The Date of the Mirabilia

“Certainly there is much about the early Arthurian tradition that might fit a god or supernatural hero. ……... This Arthur seems to belong in a world of ‘magical realism’, set apart from the abodes of man.”1 

It is generally accepted that the case for a “historical” Arthur begins with the battle list contained in the 9th century Historia Brittonum (History of the Britons, dated AD 829/30), the earliest surviving copy found in the Harleian manuscript 3859 dated AD c.1100. A single chapter provides the earliest narrative of Arthur in which he appears as ‘dux bellorum’, commander of the armies of the kings of Britain. The Harleian manuscript is the earliest and most complete text, possibly a direct copy of the original. The archetype may well have been the “certain very ancient book in the British language” presented to Geoffrey of Monmouth by Walter, Archdeacon of Oxford, which provided much of the source material for his 12th century 'The History of the Kings of Britain' (Latin: De gestis Britonum, or Historia Regum Britanniae).2

The Historia Brittonum was commonly referred to as “Nennius” following the preface in several manuscripts attributing the work to a cleric of this name who claimed to be “disciple of Saint Elved” (Elvodugus) commonly identified with Elfodd bishop of Bangor AD 755. But the assertion of Nennian authorship has largely been rejected ever since David Dumville claimed the preface to be a secondary addition, however, some scholars still argue for retention of Nennius’s authorship. 

Contained within the same manuscript and appended to the first sixty-six chapters of the Historia Brittonum is a collection of British genealogies, the Welsh Annals (Annales Cambriae), a list of the Cities of Britain (Civitates Brittaniae) and an untitled list of topographic marvels called the 'Wonders of Britain' (Mirabilia Britanniae). 

From the earliest references a figure called Arthur, we assume the same character as featured in the battle list, became attached to marvellous landscape features. Two such items are recorded in this list of ‘marvels’ or ‘wonders’ appended to the Historia Brittonum as chapters 67-76. The first marvel is the story of Carn Cabal (Cafall) a stone bearing the footprint of Arthur’s dog made when hunting the giant boar Twrch Trwyth:

1. “There is another wonderful thing in the region which is called Buelt. There is in that place a heap of stones, and one stone superposed on the pile with the footprint of a dog on it. When he hunted the boar Troynt, Cabal, who was the dog of Arthur the soldier, impressed his footprint on the stone and Arthur afterwards collected a pile of stones under the stone, whereon was the footprint of his dog, and it is called Carn Cabal. And men come and carry the stone in their hands for the space of a day and a night, and on the morrow it is found upon its pile”3 

Carn Cabal (illustration from Guest's Mabinogion)

The second records the ever changing size of the tomb of Arthur's son Amr (Anir): 

2. “There is another wonder in the country called Ergyng (Ercing). There is a tomb there by a spring, called Llygad Amr (Licat Amr); the name of the man who was buried in the tomb was Amr. He was the son of the warrior Arthur, and he killed him there and buried him. Men come to measure the tomb, and it is sometimes six feet long, sometimes nine, sometimes twelve, sometimes fifteen. At whatever measure you measure it on one occasion, you never find it again of the same measure, and I have tried it myself.”4

From this evidence as presented in the same manuscript it would appear that there were two Arthurs; a battle leader (the dux bellorum) of the post-Roman period, and a mythological figure associated with landscape features as seen in The Mirabilia. Clearly, the Arthur of The Mirabilia is at odds with the Arthur of the Nennian battle list; however, if the battle list in the Historia Brittonum was not a historical record but a legendary account then the two Arthurs become compatible.

Thomas Green has suggested that some of the entries in the Arthurian battle list have a mythological characteristic to them, arguing that a good case can be made for seeing Cat Coit Celidon (‘Battle of Coed Celyddon, the Caledonian Forest’) as the entirely mythical 'Battle of the Trees' recorded in the poem 'Kat Godeu' from the Book of Taliesin. Similarly, Green suggests the attack on the fort of Guinnon has all the appearance of an Otherworldly battle of the sort recorded of Arthur in the 8th-century (or earlier) Preiddeu Annwn. And the tenth, the ‘battle on the bank of a river which is called Tribruit’ is suspiciously similar to the Arthurian battle Traeth Tryfrwyd recounted in the poem Pa Gur? (What Man is the Gatekeeper?) in which Arthur is fighting werewolves (dogheads) and Garwlwyd who is also found in the Triads as Gwrgi Garwlwyd ('Man-dog Rough-grey') who killed one of the Cymry every day, and two on Saturday to avoid killing on Sunday. The fact that a battle from a mythical Welsh poem fought against supernatural creatures on the shore of the Tryfrwyd also appears in a so-called 'historical record' of the Historia Brittonum speaks volumes as to the author's sources and must cast serious doubt on the authenticity of the Arthurian battle list contained at Chapter 56.

There has been much debate as the dating of Pa Gur? Current opinion favours a 10th-century composition, although arguments have been forwarded for a 9th or even 8th-century origin. However, the poem clearly reflects an early mythological Arthur, fighting witches and supernatural monsters of a similar vein to the Arthur of The Mirabilia and Culhwch and Olwen. Patrick Sims-Williams suggests the poem may have been written near the south-east Welsh border, incidentally where the core of The Mirabilia are concentrated. If Sims-Williams is correct in locating the author of Pa Gur here it confirms a body of Arthurian lore was extant in this region from an early date, prior to the composition of the battle list in Historian Brittonum chapter 56.

Celtic scholar John Rhys,5 amongst others of the period, saw the primary figure as mythological from the pre-Christian Celtic world, and the leader of the sub-Roman period as a secondary development. Few Arthurian scholars would agree with Rhys today as the pendulum swings back favourably towards belief in a historical Arthur.

In a recent series of articles we have looked as evidence for a pre-Nennian Arthur, that an Arthur extant prior to the battle list of the Historia Brittonum which recorded Arthur the dux bellorum as victorious in twelve battles culminating in the battle of Badon. We find evidence for the existence of an Arthur by four individuals all so named ‘Arthur’ in the period AD 550-650.  All these individuals had links with Irish royal families in western areas of Britain but none could be considered 'THE ARTHUR' of Badon from the Historia. However, it must be noted that that the 9th century Historia Brittonum and 10th century Annales Cambriae are the first texts to associate the battle of Badon with Arthur; contemporary sources such as Gildas (later followed by Bede) failed to link the victory with Arthur.  

It is commonly believed that these four men of Gaelic descent must have been named after the memory of some military superhero from previous generations. The contemporary Welsh would not use the name ‘Arthur’ as if it carried some taboo, but this clearly did not apply to the Irish immigrants. Were these four men named after the battle leader at Badon or a surviving tradition of a mythological figure with origins in the pre-Christian world of the Celts, as envisaged by John Rhys?

To determine if Rhys was correct in arguing for a mythological primary figure we must consider which came first; the battle leader or the figure of the supernatural world? In doing so we are pitching Arthur the dux bellorum of the Historia against the Arthur of supernatural wonders of The Mirabilia.

Nicholas Higham poses the question, “If a ‘mythological’ Arthur was already the subject of storytelling in Wales when the Historia was written in 829–30, then might this be the ultimate source of the ‘historical’ figure who first appears therein? In that case Arthur’s emergence into history could be secondary to the figure of pagan mythology.”6

The case for a mythological Arthur rests on a body of Medieval Welsh literature where we find this same figure of Arthur leading a band of warriors raiding the Otherworld and stealing magic cauldrons, fighting giants, witches and hunting supernatural monsters. This Arthur was able to pass at will between the realms of myth and man; clearly this is clearly beyond the capabilities of a historical figure.

The essence of this mythological Arthur is found in the tale Culhwch and Olwen, poems such as Pa Gur? and Preiddeu Annwn, and of course The Mirabilia. Although these works are contained in later manuscripts their origins have been dated to the 9th and 10th centuries. It is significant that these early tales and poems include references to Arthur the ‘soldier’ (Latin = ‘miles’); yet this figure is a far cry from the King Arthur of Britain who becomes conqueror of Europe as portrayed by Geoffrey of Monmouth in the 12th century. In the vast majority of pre-Galfridian (before Geoffrey) texts, Arthur is involved in predominantly supernatural adventures. For this reason alone it is essential that Geoffrey’s impact on the Arthurian legend is fully appreciated and not underestimated.

The Wonders of Britain
The list of topographic marvels known as the Wonders of Britain (Mirabilia Britanniae) has been generally accepted as part of the original content of the Historia, but it is separated from the first sixty- six chapters in the earliest complete text (the Harleian MS 3859) by British genealogies, the Welsh Annals (Annales Cambriae), and the list of the Cities of Britain (Civitates Brittaniae).

The Historia has come down to us in several versions in over fifty manuscripts including, Harleian 3859, Chartres 98, Vatican Recension, Sawley Recension and even an Irish version, but unfortunately the archetype is no longer in existence. The Mirabilia does not appear in the Vatican recension but it is otherwise a near ubiquitous element of the Historia in its later recensions. Higham considers that The Mirabilia is unlikely to be of the same hand as the Historia with both the Civitates and The Mirabilia being very different in both style and purpose to the first sixty- six chapters of the Historia. Higham argues that The Mirabilia is therefore better treated as a separate text which was assimilated to the Historia.7

In most current versions of the Historia, The Mirabilia can be found as chapters 67-75, consisting of twenty marvels, immediately following chapter 66. The first mirabile listed is found at Loch Lomond in Scotland,8 however, without doubt the primary group are located in south- east Wales, the Severn valley, Ceredigion and the southern March (2–7, 9–14). Found in the midst of these is a miraculous ash tree found beside the river Wye said to bear apples (8). A further four have been appended (15–18) which are found in Anglesey with the final two relating to Ireland completing the list. 

Locations of The Mirabilia (after Higham, 2018)

In writing his chronicle of the The History of the Kings of Britain Geoffrey of Monmouth drew on many sources. It is evident from his use of certain mirabilia that he used a copy of the Historia Brittonum or the original thereof. Geoffrey has Arthur in Scotland following his victory over the Scots and Picts, in which he blockaded his enemies for a fortnight at Loch Lomond, where his companion Hoel is amazed by the sixty rivers, islands, crags, and eagles’ nests of the loch. This is a direct lift of features attributed to Loch Lomond in the Historia Brittonum. Hoel is then told about two other wonders by Arthur, which also derive directly from the Historia Brittonum. It is likely Geoffrey used these wonders simply to show some knowledge of Scotland.9 

The Historia goes South
The author of The Mirabilia seems to have drawn from existing groups of marvels in compiling his list, drawing the first four marvels, the Anglesey and the Irish items, probably from separate sources. Yet, the central group, marvels 5 to 14, are noticeably different in their presentation; each are are short individual narratives on their own merit, rather than a simple annotated list. This suggests that the author was personally acquainted with this group of marvels from south-Wales and the Marches. For example, in reference to measuring Amr ‘s tomb he claims “I have tried it myself”.10 

This bias suggests that the text originated from this area and the author likely from this area himself. Whereas it is clear that the first sixty-six chapters of the Historia Brittonum are a product of North Wales for Merfyn Frych, king of Gwynedd. Indeed, Higham sees The Mirabilia, as appended to the Harleian manuscript, as a work complete in itself and copied from a south Welsh archetype written no earlier than the mid- 950s - over a century later than the Historia.11 

Higham argues that although the genealogies attached to the Historia, as noted above, seem to derive from a north- Welsh genealogical collection, this was then “fleshed out and adapted to the needs of Dyfed’s court” in the 950s. Likewise, the Annales Cambriae were compiled in their extant form in the south, no earlier than 954. On that basis, he argues, the Harleian is best understood as a text copied from a manuscript written at St David’s in or after the mid-10th century. Higham suggests that both the Civitates and The Mirabilia are likely to have been attached to the Historia there and should be treated as independent texts written in southern Wales. He argues that it is very likely that the archetype of the Historia was limited to the first sixty- six chapters.

This is nothing new, it was suggested by Joseph Stevenson back in the early 19th century, who wrote in a the Preface to his translation of the Harleian Historia Brittonum, “The genealogies of the Saxon monarchs, and the Treatise de mirabilibus Britanniæ, formed, as we may believe, no part of the original work; but being of remote antiquity, and found in the greater number of the manuscripts, they have been inserted in the present edition.”12 

Higham concludes that, “It seems reasonable to conclude, therefore, that the Civitates and The Mirabilia are very unlikely to have been written by the Historia’s author. If it had not come south earlier, then the Historia may have reached Dyfed with Merfyn’s great- grandson, Hywel Dda, when he became king c.904”13 

There is no reason why we should interpret Higham’s argument as dating The Mirabilia to the mid-10th century, however, we can accept the case he presents for the date The Mirabilia became attached to the Historia. As Higham points out, it is very likely that The Mirabilia was in existence in south Wales as a separate, independent text long before being appended to to the Harleian manuscript. As we have seen above Stevenson regarded The Mirabilia as “being of remote antiquity.”

In truth we do not know when The Mirabilia was first written down, however as Rachel Bromwich states, the tradition of Carn Cafall and the hunting of the Twrch Twyrth was certainly already ancient by the 9th century.14

The two Arthurian marvels are both in origin folkloric explanations of topographical features. However, Brynley Roberts suggests the specific Arthurian elements appear to be later accretions, providing evidence that “by the 9th century Arthur had become a popular hero inasmuch that folklore motifs were being attached to his name and that he was a figure of sufficient fame to attract local legends into his orbit.” 15

Roberts adds that this was a feature of the later development of the Arthurian legend and examples of Arthurian topography are widespread, however, these two Arthurian marvels (Carn Cabal & Amr’s tomb) are the earliest examples which can be securely dated. Indeed, these two Arthurian marvels bear all the hallmarks of being a direct product of their author who had first hand knowledge of the topographic folklore of south-east Wales and along the English border long before their attachment to the Historia Brittonum.16

Individually these two marvels may not provide substantial evidence for Arthurian origins but when  combined together with other evidence, such as the hunt for the Twrch Twyrth, they achieve greater significance in the study of the provenance of the Arthurian legend.

Notes & References
1. Nicholas Higham, King Arthur: The Making of the Legend, Yale University Press, 2018, pp. 149-50.
2. Geoffrey of Monmouth claims to have translated the “certain ancient book in the British language” into Latin for his History of the Kings of Britain. No doubt much of his source material was taken from the Historia Brittonum, however, the extant versions of the HB are all in Latin, not the British tongue. Walter, Archdeacon of Oxford, had obtained this “ancient book” from Wales (Bern MS).
3. John Morris, ed & trans, Nennius: British History and the Welsh Annals, (Arthurian Period Sources, 8), Phillimore, 1980
4. Ibid.
5. John Rhys, Studies in the Arthurian Legend, Clarendon Press, Oxford, 1891.
6. Higham, 2018, p7.
7. Higham, 2018, pp. 179–80
8. Morris, 1980; in his translation of the Historia Brittonum, incorrectly translates stagnum Lumonoy as “Loch Leven” when it should read as “Loch Lomond”.
9. Ben Guy, Geoffrey of Monmouth’s Welsh Sources, in A Companion to Geoffrey of Monmouth, Brill, pp.31–66.
10. Brynley F. Roberts, Culhwch Ac Olwen, The Triads, Saints’ Lives, In Rachel Bromwich, A.O.H. Jarman, Brynley F. Roberts, Editors, The Arthur Of The Welsh: The Arthurian Legend In Medieval Welsh Literature, University Of Wales Press, 1991 (Second Edition, 2008), p.89-90.
11. Higham, 2018, p.226
12. Joseph Stevenson, Nennii: Historia Britonum, Preface, p.xviii, 1838.
13. Higham, 2018, p.180
14. Rachel Bromwich & D Simon Evans, Culhwch and Olwen: An edition and Study of the Oldest Arthurian Tale UWP 1992, Introduction, p.lxvi.
15. Roberts, 1991, p.92.
16. Ibid.

Edited 8/12/22

* * *