely no. A (modern) W form _bad_ etc. would have been spelt in the W of the ancient period as _bat_ and there can be no connection since _Bad(on)_ is what we find." Other noteworthy Celtic linguists, such as Dr. Simon Rodway of Aberystwyth University, Dr. Richard Coates of the University of the West of England and Professor Ranko Matasovic of the University of Zagreb, agree with Isaac on this point. Matasovic adds: “Professor Isaac is right; since we have references to Badon in Early Welsh sources, the name would have been spelled with –t- (for voiced /d/). The spelling where the letter <d> stands for /d/ and <dd> for the voiced dental fricative was introduced in the late Middle Ages.”]

Graham Isaac has the following to say on the nature of the word Badon, which I take to be au-thoritative.

His explanation of why Gildas's Badon cannot be derived from one of the Badburys (like Liddington Castle, often cited as a prime candidates for Badon) is critical in an eventual identification of this battle site. Although long and rather complicated, his argument is convincing and I have, therefore, opted to present it unedited:

"Remember in all that follows that both the -d - in Badon and the -th- in OE Bathum are pronounced like th in 'bathe' and Modern Welsh - dd-. Remember also that in Old English spelling, the letters thorn and the crossed d are interchangeable in many positions: that is variation in spelling, not in sound, and has no significance for linguistic arguments.

It is curious that a number of commentators have been happy to posit a 'British' or 'Celtic' form Badon. The reason seems to be summed up succinctly by Tolstoy in the 1961 article (p. 145):

'It is obviously impossible that Gildas should have given a Saxon name for a British locality'.

Why? I see no reason at all in the world why he should not do so (begging the question as to what, exactly, is the meaning of 'British locality' here; Gildas is just talking about a hill). This then becomes the chief crutch of the argument, as shown on p. 147 of Tolstoy's article: 'But that there was a Celtic name ‘Badon’ we know from the very passage in Gildas under discussion'.

But that is just circular: ' "Badon" must be "Celtic" because Gildas only uses "Celtic" names'. This is no argument. What would have to be shown is that 'Badon' is a regular reflex of a securely attested 'Celtic' word. This is a matter of empirical detail and is easily tested; we have vast resources to tell us what was and was not a 'Celtic' word. And there is nothing like 'Badon'.

So what do we do? Do we just say that 'Badon' must be Celtic because Gildas uses it? That gets us nowhere.

So what of the relationships between aet Bathum - Badon - Baddanbyrig? The crucial point is just that OE Bathum and the Late British / very early Welsh Badon we are talking about both have the soft -th- sound of 'bathe' and Mod.Welsh 'Baddon'. Baddanbyrig, however, has a long d-sound like -d d- in 'bad day'. Both languages, early OE and Late British, had both the d-sound and the soft th-sound. So:

1) If the English had taken over British (hy-pothetical and actually non-existent) *Badon (*Din Badon or something), they would have made it *Bathanbyrig or the like, and the modern names of these places would be something like *Bathbury.

2) If the British had taken over OE Baddanbyrig, they would have kept the d-sound, and Gildas would have written 'Batonicus mons', and Annales Cambriae would have 'bellum Batonis', etc. (where the -t- is the regular early SPELLING of the sound -d-; always keep your conceptions of spellings and your conceptions of sounds separate; one of the classic errors of the untrained is to fail to distinguish these). 

I imagine if that were the case we would have no hesitation is identifying 'Baton' with a Badbury place. But the d-sound and the soft th -sound are not interchangeable. It is either the one or the other, and in fact it is the soft th -sound that is in 'Badon', and that makes it equivalent to Bathum, not Baddanbyrig. 

(That applies to the sounds. On the other hand there is nothing strange about the British making Bad-ON out of OE Bath -UM. There was nothing in the Late British/early Welsh language which corresponded to the dative plural ending - UM of OE, so it was natural for the Britons to substitute the common British suffix - ON for the very un-British OE suffix -UM: this is not a substitution of SOUNDS, but of ENDINGS, which is quite a different matter. That Gildas then makes an unproblematic Latin adjective with -icus out of this does not require comment.)

To conclude:

1) There is no reason in the world why a 6thcentury British author should not refer to a place in Britain by its OE name.

2) There was no 'British' or 'Celtic' *Badon.

3) 'Badon' does not correspond linguistically with OE Baddanbyrig.

4) 'Badon' is the predictably regular Late British / early Welsh borrowing of OE Bathum.

Final note: the fact that later OE sources occasionally call Bath 'Badon' is just a symptom of the book-learning of the authors using the form. Gildas was a widely read and highly respected author, and Badon(-is) (from Gildas's adjective Badon -icus) will quickly and unproblematically have become the standard book-form (i.e. pri-marily Latin form) for the name of Bath. Again, all attempts to gain some sort of linguistic mile-age from the apparent, but illusory, OE variation between Bathum and Badon are vacuous."

It is thus safe to say that 'Badon' must derive from a Bath name. However, we must not restrict ourselves to the Southern Bath, which makes no sense in the context of a Northern Arthur.

For as it happens, there is a major Northern ‘Bath’ site that has gone completely unnoticed!

In the the High Peak District of Derbyshire we find Buxton. This town had once been roughly on the southernmost boundary of Brigantian tribal territory (thought to lie along a line roughly from the Mersey in the west to the Humber in the east). It was also just within Britannia Inferior (that part of northern Britain ruled from York), whose boundary was again from the Mersey, but probably more towards The Wash. 

In the Roman period, Buxton was the site of Aquae Arnemetiae, ‘the waters in front of (the goddess) Nemetia’. To the best of our knowledge, Bath in Somerset and Buxton in Derbyshire were the only two ‘Aquae’ towns in Britain.

But even better, there is a Bathum name extant at Buxton. The Roman road which leads to Buxton from the northeast, through the Peak hills, is called Bathamgate. Batham is ‘baths’, the ex-act dative plural we need to match the name Bathum/Badon. -gate is ‘road, street’, which comes from ME gate, itself a derivative of OScand gata. Bathamgate is thus ‘Baths Road’.

The recorded forms for Bathamgate are as fol-lows:

Bathinegate (for Bathmegate), 1400, from W.

Dugdale's Monasticon Anghcanum, 6 vols, London

1817-1830

Bathom gate, 1538, from Ancient Deeds in the

Public Record Office

Batham Gate, 1599, from records of the Duchy of Lancaster Special Commissions in the Public

Record Office.

Buxton sits in a bowl about one thousand feet above sea level surrounded by mountains and is itself a mountain spa. The natural mineral water of Buxton emerges from a group of springs at a constant temperature of 82 degrees Fahrenheit and is, thus, a thermal water. There are also cold springs and a supply of chalybeate (iron bearing) water. The evidence of Mesolithic man suggests a settlement dating to about 5000 BCE and archaeological finds in the Peak District around the settlement show habitation through the Neolithic, Bronze and Iron Ages to the time of the Romans. 

From the historical evidence we can say that Buxton was a civilian settlement of some importance, situated on the intersection of several roads, and providing bathing facilities in warm mineral waters. In short, it was a Roman spa. Place-names in and around Buxton, and Anglo-Saxon finds in burial mound excavations, suggest a continuing inhabitation of the area and probable use of the mineral waters.

It has long been speculated that we should expect to find a military installation at Buxton. However, subsequent archaeological fieldwork, including excavations, in and around suggested locations at the spa town have singularly failed to establish a military presence. A 'ditch feature' identified initially through resistivity survey and then from aerial photography above Mill Cliff, Buxton, gave rise to the almost confident interpretation of this site as being that of the fort: subsequent evaluation in advance of development, however, has shown that these features were geological rather than man-made, and the absence of Roman finds of any description from a series of evaluation trenches suggests that if

Buxton had a fort it was located elsewhere.

Today, the site of the probable Roman baths is covered by the Georgian Crescent building. In this area during the seventeenth and eighteenth century discoveries of lead lined baths, red plaster and building remains were made at some considerable depth in the sediments which surround the area of St Anne's well. In the eighteenth century, Pilkington investigated a mound overlooking the site of the previous discoveries. Here he found a structure which has been interpreted as a probable classical temple - one of only three known from Britain. In the mid-seventies, following the removal of a 20th century swimming pool, a brick structure was exposed and a deposit containing 232 Roman coins, 3 bronze bracelets and a wire clasp ranging in date from the 1st to the end of the 4th century CE was excavated.

This intriguing series of early discoveries lends tangible support to the interpretation of Buxton as the 'Bath of the North', but the character and extent of civilian settlement - and whether this was in association with a military installation or not, remains obscure. A considerable range of small finds, together with occasional glimpses of apparently Roman contexts, from the backgardens of houses has failed to provide a clear sense of the extent of Roman Buxton, let alone a soundly based understanding of its chronology and development. The dating of coinage in the 'votive' deposit from near the Crescent might be seen to indicate heightened frequencies of offer-ings during the third and fourth centuries. To what extent this might correlate with the development of settlement at Buxton is a matter of some conjecture.

At Poole's cavern, Buxton, excavations between 1981 and 1983 by Peakland Archaeological Society and Buxton Archaeological Society produced a large Romano-British assemblage containing a considerable body of metalwork including coins and brooches, rolls of thin sheet bronze, along with ceramics, a faunal assemblage and burials. The dating of the coins and fibulae point to use between the late 1st and 3rd centuries, with the majority being of 2nd century date. Indeed, reanalysis of the material has suggested that the cave saw its principal period of use between 120 and 220 CE. The excavators appeared to reveal some spatial separation of the coin and fibulae finds from the pottery and faunal remains, although this has been questioned.

Discussing the possible character of the use of the site Bramwell and Dalton draw attention to the comparative absence of spindle whorls, loom weights and bone hairpins which might be expected from a domestic site. Instead, they see the evidence as supporting the interpretation of the site as that of a rural shrine or sanctuary.

This too has subsequently been questioned and rejected. Instead, Branigan and Dawley interpret the site as essentially domestic, but with the additional refuse from a metalworker’s activities. They see a link between Poole's Cavern and the growth of Buxton as a spa centre providing a ready local market for small decorative trinkets.

The general trend of the evidence suggests that the Roman site may have consisted of a temple overlooking a set of Roman baths. At Bath we have a clear idea of the layout of a significant bath/water shrine complex which consisted of two major ranges: a temple and a religious precinct, within which lay the sacred spring; along-side this range were a line of three baths within a major building, at one end of which lay a typical Roman bathhouse or sauna. The Bath buildings were lavishly built in a classical style and the whole complex attracted visitors from outside the province.

In essence the Buxton layout mirrors that a Bath: parallel to the spring line is a temple and alongside the springs is a range of possibly Roman baths. As the Buxton temple is two-thirds the size of that at Bath we could assume the Buxton complex was somewhat smaller.

If the grove of the goddess Nemetia continued as an important shrine well into Arthur’s time (and the presence of St. Anne’s Well at the site of the town’s ancient baths shows that the efficacy of the sacred waters was appropriated by Christians), there is the possibility the Saxons targeted Buxton for exactly this reason. Taking the Britons’ shrine would have struck them a demoralizing blow. If the goddess or saint or goddess-become-saint is herself not safe from the depredations of the barbarians, who is?

A threat to such a shrine may well have galvanized British resistence. Arthur himself may have been called upon to lead the British in the defense of Nemetia's waters and her templegrove.

There may be a very good reason why Gildas (or his source, or a later interpolator) may have opted for English Bathum (rendered Badon in the British language of the day). The two famous 'baths' towns were anciently known as Aquae Sulis and Aquae Arnemetiae for the two goddesses presiding over the hot springs. As Arthur is made out to be the preeminent Christian hero, who in the Welsh Annals has a shield bearing the Cross of Christ that he carries during the Battle of Badon, it would not do for the ancient Romano-British name to be used in this context. To have done so would inevitably have referred directly to a pagan deity. Hence the generic and less “connotation-loaded” Germanic name for the place was substituted. This explanation might do much to placate those who insist on seeing Badon as a Celtic name.

And where is the most likely location for the monte/montis of the Baths/Batham/Badon, where the actual battle was fought?

I make this out to be what is now referred to as The Slopes, at the foot of which is the modern St. Ann’s Well, and the Crescent, under which the original Roman bath was built. The Slopes were once called St. Ann’s Cliff because it was a prominent limestone outcrop. The Tithe map of 1848 shows that the upper half of the Cliff was still largely covered in trees. I suspect the spring was anciently thought to arise from inside the Cliff, and that the trees covering it marked the precincts of the nemeton or sacred grove of Arnemetia.

The three days and three nights Arthur bore the cross (or, rather, a shield bearing an image of a cross) at Badon in the Welsh Annals are markedly similar to the three days and three nights Urien is said to have blockaded the Saxons in the island of Lindsfarne (British Metcaud) in Chapter 63 of the HB. In Gildas, immediately before mention of Badon, we have the following phrase: "From then on victory went now to our countrymen, now to their enemies…" Similarly, just prior to mention of Urien at Lindisfarne, we have this: "During that time, sometimes the enemy, sometimes the Cymry were victorious…" It would seem, therefore, that either the motif of the three days and three nights was taken from the Urien story and inserted into that of Arthur or vice-versa.

What is fascinating about this parallel is that Lindisfarne or ‘Holy Island’, as it came to be known, was an important spiritual centre of Northern Britain. The inclusion of the three days and three nights (an echo of the period Christ spent in the tomb) in the Badon story suggests that we can no longer accept the view that Arthur's portage of Christian symbols at Badon was borrowed solely from the Castle Guinnion battle account in the HB. Aquae Arnemetiae, like Lindisfarne, was a holy place. Arthur's fighting there may have been construed as a holy act.

Supposedly, 960 Saxons were slain by Arthur at Badon. In the past, most authorities have seen in the number 960 no more than a fanciful embellishment on the Annals' entry, i.e. more evidence of Arthur as a ‘legend in the making’. But 960 could be a very significant number, militarily speaking. The first cohort of a Roman legion was composed of six doubled centuries or 960 men. As the most important unit, the first cohort guarded the Roman Imperial eagle standard.

Now, while the Roman army in the late period no longer possessed a first cohort composed of this number of soldiers, it is possible Nennius's 960 betrays an antiquarian knowledge of earlier Roman military structure. However, why the Saxons are said to have lost such a number cannot be explained in terms of such an anachronistic description of a Roman unit.

The simplest explanation for Nennius's 960 is that it represents 8 Saxon long hundreds, each long hundred being composed of 120 warriors. To quote from Tacitus on the Germanic long hundred:

"On general survey, their [the German's] strength is seen to lie rather in their infantry, and that is why they combine the two arms in battle. The men who they select from the whole force and station in the van are fleet of foot and fit admirably into cavalry action. The number of these chosen men is exactly fixed. A hundred are drawn from each district, and 'the hundred' is the name they bear at home. What began as a mere number ends as a title of distinction." [Germania 6]

Curiously, in the Norse poem Grimnismal, 8 hundreds of warriors (probably 960) pass through each of the doors of Valhall, the Hall of the Slain, at the time of Ragnarok or the Doom of the Powers.

Osla or Ossa Big-Knife and Caer Faddon

It has often been said that the Welsh Caer Faddon is always a designation for Bath in Avon.

However, at least one medieval Welsh tale points strongly towards the ‘Baths’ at Buxton as the proper site.

I am speaking, of course, of the early Arthurian romance ‘The Dream of Rhonabwy’, sometimes considered to be a part of the Mabinogion collection of tales. Rhonabwy is transported back in time via the vehicle of a dream to the eve of the battle of Caer Faddon. Arthur has apparently come from Cornwall (as he is said to return thither after a truce is made) to mid-Wales and thence to Caer Faddon to meet with Osla or Ossa, a true historical contemporary of Arthur who lies at the head of the royal Bernician pedigree. 

As Arthur is said to progress from Rhyd-y-Groes to Long Mountain, he is traveling to the north-east via the Roman road. In other words, he is headed in the direction of Buxton in the High Peak.

While the romance is entirely fanciful, the chronological accuracy in the context of choosing Osla/Ossa is rather uncanny. Furthermore, it is quite clear that in the tradition the author of the romance was drawing from, Caer Faddon is most certainly not Bath. Ossa is known in English sources for being the first of the Bernicians to come to England from the Continent. Under his descendants, Bernicia became a great kingdom, stretching eventually from the Forth to the Tees. In the 7th century, Deira – which controlled roughly the area between the Tees and the Humber - was joined with Bernicia to form the Kingdom of Northumbria.

In its heyday, Northumbria shared a border with its neighbor to the south – Mercia – at the River Mersey or ‘Boundary River’. The Mersey flows east to Stockport, where it essentially starts at the confluence of the River Tame and Goyt. The Goyt has its headwaters on Axe Edge, only a half a dozen kilometers from Buxton in the High Peak.

If we allow for the story’s author to have properly chosen Ossa as Arthur’s true contemporary, but to have viewed Northumbria in an anachronistic fashion – i.e. as extending to the River Mersey – then Ossa coming from Bernicia in the extreme north of England, and Arthur coming from Cornwall in the extreme southwest, meeting for a battle at Buxton makes a great deal of sense. In fact, Buxton is pretty much exactly equidistant between the two locations.

Ossa would have been viewed as engaging in a battle just across the established boundary.

If I am right about this, the Welsh knew of the ‘Bathum’ or Badon that was Buxton.