Enter your email address to subscribe to this site and receive notifications of new posts by email.

Join 261 other subscribers


What are the historical sources for Arthur?

As Arthur is potentially an historical character, the documents naming him must form the starting point for any investigation. This is where the difficulties begin, as there are no contemporary documents naming him. Nor are there coins or inscriptions (we will see how several supposed inscriptions are irrelevant).

The earliest evidence

A page from the manuscript Book of Aneirin, containing the only text of Y Gododdin

A page from the manuscript Book of Aneirin, containing the only text of Y Gododdin

Y Gododdin

The oldest possible reference comes from a cycle of poems known as Y Gododdin, ‘The Votadini’, ascribed to a poet named Aneirin, who is supposed to have lived in the later sixth century. The verse in question reads:

gochore brein do aruur
caer – ceni bei ef arthur –
rug ciuin uerthi ig disur
ig kynnor guernor guaur

“He sent down black crows from the wall
of the fort – although he was not Arthur –
among men mighty in feats
before the alder grove: Gorddur”

Like so much early Welsh verse, the language is terse and full of allusions to stories that the listener would have known in much greater detail than we can. All we can reasonably infer from this verse is that Arthur was a famous warrior. And the date at which it was written is another problem: the only surviving manuscript (MS 2.81 of Cardiff Library, better known as the Book of Aneirin) dates from the late thirteenth century. To make matters worse, it is a composite of three variant texts belonging to two separate versions and the verse naming Arthur occurs in only one of them (the B2 text, for the technically minded), which is known to contain additions later than the supposed date of composition. This has led some commentators to believe that the verse was composed in the ninth or tenth century; on the other hand, the linguist John Koch sees traces of Primitive Welsh in the text, especially in its rhyming scheme, which would argue for an early seventh-century date for composition, at the latest. Whilst this remains a tantalising possibility, it is nothing more.

A page from a manuscript of the Historia Brittonum, listing the battles of Arthur

A page from a manuscript of the Historia Brittonum, listing the battles of Arthur

The Historia Brittonum

The earliest definite references to Arthur come from a complex and poorly studied text, the so-called Historia Brittonum (‘History of the Britons’), often ascribed to a Welsh scholar known as Nennius (or Nemniuus), although the preface naming him as author is only found in two late versions. The text tells us that it was composed in the fourth year of Merfyn Frych, king of Gwynned 825-844, placing its composition in 828/9; there are possible hints of a late eighth-century version. Famously, there is a section devoted to the battles of Arthur, who pugnabat contra illos in illis diebus cum regibus brittonum, sed ipse dux erat bellorum (‘fought against them (i.e. the Saxons) in those days, with the kings of the Britons, but he was leader of battles’). Twelve battles are named (one of them four times, so only nine separate locations are mentioned), including duodecimum… bellum in monte badonis, in quo corruerunt in uno die nongenti sexaginta uiri de uno impetu arthur; et nemo prostrauit eos nisi ipse solus (‘the twelfth battle… on the Mount of Badon, in which nine hundred and sixty men fell together on one day from one charge of Arthur; and no-one laid them low except he alone’). We are fortune that the Battle of the Badonic Mount is mentioned as a victory for the British against the Saxons by a near contemporary text, the so-called de Excidio et Conquestu Britanniae (‘About the Destruction and Conquest of Britain’) by a British writer known as Gildas, who is generally believed to have written in the early sixth century. Unfortunately, Gildas does not name the victorious commander (there was no reason for him to), but we can at least be confident that this was a real victory for the Britons. On the other hand, the tale that no-one other than Arthur was responsible for enemy deaths looks more like legend than history. Still, this is not any reason to discount the text.

There are other references to Arthur in the Historia Brittonum. The first reads est aliud mirabile in regione quae dicitur buelt. est ibi cumulus lapidum et unus lapis superpositus super congestum cum uestigio canis in eo. quando uenatus est porcum troit, impressit cabal, qui erat canis arthur militis, uestigium in lapide et arthur postea congragauit congestum lapidum sub lapide in quo erat uestigium canis sui, et uocatur carn cabal. et ueniunt homines et tollunt lapidem in manibus suis per spatium diei et noctis et in crastino die inuenitum super congestum suum (‘There is another wonder in the region that is called Builth. There, there is a pile of stones and one stone placed on the pile with the paw-print of a dog on it. When the boar Trwyth was being hunted, the dog Cafall (which was the dog of Arthur the soldier) pressed his print into the stone and afterwards Arthur brought together a pile of stones beneath the stone in which there was the print of his dog, and it is known as Corn Gafallt. And men come and take away the stone in their own hands for the length of a day and a night, and on the next day, the stone is found back on its pile’). The second reads est aliud miraculum in regione quae uocatur ercing. habetur ibi sepulcrum iuxta fontem, qui cognominatur licat amr, et uiri nomen, qui sepultus est in tumulo, sic uocabatur amr: filius arthuri militis erat et ipse occidit eum ibidem et sepeliuit. et ueniunt homines ad mensurandum tumulum in longitudine aliquando sex pedes, aliquando nouem, aliquando duodecim, aliquando quidecim. in qua mensura metieris eum in ista uice, iterum non inuenies eum in una mensura, et ego solus probaui (‘There is another wonder in the region that is called Ergyng. There, there is a tomb next to a spring, which is known as Gamber Head, and the name of the man who was buried in the mound was thus called Amr: he was the son of the soldier Arthur, and he himself killed him and buried him there. And men come to measure the mound and sometimes it is six feet in length, sometimes nine, sometimes twelve, sometimes fifteen. Whatever measurement you will have measured it on one occasion, you will never find it the same measurment again, and I have tested it myself’).

These are at the same time impressive and problematic pieces of text. Arthur is clearly a well known character and he is a soldier. This confirms what we have already discovered. But he is associated in the first with the hunting of the boar Trwyth, an episode found in a much later story, Kulhwch ac Olwen (‘Culhwch and Olwen’), where the boar is in Ireland and is said to have been a king transformed into the beast for his sins. The magical elements of the Kulhwch ac Olwen story already appear to be present in the apparently folkloric origin story for the dog’s footprint. It should also be noted that cafal is Old Welsh for ‘horse’, which would be an odd name for a dog, so it looks as if there is alreay a degree of confusion in the tale. The site is today Corn Gafallt, near Rhaeadr, where there are indeed some burial mounds south of the summit. With the story of Llygad Amr, we have similarly magical details. The tomb ought to be beside Gamber Head, a spring whose name derives from the Old Welsh Amr. A tomb has been found here, which is an element in a spectacular collection of Bad Archaeology investigated by Time Team in 2001. The location of the original tomb is unknown, although it has been suggested that the nearby but destroyed Wormelow Tump may have been the site of the ‘wonder’.

The Historia Brittonum is the mainstay of those who wish to present Arthur as an historical figure, but as we can see, there are enormous difficulties with those parts of the text naming him that make it impossible to read as a straightforward historical narrative. Instead, when we look at the document as a whole, we can see that it was compiled from largely legendary material. Even so, it has to be said that the other main figures it names from the Roman and early medieval periods were genuinely historical rather than mythological, even if clearly legendary stories have been woven around them.

Manuscript A of the Annales Cambrię; the entry dealing with the Strife of Camlann is at the bottom right; Badon is just above middle right

Manuscript A of the Annales Cambrię; the entry dealing with the Strife of Camlann is at the bottom right; Badon is just above middle right

Annales Cambrię

The next historical source to mention Arthur is known as Annales Cambrię (‘Annals of Wales’), a chronicle of Welsh history that survives in three separate versions. The oldest, contained in a late eleventh-century manuscript (Harleian MS 3859), extends only as far as 954, which is probably close to the date at which this version was written. It has two separate entries: at ‘Year 72’, it reads Bellum badonis, in quo arthur portauit crucem Domini nostri ihu xp’i tribus diebus & tribus noctibus in humeros suos & brittones victores fuerunt (‘The Battle of Badon, in which Arthur carried the Cross of Our Lord Jesus Christ for three days and three nights on his shoulders and the Britons were victors’), while the second, at ‘Year 93 ’ reads Gueith camlann in qua arthur et medraut corruerunt, et mortalitas in brittannia et in hibernia fuit (‘The strife of Camlan, in which Arthur and Medrawd perished, and there was a great mortality in Britain and Ireland’). The B version of the text, written in the late thirteenth century, has the entries slightly differently: the first reads Bellum Badonis, in quo Rex Arturus portauit crucem Domini nostri ihu xpi tribus diebus & tribus noctibus in humeris suis. In illo proelio ceciderunt Colgrinus et Radulphus Anglorum duces (‘The Battle of Badon, in which King Arthur carried the Cross of Our Lord Jesus Christ for three days and three nights on his shoulders. In that battle, Colgrin and Radulf, leaders of the English, were killed’). We now have Arthur called “King” and the insertion of the names Colgrin and Radulf. The second entry reads: Bellum Camlan, in quo inclitus Arthurus et Modredus proditor suus, mutuis uulneribus corruerunt (‘The Battle of Camlan, in which the famous Arthur and Modred, his betrayer, perished from mutual wounds’). Here, Arthur is now “famous” and Modred has betrayed him. Where did these additional pieces of information come from? They derive from the early twelfth-century work of Geoffrey of Monmouth, of which more below.

The dating of the annals is a source of debate: unlike many other early medieval chronicles, it does not give dates anno domini but a list of years, which, in the A text, are divided into decades. The later entries can be dated with reference to other documents, which allows its first year to be calculated as AD 444. Unfortunately, it doesn’t work that logically: ‘Year 9’ corresponds to AD 451, as this was when Pope Leo confirmed the correct method of calculating the date of Easter, which would make 443 ‘Year 1’. Some decades as counted in the manuscript contain eleven years, some only nine, so when historians calculate the date for the Battle of Mount Badon as AD 516 and the Battle of Camlann as AD 437, we must ask whether that is a correct deduction. Badon is placed five years before the birth of Saint Columba and death of Saint Brigid, which are given in Irish annals as occurring in 522 or 523, which might suggest a date for the battle around 517 or 518. At a mere seven years before the death of Ciaran, calculated by Irish annalists as 548, then Camlann ought to be dated to 541. No matter how we look at these numbers, it is clear that we cannot trust them to give us accurate dates.

Other documents

There are no other historical sources naming Arthur before Geoffrey of Monmouth (c 1100-1155) composed his almost entirely fictional Historia Regum Britannię around 1136. True, there is literature in which his name occurs (not always as certainly as we might hope), but these are folkloric literature and poetry. They can tell us about how the legend evolved, not about the history behind the name. And it is quite clear that Geoffrey of Monmouth invented much of his story, which then became authoritative and infected all subsequent literature.

7 Responses to The documentary evidence for Arthur

  • Dr. John King says:

    To further compound the issue, Gildas attributes the victory at Mount Badon to Ambrosius. For both him and ‘Nennius,’ this fight is a climactic one, coming on the heels of several battles.

    It is easy to surmise that later tellings could conflate Ambrosius with Arthos, as one has many, if contradictory, stories of Ambrosius.

    It is also of interest that Gildas, likely a Pelagian, lacks the Roman-Catholic elements of Nennius’ later Arthur.

    These texts are, in and of themselves, fascinating, and I find it a little sad that they are subordinated to the need to verify a King Arthur.

  • Angus murray says:

    You quote: “pugnabat contra illos in illis diebus cum regibus brittonum, sed ipse erat dux bellorum”
    from HB but the quote is incorrect. It should read: “sed ipse dux erat bellorum”. It really annoys me that everywhere you read that Arthur was Dux Bellorum, interpreted as a title. If the passage read “sed ipse dux bellorum erat”, one might conclude that. However anyone who can read Latin knows that it simply doesn’t say that or even imply it. In theory it doesn’t exclude Arthur as Dux Bellorum, but it certainly doesn’t say he is either. The verb would usually follow the object of the sentence (subject, object, verb or object, verb, subject ) and so “sed ipse erat dux bellorum” would mean “but the leader of the war was he himself”, giving the false impression that Arthur might have been “Dux Bellorum”, since mediaeval Latin does not generally use capitalisation for proper nouns. Bad Archaelogy!

    • Keith Fitzpatrick-Matthews says:

      Oops! That’s my fault about the word order. I’ve corrected it now.

      I agree with you, although I should point out that the author of the Historia Brittonum throws in a lot of ipses that have no reflexive sense whatsoever; I wonder if we’re witnessing an author thinking in Old Welsh whose Latin is influenced by Old Welsh grammar and syntax. There are passages in Bede, for instance, where the Latin clearly reflects an Old English structure.

  • Angus murray says:

    Well its easy to make typos. I made two – bellorum is plural hence “wars” not war and archaelogy instead of archaeology. However, your word order suggests that you might – subconsciously at least – subscribe to the Dux Bellorum theory.

    You say:

    Historia Brittonum throws in a lot of ipses that have no reflexive sense whatsoever

    But surely that’s good Latin? Ipse is properly speaking intensive, se is reflexive. (hence “nisi ipse solus ” above where the intensive is appropriate to solus).

    You say:

    ‘Year 9’ corresponds to AD 451, as this was when Pope Leo confirmed the correct method of calculating the date of Easter, which would make 443 ‘Year 1’.

    Leo was still querying the calculation in 455 and it was not until 457 that Hilarius his archdeacon referred the matter to Victorius who devised a cycle to more accurately determine the date of easter. So Year 1 was actually 449. Add this to cycle Year 93 for Camlann and you get 542, agreeing approximately with the seven year interval before the death of Ciaran in 548 as determined from AoT and showing that far from “having an inordinate love of lying”, Geoffrey actually knew what he was talking about when he gives the date as 542.

    • Keith Fitzpatrick-Matthews says:

      Oh no, I don’t see dux bellorum as anything more than the author describing Arthur as a war leader: it’s difficult to see how he could have expressed it any other way. Perhaps “warlord” would be a good way of rendering it into contemporary English.

      If you look through the Latin of the Historia Brittonum, you find a lot of unnecessary uses of the word ipse and it looks to me as if the author uses it habitually as the third person singular pronoun.

      With regard to the chronology of the Annales Cambrię, it’s quite frankly a mess. Even if we take 449 as the start date, Year 93 would be 541, not 542: remember that you need to add everything to 448 (so 448+Year1 = 449, 448+Year 2 = 450, 448+Year 3 = 451 etc.). It’s too easy to get caught up in numerological blind alleys playing these sorts of games. I suspect that Year 1 was originally intended to be AD 444, the date at which the Alexandrian tables that continued to be used by the British Church until the late eighth century were composed under Bishop Cyril: it didn’t adopt the Victorian tables or their modified Dionysian form until then. The garbled entry for Year 9 reads pasca commutatur super diem dominicum cum papa Leone episcopo romę, which doesn’t have quite the meaning I imply in the main text. Indeed, “Easter was changed onto the Lord’s Day with [by] Pope Leo, Bishop of Rome may imply that the writer disagreed with what Pope Leo had done. Needless to say, Easter had always been celebrated on a Sunday, but I wonder if the compiler of the Annales was taking a swipe at Leo’s institution of the Victorian tables.

      451, of course, is the year of the Council of Chalcedon. Easter wasn’t something that seems to have been discussed at it.

      • Angus murray says:

        You’re quite right that you have to add 92 years (the difference between 1 and 93) to 449 for year 93 to get 541. At least that agrees with AoT.
        So would you agree that 451 is a myth? is it perhaps a typo for 457 that’s been oft repeated and not anything to do with the council of chalcedon?

        I’m sure that 444 was indeed the right calibration for later years but the likelihood is that the calibration is not the same for all years. The numbers of “an” don’t of course perfectly correspond to the major decadal intervals in roman numerals. Scepticism of the new table, I would imagine is very likely since Britain stuck with the 84 year cycle for a very long time afterwards.

        • Keith Fitzpatrick-Matthews says:

          Yes, I suspect that the entry for Year 9/“451” is wrong in one (or both) of two ways: it’s a confusion of the Council of Chalcedon with Pope Leo’s ruling on Easter Tables and/or it’s a garbled summary of the ruling put into the wrong year.

          The chronological scheme of the Annales Cambrię is a mess, particularly in the section before the 790s, when selected entries were copied from a version of the Irish World Chronicle and events of British interest added, according to Kathleen Hughes.

Agree or disagree? Please comment!