The Heroic Age, Issue 2, Autumn/Winter 1999

This paper¹ defends the accuracy of three dates relating to the reign of Vortigern given in the Historia Brittonum.² We shall be concerned chiefly with Chapter 66 (Morris 1980:39), which presents the reader with an impressive set of calculations and dates:³

This highly important passage purports to date three events in fifth-century Britain; the accession of Vortigern (in A.D. 425), the arrival of the Saxons in the fourth year of Vortigern (A.D. 429) and the battle of Guoloph in the twelfth year of Vortigern (A.D. 437). None of these dates, however, agrees with the more traditional dates for these events presented by Bede in Historia Ecclesiastica Gentis Anglorum (A.D. 731; Shirley-Price 1990) and the Anglo-Saxon Chronicle.⁹ For most scholars an "early" Vortigern as opposed to Bede's dating around A.D. 449 might be acceptable, but such an early date for the coming of the Saxons would be out of the question for them as well (cf. Hawkes 1989). Even so, some scholars have expressed views that are in favour of the accuracy of these early dates, especially Henry Munro Chadwick (1954), Leslie Alcock (1971), Stephen Johnson (1980), Horst-Wolfgang Böhme (1986), J.N.L. Myres (1986), Bernard S. Bachrach (1988), N.J. Higham (1994), and Michael Jones (1996). ¹⁰

The problem is of course that these three dates from Chapter 66 are not recorded in any independent contemporary fifth-century source. In fact, no surviving contemporary insular source comments on them,¹¹ and on the continent the closest observers, such as Prosper and the Gallic Chronicles are in southern Gaul. The first British glimpse is from Gildas (Winterbottom 1978), a century after the events in question, followed by Bede (Shirley-Price 1990) in the eight century, and Nennius, the supposed author of the Historia Brittonum in the ninth century.¹²

Though Bede is the earlier, he is considered more accurate, mostly due to his use of Gildas, who is still considered authoritative for the period. Until the late sixties however, the Historia Brittonum was used freely as a source, particularly because of the details it contains. Though it seemed clear, even then, that much of its contents should be treated with extreme caution, many of the dates provided had been considered correct. This culminated in large period-studies such as The Age of Arthur (1973) by the late John Morris, whose use of the Historia Brittonum led to much criticism among historians.¹³

Long-overdue¹⁴ textual study of the Historia Brittonum from the 1970s onwards, established its untrustworthiness with regard to dates. The relatively late date of the Historia's compilation in relation to the events it describes, its indiscriminate use of myth and legend alongside more accepted history, but also the apparent inability of the anonymous compiler to separate "Anno Domini" and "Anno Passionis" dates (Dumville 1972-74:439-441),¹⁵ contributed much to this conclusion. To establish the true historical value of the text, the historical facts had to be separated from the fiction, especially where early English myth and legend were concerned. Chapters on Arthur were therefore rejected, and events with seemingly precise dates such as the accession of Vortigern, the Adventus Saxonum and the Battle of Guoloph shared this fate. Several authors came to the conclusion that these dates had no historical validity at all.¹⁶

Tolstoy

The first of these critics was Nikolai Tolstoy (1962:52-153), who suggested a possible solution to the problem of the discrepancies between the Vortigern dates in the Historia Brittonum and in Bede in a passage from Chapter 31 of the Historia Brittonum:

According to Tolstoy, the key to the problem of apparently irreconcilable dates¹⁸ was a mistaken calculation. This calculation added the forty years in which "the Britons went in fear" to the year A.D. 388 (Maximus's death), to give the supposed date of the arrival of the Saxons (A.D. 428).

Tolstoy (1962:152) considered this supposed calculation a mistake, for this forty years should clearly have been added to the real "end of the Roman Empire in Britain", namely A.D. 409. This would give A.D. 449 for the Adventus, which fitted perfectly with the traditional dates of Bede and the Anglo-Saxon Chronicle. ¹⁹ The accession of Vortigern would accordingly be pushed forward to A.D. 446, a date acceptable to most historians as it occurred together with the "appeal to Aetius"²⁰ . The conflict with Ambrosius at the Battle of Guoloph, now dated to A.D. 458, would therefore agree with Gildas and also with Bede's dating of Ambrosius Aurelianus some time after 449. Though Tolstoy's solution reconciles the dating of Nennius with the traditional dating of Vortigern and the Adventus Saxonum, it remains unsatisfactory as all the three assumptions upon which it is based are questionable:

Tolstoy argued that the dates as originally calculated by Nennius were irreconcilable with what we know of fifth-century history. But this assumption is no longer tenable. Since Tolstoy wrote in the early 1960s, archaeology has put forward evidence²¹ for a more complex and less traditional transition from the Roman empire during the first half of the fifth century. The possible dating of such a transition is perfectly reconcilable with Nennius's dates.

Tolstoy's second assumption is that the period of forty years leading up to the arrival of the Saxons was originally added to the date of the death of Maximus, i.e. to the year 388. This is by no means stated by Nennius, whose statement about the "forty years" just happens to be followed by the mention of Vortigern and the subsequent Adventus Saxonum. In my opinion there is nothing that warrants a direct chronological connection between these two sentences. The sole fact that the sentence "and after the killing of the tyrant Maximus and the end of the Roman Empire in Britain, the Britons went in fear for 40 years" is followed by "Vortigern ruled in Britain" does not mean that Nennius thought Vortigern succeeded Maximus (as both Dumville and Miller claimed, below) or that the Saxons arrived in Britain 40 years after Maximus' death (as Tolstoy claimed). This problem has received no further attention from subsequent critics.

Tolstoy's third assumption is that Nennius was ignorant of the traditional dates for Vortigern and the Adventus Saxonum as presented by Bede. Had he been so, the discrepancy between his solution and theirs would have been obvious immediately. To have accepted this discrepancy would have made Nennius look at least very careless. But Nennius was well aware of these dates, for though Bede does not seem to be included on the impressive list of sources apparently available to Nennius,²² his use of English material makes it extremely unlikely that he had no access to Bede's work or disregarded it deliberately. Therefore, either his calculation was correct or he had access to another set of dates, now lost (I will deal with this possibility below). Tolstoy (1962:152) did indeed acknowledge that there might have been such a possible source, an "original core" which provided the "fourth year" for the Adventus and the "twelfth year" for the battle of Guoloph.

Dumville

The next to tackle this problem was David Dumville (1972-1974). His conclusion was that Nennius was attempting to write a synthetic history (Dumville 1986:1) and as a result looked for the dates of the Adventus Saxonum and the accession of Vortigern. Dumville (1972-74:444) states that Nennius computed these dates himself, using only two sources, Prosper's Epitoma Chronicon (Mommsen 1894-98) and Victorius's Cursus Paschalis (Mommsen 1894-98). After he had made the calculation as to the date he wanted, Nennius subsequently added references to the contemporary Roman consuls from Victorius's Cursus Paschalis (Dumville 1972-74:445) to add authority to his text.

Like Tolstoy, Dumville (1972-74:445) saw the statement in Chapter 31 as the key to the problem. Nennius would have discovered from Prosper that Maximus's death occurred at the hands of the Emperors Theodosius and Valentinian in A.D. 388. To this he would have added the period of forty years to arrive at the date for the Adventus Saxonum as proposed by Tolstoy. But unlike Tolstoy's solution, Dumville's solution also aligns itself with a statement in Chapter 29:

This is a quotation from Prosper's Chronicle, though altered by the substitution of the word "consulibus" for "imperatoribus". According to Dumville (1972-74:444), this showed that Nennius believed that emperors gave way to consuls, a view supposedly confirmed when he found in the Victorian Cursus Paschalis that Valentinian and Theodosius were consuls in A.D. 387 and 388.

This in turn led to Dumville's explanation of the impressive chronological accuracy of the equation of the Adventus Saxonum with the fourth year of Vortigern. As Nennius supposedly wished to establish the accession of Vortigern, Dumville states that Nennius turned again to these tables to find both consuls who brought the reign of Maximus to an end, and found both Valentinian and Theodosius mentioned together for the first time in Anno CCCXCVIII (A.D. 425). Nennius then supposedly equated these consuls in A.D. 425 with the emperors in A.D. 388. Hence, since Vortigern succeeded Maximus, his accession occurred in A.D. 425. This meant that the Adventus Saxonum took place in A.D. 388 + the "Forty Years of Fear" = A.D. 428.

Thus the Adventus Saxonum occurred in the fourth year of Vortigern. Dumville considered the appearances of the "1st year" and the "12th year" as useless, for they were only generated by Nennius, solely for the purpose of establishing the dates. Dumville (1972-74:445) concluded that this impressive chronological precision in dating the Adventus Saxonum was illusory and that there was no possiblity of some authentic native record behind the synchronisms. Yet this solution is not satisfactory. It rests upon a number of mistakes and assumptions that can be disproved.

First, Dumville states that Nennius equated the accession of Vortigern "with the period following Maximus death",²⁴ for he looked for the first occurrence of Theodosius and Valentinian together as consuls to find the date for the accession of Vortigern. I believe this to be a fundamental mistake. Nennius clearly did not equate the death of Maximus (or even the "period" thereafter) with the accession of Vortigern. The enormous discrepancy in the length of the period - no less than thirty-six years - would of course have been clear to Nennius. He clearly proved this with his statements that the Adventus Saxonum was in the fourth year of Vortigern and that there was a period of "forty" years between Maximus's death and the accession and/or the Adventus Saxonum. Even more to the point, Prosper, who was one of the sources used by Nennius, ascribed the fall of Maximus to Theodosius and Valentinian "sub anno" 388. How Nennius would have managed to use Prosper's dates and still choose 425 (unless he was being deliberately fraudulent) is inexplicable.

Second, in proposing his solutions Dumville assumes that Nennius, apparently muddling his A.D. and A.P. dates, fails to distinguish between the emperors of the fourth and the fifth century, confusing the emperors Theodosius and Valentinian with their successors. But Nennius would have surely noticed in his source (the Victorian Cursus Paschalis) the names of "Stilicho" and "Aetius" before the year he supposedly equated with the death of Maximus. He would then also have known that the "consuls" he found in A.D. 425 were not the Theodosius I and Valentinian II of the fourth century, but the Theodosius II and Valentinian III of the fifth. Nennius is not likely to have made such an enormous mistake, for he states in Chapter 29 that Theodosius and Valentinian ruled together for "eight years".²⁵ Yet the Theodosius and Valentinian he supposedly found in the Cursus Paschalis appear over a period of "ten years".²⁶ This could be explained as a simple mistake, but taken together with the rest of the evidence I propose that Nennius indeed knew what he was writing about.

Third, Dumville assumes that Nennius looked in the Victorian Cursus Paschalis for the appearance of Theodosius and Valentinian together to deduce the accession of Vortigern. But Dumville fails to explain why Nennius specifically had to choose for the year A.D. 425 for this event, instead of, e.g., A.D. 426, where both names occur as well, as they do in other years.²⁷ The names of Theodosius and Valentinian also appear in several other years by themselves,²⁸ as they of course do before the death of Maximus.²⁹ As there is no reason why Nennius would not have chosen any of these dates, the supposed choice of this date now becomes difficult to explain.

As I have shown above, the suggestion for the calculation of the year A.D. 428 is not sufficient. The solution for the date of the accession of Vortigern in the year A.D. 425 has also been disposed of. Therefore Dumville's proposed solutions for the mechanics behind the supposed calculation of dates by Nennius' remain unconvincing. Dumville has never attempted to explain the origin of the remaining date, i.e. the twelfth year of the reign of Vortigern, the battle of Guoloph.

Miller

Molly Miller is the third and last of the authors to be discussed here. Her solutions to the problems of Chapter 66 are quite different from those of her predecessors. Not only did she attempt to explain the "twelfth year" of Vortigern and the seemingly unrelated battle of Guoloph, but she developed a solution based on the text itself instead of looking at various calculations within the text. Though Miller agreed with Dumville that Nennius (or a source) might have believed Vortigern to have succeeded Maximus directly in A.D. 425 (1977-78:52), she proposed two different reasons to dispose of the resulting supposed accuracy.

Miller's first reason is that the dates in Chapter 66 of the Historia Brittonum might have been based on the pedigree of Gwrtheyrnion contained in Chapter 39, which is probably the only evidence that might confirm them. This pedigree would support the adult lifetime of Vortigern falling within or overlapping the years from 425 to 462-63.³⁰ Though this source is not contemporary and the pedigree uncertain, it cannot be completely disproved (Miller 1978:52-55).

Miller's second, more intricate, explanation deals with a possible mistake made by Nennius. Miller based this upon a possibly ninth or tenth-century document, Chronica Minora I (Mommsen 1892:515-566), containing several manuscripts,³¹ especially the Gallic Chronicle, erroneously attributed to Prosper.³² The entry in this text under Honorius XXX is a duplication of Maximus's death: "Maximus tyrannus de regno deicitur" ("Maximus the usurper was driven from his reign").³³ Since the fifth entry before this one³⁴ clearly mentions the end of Maximus, this one is not correct. Yet it seems possible that this text, attributed to Prosper, could have led to the confusion of Nennius (or an earlier source)³⁵ in choosing the date for both the death of Maximus and the accession of Vortigern.³⁶

This solution remains unconvincing, because Miller fails to create a link between the sources used by Nennius and this manuscript. Nennius is known to have used Prosper, who does not make the mistake of erroneously duplicating Maximus's death. No evidence exists that Nennius used or even had knowledge of this Gallic Chronicle.

Miller further tries to explain the calculations of Chapter 66 by attempting to "reconstruct" the text itself. This reconstruction is based on the assumption that the original text once consisted of two columns (Miller 1980:26). Because of this, the entry of "et a regno Guorthigirni", originally in the middle of the chapter (above), should in fact be placed at the very end. Furthermore, the word "regno", originally translated as "accession", which to Miller (1980:27) cannot be right, has been changed by Miller to "reign". To do this, the "a " ("et a regno") has to be removed as a copyist's false correction. Miller now translates the entry as "And for Vortigern's reign are twelve years". This not only justifies the removal of the entry to the end of the chapter, its new translation immediately explains the connection to the Battle of Guoloph: the "discordia" that in this explanation marked the end of Vortigern's reign.

Miller's arguments look very impressive: Nennius, a source, or someone else responsible for transcribing the Historia Brittonum appears to have been convinced that Vortigern ruled for twelve years, a reign ended by a "discordia" at the Battle of Guoloph. She herself urges caution,³⁷ for though Miller has solved the problems, she admits that her solution remains, unfortunately, little better than juggling with words, as too many scribal errors and other such assumptions have to be taken into account. Therefore, apart from possible confirmation by the pedigree, Miller supplies no new facts. The date of the Adventus Saxonum in A.D. 428 remains unexplained.³⁸

Annals

This leaves the question as to what does explain the origin and dating of the entries for Vortigern's first year (accession), his fourth year (Adventus Saxonum) and his twelfth (the battle of Guoloph)? Explanations include a simple mistake (Tolstoy), ghost-dates calculated by Nennius (Dumville) or a muddled transcription of the text (Miller). But when examined, none of these seem satisfactory. A last remaining option is one of of a real underlying source for the dates in the Historia Brittonum, possibly a contemporary set of annals.

As we have seen above, Tolstoy (1962:152) acknowledged there might have been a "core", that stated that the Adventus Saxonum occurred in the fourth year of Vortigern. Even Miller believed it certain that the dates listed in chapter 66 came from an older document incorporated into the Historia Brittonum.³⁹ Though this option was rejected by Dumville (1972-74:445) we have seen that his reasons for this rejection are not satisfactory. The advantage to this solution is that no change to Nennius methodology is neccesary. Nennius may have found the year elsewhere and provided the consuls accordingly, instead of finding the consuls first and only then finding the year. The problem is of course that hardly any evidence of such a source exists.

The evidence that we have is very slight, nor does it point to a set of annals that might underlie the list behind chapter 66. Yet it does provide evidence that there were contemporary insular sources with knowledge of Vortigern. It was Dumville (1973:312-14) who pointed this out with reference to a ninth-century chronicle-fragment⁴⁰ that provides a possible source. Dumville thought it a short British chronicle which shows an early form of the name "Vortigern":

This passage is part of only twelve enties, starting at the year 60 B.C. and cut short after A.D. 565, where Bede continues. The chronicle is based (Dumville 1973:312) on Bede's annalistic summary in his Historia Ecclesiastica V 24 (Shirley-Price 1990:325-6), but with additional information (in italics here). Though dating back to a French scriptorum in the second half of the ninth century the chronicle is almost certainly English, originated up to a century earlier (Dumville 1973:312). However, the sources of this chronicle might have included a very early British set of annals.

The evidence for this is the form of the name of Vortigern. The very early version of Vortigern's name (Uuertigerno) is considered by Kenneth Jackson to be evidence for a possible fifth-century source for the chronicle fragment.⁴² Though the spelling of Vortigern's name might have been drawn from another version of Bede's De Temporum Ratione, where the form Vertigerno is used, this certainly need not have been the case. Bede's form of Vertigerno probably went back to a British original of *Wortigernos (Jackson 1982:36). Though it certainly proves that Bede used very early sources, it must be stressed that, apart from Gildas, such sources in fact did exist.

We may assume that, based on such early forms of the name of Vortigern, at least in theory, sources contemporary with Vortigern existed in the period immediately before Bede and Nennius wrote. Though Nennius did not use this form of the name, we need not conclude that he had no access to such early sources. We know he used Isidore of Seville and very probably Bede as a source and must therefore have been aware of early forms as Uurtigernus or Vertigerno. That he used Guorthigirnus instead was probably only for the benefit of his (ninth-century) Welsh audience.

Conclusion

The purpose of this paper was to re-examine the validity of the dates in Chapter 66 of the Historia Brittonum, by examining the criticism of Nennius supposed methods. In the light of the enormity of the mistakes neccessary, we can safely rule out the idea that Nennius knowingly looked in the Victorian Cursus Paschalis for the names of Valentinian and Theodosius to establish the date of Vortigern's accession. He did not need to, for we can safely assume that he did not equate the accession of Vortigern with either the end of Maximus's reign or the "forty years" that supposedly followed it. What he did add from this and other sources was the names of the consuls, which in all probability were not present in any early British source. He also made several calculations, thereby muddling several sets of dating and showing his lack of knowledge and insight into these sources. We have no grounds for thinking that he carelessly or deliberately corrupted his sources in any way. Furthermore, though he was aware of Bede and the (later) dating of Vortigern in that source, he still chose to reach a very different conclusion.

We can therefore no longer totally rule out the idea that the list of dates in Chapter 66 of the Historia Brittonum was indeed based upon an underlying earlier source, now lost. Concerning the accuracy of the dates for Vortigern in Chapter 66 of the Historia Brittonum, we may therefore conclude, at least speculatively, that they could very well have been taken from a fifth-century British source.