The story of Brutus is a most remarkable example of how an elaborate legend can develop out of practically nothing, under the circumstances that prevailed then. A few statements here and there in the Chronicle of Eusebius and elsewhere were taken out of context and misunderstood and synthesized into an explanation of the origin of the Britons as it related to the origin of the Romans. In the struggle to understand this and reconcile it with known history, further editors conflated several "persons" who could be reckoned as an eponym of Britain into a character (named Brutus in one place and Britto in another), who fit into the story of the Aeneid origin of the Romans. Still further editors interpreted all this as any reasonable but ignorant person could be expected to, and formed it into as coherent a story as possible. We can see how this developed only by comparing the wording and structure of the final product to the classical sources we possess on which it clearly depended. When we turn to other legends, whose sources we do not possess, we can only wonder whether or not they have suffered from the same sort of phenomenon.
It seems that no one who built upon the work of Nennius gave the proper credit, except the Irish recensions, the only ones in which his name is preserved. Still, we can clearly see the debt owed by HA, HRB, and, if we look very closely, certain genealogical works.
The Historia Anglorum (HA) of Henry of Huntingdon was an effort to synthesize all the available knowledge on the history of Britain. It draws heavily from HB, from Bede, and from common classical sources, but shows almost no knowledge of anything that might have been added to the statments of Nennius, except in a very small number of cases. For example, it knows of Lud, and has identified St. Helena as a daughter of Coel of Colchester [2].
A few years later appeared the Historia Regum Britanniae (HRB) of Geoffrey of Monmouth. Geoffrey delivered not only the additions found in HA, but also much, much more, which even now defies explanation, unless, as is often supposed, he was a liar with a brilliant imagination. Geoffrey claimed in his preface to have been given a vetustissimus liber, an ancient little book in the British language, which was his principal source, though he also drew from Bede, Gildas, and others. The dependence on HB is obvious, but is never explicitly acknowledged. So, it has often been supposed that the mysterious book in question was none other than HB, and the copious additions were fictions created by Geoffrey.
The mention of Gildas is a curious point. Both HA and HRB, as well as other historians for centuries to come, refer to Gildas as the source for information that in most cases is obviously quoted from HB. In several cases HRB specifically attributes statements to Gildas that are found nowhere else. The only known historical work of Gildas was his De Excidio Britanniae, more of a diatribe really than a useful history. It was known to Bede, and in one place HRB correctly quotes from it, though it is difficult to believe that the compiler of the latter himself had access to any more of it.
Seeing that HA contains certain small additions not found in any extant versions of HB but present in HRB, it would seem that both HA and HRB drew from a slightly more developed recension of HB that has not survived. This recension would have mentioned Lud and Coel, as well as Brutus and Arthur and everything else for which HB is famous, but would apparently have been attributed to Gildas. This is the work that HRB cites as a source in its preface, and is distinct from the vetustissimus liber, although the latter may have drawn from it. There also are two late manuscripts of HB (called P and Q) that purport to have been written by Gildas. During the renaissance, it therefore seemed that HB must have been written by a man named Gildas, who was been distinct from three earlier men of that name, and was therefore called Gildas Quartus, whom Bartrum labels a fictional person [3].
One possibility is that the name of Gildas became attached to HB without much change to the text, and later an important recension of it was made, with the aforementioned additions, and used by HA and HRB. Another possibility is that this recension first took on the attribution to Gildas and became widespread and well known, and later when a copy was made of the earlier recension it was recognized as the work of Gildas and so identified. Either way, it would seem that the name of Gildas found its way into P and Q purely by reputation. With the first possibility, Gildas would be the famous sixth century historian, so the reasons for associating him with the work are not very strong, but with the second possibility, it may well be that the author of this new recension was in fact named Gildas, which would be a very good reason for inserting the name and would inevitably lead to the confusion observed. So, for the sake of discussion, let us call this recension the Gildas recension, and its author Gildas Quartus.
Subtle but unmistakable traces of HB also continue to turn up in the genealogies. The relevant section is the pedigree of Brutus in HB 18 (here quoting the Vatican MS):
Brittones a Bruto. Brutus filius Hisicionis, Hisicion Alani, Alanus filius Raea Silueae, Rea Siluea filia Numae Pampilii, filii Ascanii, Ascanius filius Aeneae, filii Anchisae, filii Troi, filii Dardani, filii Flise, filii Iuuani, filii Iafeth... filii Noe....Anyone with access to standard classical texts could see that there is an error here in the Trojan kings, which should be emended to show: Anchises son of Capys son of Assaracus son of Tros son of Erichthonius son of Dardanus, said to be a son of Zeus (or Jupiter). However, none of the extant manuscripts of HB show any attempt at such an emendation. This genealogy, along with the related genealogy of Britto in HB 17, were superseded by the account of HRB, which, as in HB 10, showed Brutus as a son of Silvius. It appears one last time in a late miscellaneous compilation, ByA 33 (see EWGT):
Elen f. Eudaf m. Karadawc m. Bran m. Llyr lletieith m. Garar m. Gerein Hir m. Secuyn m. Keit m. Arth m. Moruran m. Kerint m. Cridol m. Dingarth m. Annun m. Lainus m. Brutus m. Ericonius m. Alanius m. Reesilue m. Nenne m. Parapilius m. Ascanius m. Eneas m. Anchise m. Troys m. Dardan m. Iauan m. Iaphet m. Noe.After HB, the next time we see a full pedigree of Brutus is in GaC 2:
...Brut tywyssauc o Ruvein m. Siluius m. Ascanius m. Eneas ysgwyt wyn m. Anchises m. Capis m. Assaracus m. Trois m. Herictonius m. Dardanus m. Iupiter m. Sadurn m. Celius m. Cretus m. Ciprius m. Ketun m. Iauan m. Iaphet m. Noe hen....The emendations have been effected, and the Trojans are now given a new longer pedigree back to Noah. No other extant sources even attempt to make such a link [4]. Lebor Bretnach, the Irish recension of HB, greatly modifies the line, making the aforementioned emendations and tracing the line through Ham rather than Japheth with further spurious names [5], but in Britain there seems to be no possible basis for this pedigree besides HB or a derivative. The modifications observed here might have been effected after the genealogy was extracted from HB, but it seems more likely they were effected within a later recension of HB itself, perhaps that of Gildas Quartus.
The most extraordinary thing about this is that a nearly identical bit of genealogy turns up in the Icelandic Langfeðgatal, which tabulates the ancestry of Óðin as follows:
Óðin Frjálafsson Finnsson Goðólfsson Játsson Bjáfsson Skjöldarson Hermóðsson Itormánsson Aþrasson Bedviggsson Seskefsson Magason Móðason Vingenersson Vingiþórsson Einridasson Hlóriðasson Þórsson Tróánsson Priamsdóttir Lamidonsson Ilusson Ereásson Erichonísson Dardansson Jupitersson Saturnisson Celísson Cretisson Ciprisson Zechimsson Javansson Japhetsson Nóasson....It is all but inconceivable that this ancestry of Dardanus has been copied from GaC, and equally unlikely that GaC has imported it from Iceland. What we do know is that Iceland received at least one copy of the Saxon genealogies, which contain the sequence from Woden (Óðin) to Sceaf (Seskef, apparenty a misreading of se Sceaf in the ASC text that proceeds to make him a son of Noah) [6]. From there to Þór (i.e. Thor) is a sequence of seven names that are each mentioned elsewhere as bynames of Þór. Tróán daughter of Priam is not recorded in classical sources and appears to be nothing more than a corruption of the word Trojan. This is the one other instance in medieval genealogical literature, at least before the forgeries of Annius of Viterbo, where a royal line is explicitly traced to the Trojan kings and thence to Noah, and the pre-Trojan portion of the pedigree is essentially identical.
The first appearance of this ancestry of Woden is in the Saxon genealogies [7], which go progressively farther with later versions. A very early version found its way into HB, one which lacks the Wessex pedigree and goes only as far back as Geat (i.e., Ját above), who is acknowledged as a pagan god. The extension to Sceaf was widely available in the time of Nennius, and not long after, ASC had implicitly identified Sceaf with Shem, making him a son of Noah and thereby providing a route all the way back to Adam. The Icelandic version of the ancestry of Óðin first appears in the Prose Edda, about the same time that GaC was composed, but there Sceaf is no longer a son of Noah, but is given an extended pedigree back to Priam, whose ancestry in turn is not discussed at all.
It has long been supposed that Snorri forged the additions himself, while the further additions in Langfeðgatal are left an enigma. However, there grave difficulties with the theory that Snorri is responsible for the genealogy in his prologue. He leads himself into a contradiction, since later in the Edda he tells the creation myth where Óðin is a son of Bor son of Buri, the first man, while the prologue makes him a son of Frjálaf Finnsson and continues with a long pedigree. Further, if Snorri had any interest in linking Óðin to the Trojan kings, it is odd that he did not even bother to state the ancestry of Priam, which was readily available to any scholar who knew who Priam was. It seems rather that Snorri, like so many Icelanders of his day, has no great interest in such things and is merely repeating this genealogy in a prologue mainly for the sake of completeness. Since the kernel of the genealogy was already circulating in Iceland before the time of Snorri [8], there was enough opportunity for it to develop into the longer form, which Snorri simply copied.
Since GaC and Langfeðgatal seem to have drawn from a common source, and that source seems to have drawn ultimately from the extant versions of HB, it is natural to suppose that this intermediate source is one or more later recensions of HB, now lost. Let us suppose that shortly after the Nennius recension, a new edition, which we shall call the Gildas recension, was made, in which the Roman period material was expanded a little and the material on the prehistoric genealogies was brought into better order, by emending the section on the Trojan kings and finally settling on the name Brutus (rather than Britto). The Gildas recension or its derivatives would then be used by HA, HRB, and GaC, while one particular derivative would be picked up by the English.
This English recension would be molded by the different viewpoint and source materials held by the English. In a way, it was doomed from the start, since a book that glorified the illustrious history of the Welsh and exalted their lineage and thus their claim to the island would not go over well with English audiences. But it did tell the story of Hengest, and give a substantial section on the Saxon genealogies, including parts that had been omitted entirely in ASC, and provide interesting information on the history of the land. The English editor, having the longer ancestry of Woden then current but not yet the final version in ASC, would replace that section with his own better information, and see an obvious point for comment. If the native British people were descended from the Trojans, and as such were the legitimate rulers of the land, the English claim could only be good if the English people shared in such a descent.
We have seen with the example of Brutus how small, innocent statements can be lucky enough to snowball into a grand legend if they are of the right flavor. If the English editor said anything about how the descendants of Woden had a right by blood to rule the land, another editor might then make this more explicit by saying that Woden was descended from the Trojans, and another might add more to this, until the ground was laid for the leap to be made by the Icelandic recension (or, for all we know, a Danish recension, though it seems less likely). The source for the Icelandic recension, which by now might have shed much of the uninteresting Welsh material, gave in one section information on Brutus, the Trojan kings, their descent from Noah, and so forth, and in another section the ancestry of Woden back to Sceaf, some notes on Sceaf, some notes on the ultimate descent from Priam, and some notes that could be interpreted as referring to Thor. This editor then had a simple task of filling out the genealogy, by arranging the seven names of Thor (from his own sources) after Seskef, mentioning Memnon and a Trojan daughter of Priam where appropriate [9], giving the known ancestry of Priam, and finishing by copying the ancestry of the Trojan kings already given. Even if this editor did not intend for the sequence he wrote to be a pedigree precisely, those who extracted genealogies from it would inevitably assume so.
Once this was done, the Icelandic recension had lost its raison d'être and perished. The extension of the pedigree of Woden made its way to Snorri, who begrudgingly reproduced it as far as Priam in his prologue. In its full form it was included in the Langfeðgatal. Once a mythical genealogy has reached Noah, it is then more or less solidified in its final form, having nowhere left to grow. The idea of linking Óðin to the Trojan kings and thence to Noah would never have occurred spontaneously, or in any case have succeeded, if it were not for the catalyst provided by HB, but once it was done, the result was so appealing that it could never be undone.