"A" saw the massacres as the central dominant element in the relationship of Britain and Rome. The Massacre became a reality across time, repeated thrice at three fundamental moments, at each of which it alters the position of the two peoples; and the notion of a Massacre of Roman duces reached further than A, haunting Welsh legend throughout its history. A later development gave us the better-known and certainly more durable version that developed into three surviving accounts, Nennius' Seven Emperors, Geoffrey's Roman stories, and The dream of Maxen Gwledig; it knew of only one massacre at the very end, and personalized the crisis of Roman power by incarnating it in one man, Maximus/Maximianus/Maxen, who put an end to Roman power by taking the Host planted by the Romans back to the continent. This version places its more limited view of the massacre firmly at the end of the Roman period, but is no less clear that it happened and that it was a decisive moment in the island's story^[1].

The enormous importance of the three massacres in A suggests a world that still felt relations with the Romans as a serious issue; the use of a single massacre as a historical marker with no extra significance in the Seven Emperors suggests no such strong concern. In the legend of the Seven Emperors, the Romans seem to keep away from Britain out of fear; in A, out of disgust - they have already come back to help the British once, and their reward has been another massacre. This implies that they could come if they wanted to, that is, that they are regarded as a real and present Power; whereas in the Seven Emperors version, they are not regarded as so powerful or as so menacing that one good Massacre will not make them think twice - in other words, their power and presence is felt far less.

This shows the mind of the age of Nennius, in which what was known as the Roman empire was (were) weak and far away. The fifth and sixth centuries moved from the fading but awesome memories of the Western empire, whose tail was still thrashing in Gaul by the 460s and 470s, to the threatening reality of Justinian's seaborne empire, with its intense trade with British harbours. By Nennius' time, however, the empires were two, both kept very busy by Arabs and other barbarians, and neither with the power or inclination to have much to do with Britain. The fact that Nennius manifestly favours the legend of the Seven Emperors and manifestly neglects A speaks for itself; the former represents contemporary thought, the latter is an ancient oddity. He deserves our thanks for mentioning it at all.

But though every story includes a massacre - one might almost say: a massacre, any massacre - its position in the story is as shifting as quicksilver. We have seen that the legend of The dream is close to Gildas' in many ways, but Gildas places his only massacre at the beginning, not at the end of the Roman period. The legend of the Seven Emperors, on the other hand, places it at its very end. This is actually closer to A than to The dream, where a massacre of Roman leaders is the precondition of Maxen returning to power; in Nennius 28, it comes only once Maximianus is dead.

This is best explained by the triumphalistic atmosphere of The dream, which, as we have seen, reverses the meaning and content of many of Gildas' points. Its every point is meant to glorify Maxen and the Welsh nation, and part of it was to turn the massacre of the magistrates into a glorious episode. But if the author felt the need to whitewash the massacre, it means that in his time (and The dream is certainly many centuries later than Gildas, let alone A) it was particularly linked with Maxen, who always is, wherever he appears - in Gildas, in Nennius, in triadic tradition, in Geoffrey - the last Roman emperor in Britain.

Gildas censored the closing massacres from his version. His reasons were both his detestation of British treachery, and his acquaintance with the reliable, Continental historical sources I have called B, which told him that continental records knew of no such massacre; but his version is secondary to that of A, in which the massacre was nothing less than a fundamental element of the whole relationship between Britain and Rome - and which probably had nothing to do with Maximus. He did not manage to suppress the idea that Rome left Britain because of a Massacre, and therefore the history of subsequent developments cannot be said to depend entirely on him.

On the other hand, I suspect that it may be he who introduced the figure of Maximus and characterized him as the jumped-up British tyrannus. There is nothing in the later Maximus/Maximianus/Maxen legends that cannot be led back to Gildas' negative portrait of the usurper, certainly including the dynastic fiction of a Good Maximus set against Bad Maximianus; and there is nothing in Gildas which cannot be led back, either to Sulpicius Severus' historical record, or to Gildas' own design. He had three clear intentions: a), to get rid of the ridiculous idea that the Romans could be scared away from Britain by a simple massacre, b), to vigorously make the point that the high and royal Roman race no longer exists in Britain, having left the island in 388, and, c), to attribute to the succeeding political leadership - except for Ambrosius, "almost the last of the Romans", and his descendants - the character of "native" tyranni or teyrnedd. While we cannot show that someone before him had already made use of the historical Magnus Clemens Maximus, later usages bear the mark of these typically Gildasian concerns, concerns in which he was in rebellion against the whole mentality of his time. In other words, the later Maximus-legends bear the marks of Gildas’ own individual revolt against his age. Even so late a product as Maxen seems to purposely reverse Gildas' point: Maxen defeats the representative of the native tyranni, Beli son of Minogan, and represents the higher, uniting Roman monarchy, with his bride Elen building roads. This suggests that someone in the thirteenth century still understood the point of dividing monarchy between a higher (Maxen) and a lower (Beli) rung, though the specific terms teyrn and gwledig seem to have lost their differential meaning.

Now the native legends of the Massacres are not without an echo outside Britain. The notion of a nativist British revolt against Rome^[2] has long had a footing in what one might call "believed history", the history for which there is no evidence but which sober professional historians - rather than cranks - think credible; thanks to a single passage in the Byzantine historian Zosimus, which, apparently, no member of the congenitally sceptical tribe of historians has ever thought of challenging. Zosimus’ unfinished New History, the same work which mentions the Rescript of Honorius, states that the British expelled the Roman magistrates and got rid of Roman law, reverting to their own customs, in the course of a great war against "the tribes beyond the Rhine" in which they are said to run the most fearful risks (6.5.3/286-7).

In view of what we have seen of the nature and origin of the legend, do we believe this? This revolt against Roman law is unknown to every other writer who had occasion to write about fifth-century Britain; which, considering the evil reputation Britain enjoyed at the time, is at least surprising. Quite to the contrary, the church writer Prosper Tiro of Aquitaine called it, as late as 434, "the Roman island", as compared with Ireland, which was "barbarian": a clear description, from someone who, as we will see, can be shown to have taken a close interest in British matters. In the same period, the compiler of a list of official posts known as the Notitia Dignitatum regarded Britain as part of the Roman Empire^[3]. It is also in direct contradiction with Zosimus' own mention of the Rescript of Honorius (6.10.2/291), where he clearly dates the Rescript as a measure taken during Alaric's invasion and the end of Constantine III's adventure, among a list of specific Imperial enactments dealing with the after-effects of the disastrous year 410.

The war against the Saxons and the rejection of Roman laws and magistrates come five chapters earlier and are in a different world altogether. With the Rescript we have a clear date, a clear political context, mention of the issuing authority; the ejection of the magistrates could not be more vague, with not a single date, name or place. "As they advanced, the barbarians from beyond the Rhine gained control of everything^[4], and brought the men of the British island and some who dwelled among the nations of the Celts [oikontas twn en keltois eqnwn enia] to the necessity of rejecting/revolting against [aposthnai] the Roman principles/government [archs] and live in their own way without submitting to their [the Romans'] laws. Those in Britain therefore armed themselves and ran terrible danger to free their ciuitates from the menacing barbarians, and all Armorica and other provinces of Gaul imitated the British and likewise made themselves independent, threw out the Roman magistrates and set up their own independent government."

Zosimus is universally admitted not to be a clear-headed or insightful historian, and this fairly evident self-contradiction seems typical. On the other hand, much of his information comes from Olympiodorus of Thebes, as reliable a source as he himself is unreliable: "an active politician, ...[whose] diplomatic missions included several embassies to the Huns and other barbarian peoples, who held him in great esteem. Olympiodorus describes his work not as a history, but as source materials for a history... [it] covered the years 407 to 425... composed in annalistic fashion, using consular years for dating events. John Matthews points out the great precision that Olympiodorus used in technical matters, accurately transliterating Latin bureaucratic directly into Greek."^[5]

Now, of the two passages, the one that is far likeliest to come from Olympodorus is 6.10.2, mentioning the sort of materials which Olympiodorus collected - to wit, an imperial decree dealing with a major matter of foreign policy - and placed among a number of other dispositions by Honorius, dealing with the effects of the crisis of 410. 6.5.3, By contrast, shows no sign of depending in any way on official reports or decrees; the manner is rather that of a strikingly vague summary of a narrative, and nothing suggests an official record - not a name, not a date, not a clear event. It is as fuzzy as Zosimus' own mind; while 6.10.3 is as clear as that of Olympiodorus. If we have to choose, 6.10.2 is the obvious choice.

What is more, 6.5.3 contradicts not only Zosimus, but everything else that is known about the period. Not only is there, as I said, no other mention of this supposed anti-Roman rebellion whatever, but, in a well-reported, multi-sourced period of history, its supposed spread cuts across Roman and barbarian activities in several provinces, and yet no contemporary historian - not Orosius, not the Gaulish Chronicles, not Constantius of Lyons nor any other contemporary hagiographer, not Procopius, not anyone else - reports anything about it. It cannot be identified with the revolt of Constantine III, which took place in a largely different area - in particular, it seems to have had little or nothing to do with Armorica - and it involves ideas completely alien to him. Zosimus' rebels wanted to establish a separate government, while Constantine wanted no more than to relieve Honorius of his throne; they rejected the laws of Rome, while Constantine went out of his way to mint his own coins in imitation of Roman emperors! Britain and Gaul were the usurper's power base until his death; but Britain and Gaul are also - and at the same time! - the epicentre of a revolt which is not an usurpation but a complete rejection of Roman law and institutions, installing what sounds, from Zosimus' vague statements, a number of more or less independent local governments - "the people of Britain... and all Armorica and other provinces of the Gauls... made themselves independent... setting up their own independent government": of whose rise and eventual fate we hear nothing, though, whatever the case with Britain, we have plenty of evidence about Gaul in this period. Would we not expect to hear something, if such a revolutionary notion as rejecting Roman laws had prevailed through much of Gaul about 410?

Nor, in spite of E.A.Thompson's valiant but doomed attempt, can it be identified with the bacauda taking place in Armorica seven years later, in 417; not only is that much too late, but no British connection can be found. Thompson also makes much of the fact that one Gaulish Chronicle says that a later Bacauda, probably to be dated to 437, "broke off from Roman fellowship"; a statement directly contradicted by the fact that its leader Tibatto immediately approached Saint Germanus of Auxerre to open negotiations with Aetius and the Emperor for a pardon - ooh, some great anti-Roman rebel this is! Germanus died in the course of the negotiations, which may have doomed Tibatto's cause; the next we hear, he has taken refuge among the Huns, which is probably the reason why a later and unsympathetic chronicler made him a breakaway from "Roman fellowship" in the first place. But there is no evidence whatever that his or his followers’ original intentions were anti-Roman in terms of setting themselves up as a separate national identity^[6].

Conversely, while no contemporary account supports any feature of Zosimus 6.5.3, every contemporary one that mentions Britain at all supports 6.10.2, however vaguely. The Gaulish Chronicle of 452 and the Narratio de imperatoribus... agree that Britain was lost to Rome in 410; but - contrary to 6.5.3 - Prosper regards the country as still essentially Roman in 434 or thereabouts. This supports my reading of the Rescript, that is that it was not an imperial reaction to a revolt against Roman law and government - which had not happened - but a rebuff aimed at an essentially Roman province. It is only a century later that we find Procopius regarding British ambassadors as barbarians.

In fact, every aspect of Zosimus 6.5.3 suggest that it is an intrusive item, unrelated to any other known event. It weighs like a brick in every telling of the events of 406-410 - themselves sufficiently dramatic and crowded - forcing unnecessary complications, brain-twisting hypotheses and general confusion. Generations of historians have been pushed to the aspirin cabinet trying to reconcile 6.10.2, where the independence of Britain is the result of Honorius' sovereign, Roman enactment, with 6.5.3, which declares Britain's complete separation, not only from Ravenna, but from the ius Romanum^[7].

On the other hand, 6.5.3 looks like it might have something to do with A. In A, the cause of the break between Britain and Rome is the iugum iuris - the very thing that Zosimus’ rebellious Britons and Armoricans reject. This suggests that we might want to look to Britain for its origin.

Nothing, however, could be further from A's picture than the notion of a Britain already abandoned by Rome, beset by German invaders, valiantly taking up arms against them, and then consciously rejecting Roman ways and resorting to native law. A may be written in the shadow of Saxon power, but its great enemy are the Picts and Scots, and they alone. Also, while A does imply a rejection of Roman political power, the break is not through any reversal to "British" ways, but in the replacement of the earthly lordship of the no-longer-willing Romans with the heavenly lordship of the Christian God. And this lordship involves at least a residue of Roman power, through the Roman possession of the sacred substances oil and wine - in other words, it implies religious dependency on Rome, Roman Catholicism. A does implicitly argue against ius Romanum, at least in the political sphere, but has no clear positive political alternative, and its Great Enemy belongs to an earlier stage of island history, before the Saxons became a major power there.

I have already argued that A is a historical/legendary production, looking back to a partly imagined, partly historical past, and that the Rescript of Honorius and the defeat of the Third Pictish Invasion are not only the climactic, culminating moments, but also the point in which it hits the ground of clearly described, reliable history. In other words, it was written long after these events took place - long enough, indeed, to have forgotten the permanent Roman presence on the Wall. On the other hand, Zosimus' notice does not seem to have the clarity, definiteness and rounded outline of legend. There are fuzzy elements - in particular, the vague mention of Armorica and other areas of Gaul - and on the other hand, an attention to societal features that does not often turn up in legend.

The most interesting of these is his distinction between common or garden "dwellers in the island of Britain" and people who "lived among Celtic tribes". Zosimus being a Roman, his notion of common or garden normality must mean Roman ways; while "living among Celtic tribes" must mean living in a different ethnic world. The description of individuals "living among" the tribes of the Celts and who are politically active - since it seems that it is they who decide to reject Roman ways - reminds me of the Roman-originated dynasties settled after 367 among the Celtic tribes of the north, suggesting that it was their descendants who were the motive power in this political development. Greek, like Latin, allows remarkable precision and clarity of expression, and even a muddle-head like Zosimus would not have described - in prose, as opposed to poetry - native Celts as living "among" the tribes of the Celts, as if they were guests in their own country.

Zosimus is saying that two groups, the ordinary Romano-British of the south, and the Roman-originated tribal dynasties of the north, made common cause to reject Roman laws and Roman ways and to fight “the barbarians from beyond the Rhine”. In Britain there are people who are simply described as “living in Britain” - “living” with no connotations, “living” in the way that Zosimus thinks ordinary”; and people who "live among the Celtic tribes". In other words, the information received by Zosimus included a clear ethnic distinction between "the men of the British island" in general, and some who dwell among tribal groups. He is speaking of a Romano-British nation some of whom live among Celtic eqnwn, "peoples", clearly described tribes; in other words, those Roman dynasties - still clearly understood to be of Roman blood by the time A was written - who "lived among" the tribes over which their fathers or grandfathers or ancestors had been imposed by earlier Roman power.

The barbarians are another group of enormous interest. Nobody seems to have noticed that Zosimus' phrasing implies that the ciuitates of Britain are already under the control of "the barbarians beyond the Rhine" when the war starts. The barbarians had seized everything, and the men of Britain had to free their ciuitates from their threatening presence. In effect, what he is describing is a war against a Saxon occupier in the course of which the British reject Roman ways and laws. It is only when the struggle against the barbarians spreads to Gaul that we hear of Roman magistrates - as opposed to that abstract entity, Roman arche, "principles" or "government" - being expelled; and as this marks a clear chronological and territorial stage of the struggle, it must follow that "those who dwelled among the nations of the Celts" and who took part in the beginning of the anti-Roman movement, must have "lived among" British Celtic tribes. It is in Britain that the movement starts; only afterwards does it spread across the Channel.

Whenever this political development took place, therefore, it must have been after the aduentus Saxonum, and therefore long after 410. If it implied, as Zosimus said, a definite and conscious rejection of Roman ways and Roman rulers, then it is possible that A, the legendary history, represented a charter legend for this nativist revolution, supplying an illustrious precedent and a religious justification. I think we can go even further: if the Picts in A, a legend of the historical past, are meant to figure as something as a counterpart or foreshadowing of the Saxons, and if the latter were the contemporary Great Enemy, then the author of A could not fail to realize, even as he celebrated the great victory over the Third Invasion, that the Picts were not in fact destroyed, but rather remained in their Northern home, regained their independence, started threatening Britain again; caused the Saxons to be sent for; and for all we know, were still a live threat as the author of A was writing. In other words, if he saw the Picts as predecessors of the Saxons, then he must have been writing from a point of view in which the Saxons had been in Britain for a long time already, had already suffered at least one disastrous defeat, but were still a threat; which, to say the least, pushes it even farther from 410AD.

Now, I have a definite theory about this narrative; but I cannot introduce it without a break in the pace of explanation. I must leave Zosimus altogether, and start again from a peculiar expression in Nennius, 300 years and half a continent away.

Nennius has a schoolboy habit of making indiscriminate use of Latin expressions that strike him as unusual; for instance, cyulae for warships. Gildas uses this word only once, in his account of the Saxons; it is in fact an English word, modern keels, and stands only for the warships of the barbarians themselves, easily distinguished, by their design and apparel, from Roman craft. Nennius has found it in Gildas, failed to see its specific meaning, and injudiciously started to use it as standard for men'o'war in the most inappropriate contexts, such as the emperor Claudius' naval raid upon the Orkneys.

Now in his chapter 20, which conflates Welsh legend and Bede's historical account of Caesar, Nennius speaks of the magnum discrimen that the stakes planted by the British in the Thames were for the invading Roman troops of Caesar. The same bizarre expression turns up in Gildas: when striking their Devil's deal with the British, the Saxons - who are, in Gildas' view, shortly to betray their masters - say that they were magna... discrimina pro bonis hostibus subituris, there to undergo great dangers, magna discrimina, for their good hosts.

Magnum discrimen is a rather forced, not to say ungrammatical, way to say "great danger". The standard word for danger is periculum; what is more, a discrimen cannot properly be magnum, since it stands for a "distinction, separation" (the way in which it comes to mean "danger" is the expression in discriminem uitae, "to the distinction of life [and death]"), an abstract concept that cannot carry an idea of size. The expression stands out like a sore thumb to anyone with even a moderate acquaintance with classical Latin; it would have convulsed Cicero and Quintilian, and, by the same token, is hardly in keeping with Gildas' limpid grammar - or vocabulary - or usage.

Finding a strange ungrammatical expression in Nennius would not be particularly surprising; finding the very same terms in Gildas, a writer who would be physically incapable of the solecism of calling magnum a discrimen, suggests a particular purpose. Is it possible that he is quoting? It certainly is, and it must be a sarcastic quotation. The context is savagely ironical, doubly ironical: the Saxons call boni hospites those they are soon to betray and butcher, and the boni hospites themselves are anything but - they are so corrupt, and so stupid, that the vials of God's wrath are already being poured upon them. Sarcasm of this kind is meant to explode on the hearers like a fireball, like mention of Adolf Hitler's good faith and Josef Stalin's humanitarian principle; and that strange expression magna discrimina must be a part of it.

It follows that there was a piece of writing, known to both Gildas and Nennius, that used the expression magnum discrimen or magna discrimina to describe the dangers of war. Both Nennius and Gildas resort to it when describing great wars of invasion that will determine the future of the island: in Nennius, Caesar's Romans are fighting to impose Roman ius on the tyranni et tumidi British leaders; in Gildas, the Saxons will fight to save Britannia from the Picts - though the great writer is foreshadowing, none too subtly, the war of conquest they were soon to wage themselves. That is to say, the expression must be closely connected with the defence of the realm. And remember the sarcasm. Gildas is putting in the barbarians' mouths an expression he knows will resonate in his listeners' ears and minds. He wants his audience to feel disgust. A quotation in this context is bound to be sarcastic, placing noble words, words that are all but sacred, in mouths so vile that even to use them is a pollution, like (in modern terms) a neo-Nazi spouting the Gospel, or the Declaration of Independence, or the Gettysburg Address, or Churchill's Dunkirk speech. And if the quotation has to do with the defence of the realm: to put it in the mouth of treacherous mercenary barbarians, foolishly brought in to defend it because the British did not have the strength, only to prove the worst danger, discrimen, periculum, that Britain ever ran - this would be supremely pointed and effective rhetoric, indeed worthy of this great writer.

We have already seen that Gildas lamented that it has not fallen to him to describe "the dangers of most valiant soldiers in grim war", and that this argues that a piece of writing existed that described exactly that; that the young Gildas, early recognized as a wonder-talent and rightly ambitious, had yearned to measure himself against the classic account of the defence of Britain; and that his life had been blighted by the progressive, disheartening realization of how little his age had of the heroic. This closes the circle; for it can hardly be a coincidence that Gildas opens his own "little work" with what seems precisely a gentle correction of the expression magnum discrimen. Quia non tam fortissimorum militum enuntiare trucis belli pericula mihi statutum est... "As it was not granted to me so much to announce the dangers of bravest soldiers in grim war..." - the proper word is not discrimen, it is periculum: it was not given to me to me to speak of - ahem! - the pericula of bravest soldiers in grim war. After ten years of silence or more, Gildas has finally taken his predecessor's challenge, but according to the literary and artistic requirements of an age of desidiosi, cowardly lazy bums. Gildas was in the mood of Thomas Mann's Adrian Leverkühn, who, faced with the collapse of "the good and noble... what we call the human, although it is good and noble", resolves to "take back the Ninth Symphony", and, measuring himself against the greatest of all his predecessors, writes a Lamentation of Dr.Faustus that is meant to deny, point by point, Beethoven's majestic assertion of the worth of human life. Only Adrian Leverkühn is a fictional character; Saint Gildas was a real person and a real genius, really faced with the collapse of everything he loved.

I suggest that the expression magna discrimina was found in the first few words, both because it is there that it would be most memorable, and because Gildas opens his own work with that correction - and leaves magna discrimina to mendacious barbarians with as much Latin as honesty. I will call this lost document L. It would have opened something like: "my subject is the magna discrimina that the most valiant soldiers of Britain ran in their grim war to free their ciuitates from the barbarians..."^[8]

Now we get back to Zosimus; for it is he, of all people, who shows what seem to me unmistakable echoes of L. In 6.5.3, he all but translates what I surmised as the work's opening words: "Consequently, the people of Britain armed themselves [fortissimorum militum?] and took their lives in their hands [magna discrimina subituris?] in order to rid their cities of the barbarians who were menacing them". Dating Zosimus is a notoriously thankless task, but, as an overt pagan, he must have lived before Justinian's suppression of pagan centres of learning; on the other hand, he says that the British "reverted to their ancestral customs", which the British of 410-440 definitely did not. He must be thinking of the tribalized, Celtic sixth-century Britain of Gildas. The fact that he speaks of poleis acting independently rather than of coordinated action also suggests the fragmented polities of Gildas; and the apparition of Armorica, which fits so badly the picture of 410, would be much better placed close to the sudden apparition of a British-colonized Brittany in the sixth century. Zosimus is flagrantly describing the beginnings of Gildasian Britain.

What is more, his historical scheme and that of Gildas show an exactly similar mistake. Like Gildas, he describes one single barbarian assault, a Saxon assault ("tribes from beyond the Rhine"), which places the very existence of the "British poleis" in danger; I suggest that he has conflated the great raid of 409-411 with the Saxon settlement and rebellion of 442, which is what L was speaking of. Gildas knows nothing of it either: the only war fought towards the end of Roman power is the Third Pictish Invasion, and while Gildas knows of the existence of the Saxon Shore forts, he never shows them being used, as they certainly would have been during the great raid of 409-411.

Now while the raid of 409-411 was large enough to deserve mention in Gallic chronicles, it did not actually lead to any permanent settlement, either in Gaul or in Britain, let alone threaten to conquer the whole island; the revolt of 442 did both and more. And while Zosimus confusing two wholly separate Saxon episodes would be quite in keeping with his notorious ways (E.A.Thompson,^[9] who was defending him, said that he had "an unsurpassable claim to be... the worst of all extant Greek historians of the Roman empire"!), I think we should notice that Gildas knows nothing of it either. Until the Saxons are called in to defend Britain against the Picts, they are never mentioned: it is the Picts who are the great danger, even though, when the Saxons appear on the scene, Gildas starts speaking as though they had been the most dreaded of peoples all along. And if two wholly separate historians make the same mistake, then I submit they had the same source - one which conflated the end of Roman power in Britain, and the Saxon raid which followed it, with the Saxon war, or else so described the beginnings of the Saxon war as to make it seem that it followed upon the end of Roman power in Britain.

This is bound to have something to do with the different sources L is using. L, I believe, is not A, and may contradict it in some aspects. A concentrates on the Picts, and, so far as I can see, ends with the third Pictish invasion and the final British victory, after which Gildas summarizes an unmistakably historical and contemporary account in ch.21; L, like Gildas' own answer to it, is concerned with the Saxons, and may well, indeed must, have contained allusions to their old and terrible reputation. But Zosimus knows nothing of the Picts, and only ever describes a Saxon danger to Britain.

L’s picture of Britain does not seem to agree with A’s, of a consistently non-Roman Britain dominated and protected by a distant and foreign Rome, and eventually losing that protection with the Third Massacre. A, in spite of the probably historical nature of the Third Pictish Invasion, is a legendary picture; L, in spite of the mistake about the early involvement of the Saxons with Britain, must have been primarily historical. A is based on what is, in my view, a purely local history of the over-Wall tribes, and it is not clear whether it itself was aware of the fact; it seems likelier that, by projecting their comparatively recent and certainly local history upon the collective past of all Britain, it left no space for the existence of a Romano-British world. On the other hand, if L was a contemporary account, it must have been clearer than Gildas allows about Britain's mutation from Roman customs to resurgent Celticism, and even about its ethnic roots in the north of the island. Zosimus is quite clear, both on the fact that there were Celtic tribes with Roman residents among them, and that the British had "reverted to their own customs"; and, to judge by the Celtic nature of the British settlement in Brittany, that was a historical fact. However, while Zosimus knew of this development, his probable contemporary Procopius has no idea of it. He regarded the British ambassadors to Justinian as “barbarians”; he knew that Britain had once been Roman; but he had no idea at all of how these two facts were connected. His idea of Britain came half from wholly literary sources and half from tall tales from Frankish ambassadors; and it is the despised Zosimus, to the astonishment of any historian who had the misfortune of dealing with him, who has a sense of the mutation that had been taking place.

Equally important is that, to judge from Zosimus, the changeover was a conscious decision, not a gradual process or drift into Celtic ways. As a result of Saxon oppression, he says, the British decided to make use of local ways, to forsake the unprofitable Roman habits of the past and accept British custom. As I pointed out, this is echoed in the iugum iuris of A, in which it is Roman law that is intolerable. The same idea, though not so clearly stated, dominates Gildas from end to end: his Britons are incapable and unwilling to accept law from the Romans, and that is the cause of all their disasters.

Can we take the Massacre for a historical event? The date, at the very least, must be altered. A, the legend of the Seven Emperors, and (indirectly) The dream of Maxen Gwledig, all insist on a direct and immediate connection with the end of Roman rule in 410; but Zosimus, who only knew L, drew from it a notion that the Massacre was directly connected with the Saxon war; and we know, though he apparently did not, that it started in 441/2. What is more, he only describes it as a result of what, stripped of the talk about the Armoricans and other Gauls "imitating" the British, amounts to a description of large-scale British intervention in Gaul, attended by the expulsion of local Roman magistrates. No such intervention is on record; but in the history of western Gaul in the late fifth and early sixth centuries there are holes you could drive a tank through.^[10].

I take the view that L, as reported by Zosimus, was right, and that any expulsion or massacre of magistrates took place much later, in the course of the Saxon cycle of wars. The fact that every other document places it at the end of Roman power only proves that Welsh legend held it to be so. It must have been a genuine political revolution. Political revolutions have victims, and there is no reason not to think that some surviving supporters of Roman law were got rid of, one way or another, in the course of the establishment of a resurgent Celtic order. Its backdating to 410 is easily explained as a mythical justification for actual political contemporary developments: the Celticizing British party would convince themselves that the end of Roman power in the already distant past actually supplied a precedent and a justification for their political actions in the present. A seems to have been an effort in that direction: when a mythological precedent for a political revolution against the application of Roman law in Britain was required, the memories of one treacherous change of allegiance - the areani's sell-out in the 360s - and three large-scale revolts - those of Magnentius, Maximus, and Constantine - were conflated with the expulsion or massacre of a class of Roman governors, making the latter a political principle.

This would account for the extraordinary potency and persistency of the legend, which has no parallel in Ireland (though Ireland is hardly without invasion legends) but which clung to Welsh myth even to the point of being altered in a self-flattering nationalistic manner in The dream of Maxen Gwledig, centuries after it was formed. It is also relevant to the extraordinary fact that A, who has organized the Massacre of the Magistrates into a structural, almost a cosmic, principle of British-Roman relationship, also makes the Romans leave Britain for the last time before the Rescript of Honorius: the defence of the realm against the Picts, that motor of the whole action of A, is a British, specifically a North British concern. The story of how the British realms consecrated themselves to God before the final victory against the Third Pictish Invasion is probably part of the same legend, a process of self-justification for the victorious Celticizing party: ever since the final victory over the Picts, God, not the Roman emperor, had been the over-king of the British. Some episode of national prayer and repentance - such as we cannot doubt must have taken place, perhaps often, in the course of a terrible struggle for survival - may have been dug up from fading memory, and transformed into a complete rejection of any merely earthly domination, placing the victorious armies of 410 under His hand before they went to avenge their country's wrong on the Picts. That would, no doubt, be a most prestigious precedent.

To our enormous surprise, then, we find that Zosimus (of all people) had some sort of access to a document we can only otherwise trace from the single surviving work of literature from Gildasian Britain. It must have reached him in the keel of one of those Greek ships that regularly plied the tin route to the Cassiterides. We know that books travelled regularly from the East to Britain and Ireland; why not the other way? - except, of course, that books from the islands would not be remotely so interesting to the Greeks as Greek writings to the islanders. But as we surmise from Gildas that L was regarded as a national literary masterpiece, one can imagine a British visitor showing or reading it to Zosimus with a sort of nervous pride - you are a Roman historian, sir, a very educated person... but we, too, have one of our own.

Notes

^[1]Scholars of mythology and popular tradition will be fascinated to notice that the wilder and more complex elaboration upon the idea, the one that is furthest from whatever facts may have taken place and less useful to reconstruct them, is by far the earlier, old enough by Gildas' time for him to revise it, obsolescent by the age of Nennius. And it is marvellously ironic that the current standard edition of Gildas should carry notes by John Morris that identify Gildas' one surviving massacre with - the wrong historical event: to wit, the revolt of Boudicca. I doubt historians do that sort of thing so much any more, but this is still an object lesson to all of us who rush in too eagerly to pluck supposedly historical data from "popular memory", out of context.

^[2]Though John Morris has made a valiant effort to find a reference to it in a contemporary Pelagian letter ascribed, with varying degrees of likelihood, to a British Pelagian bishop called Fastidius, or to Pelagius himself. Age of Arthur, 45: cf. the pertinent criticism of R.P.C.HANSON, Saint Patrick: his origins and career, Oxford 1968, 40-44.

^[3]There are a number of problems connected with the Notitia Dignitatum, to unravel which would require a far greater knowledge of the late Roman Empire than I have. As far as Britain is concerned, scholars who know such things tell us that its list was apparently some decades out of date. To me, this suggests that although Britain was regarded as Roman, there was no direct or constant communication with it at the level of ordinary day-to-day civil service work. I think this agrees with my theory that Britain was run by a Roman but independent government; it may not necessarily have communicated with its “colleagues” in the Western and Eastern Empires - especially if the Western and Eastern administrations doubted or denied its legitimacy, leading to a policy of non-recognition and non-collaboration.

^[4]Zosimus' description of the barbarians "seizing everything in their advance" [panta kat exousian epiontes] seems to echo, consciously or unconsciously, Heraclitus' description of the fire of Doom: "everything, in its advance, will Fire judge and seize" [panta gar to pur epelqon krinei kai kataleyetai] (Diels-Krantz 66). This suggests that, in whatever form the story reached him, the emotional content of the barbarians' advance was little short of apocalyptic; it made him think of, or at least made his sub-conscious echo, a pagan description of the end of the world. (Zosimus was a pagan himself, and likelier to think of Stoic and Heracleitan apocalypses than of St.John the Divine.) It may also have had a content of national moral condemnation of the Salvianus/Gildas kind, since the fire of Herodotus is not only a destroyer, but a terrible judge of the inadequacies of everything it reaches.

^[5]SNYDER op.cit. 33.

^[6]E.A.THOMPSON, Saint Germanus of Auxerre and the end of Roman Britain, Woodbridge 1984.

^[7]It is worth insisting on the strict words of Zosimus 6.10.2. Many historians state that the Rescript was a response to a plea for help from an unspecified party in Britain, but Zosimus does not say so - nor, for that matter, deny it. All he says is that such a document as the Rescript was issued. This does not actually forbid us from thinking of a British appeal for help, but the only source that mentions such an appeal is Gildas - in a part of his story which indubitably depends on A.

^[8]As in the adaptation of Rufinus that forms ch.8, Gildas shows powers of subtle adaptation, rethinking and recasting other men's work. He reworked Rufinus' sentence about the true Sun to make the true Sun shine specifically over frozen Britain, a clever and poetically beautiful use of his predecessor's none too inspired simile. Likewise, when rewriting A's account of the last British legatio to Rome - the sablonibus episode - he put in the legati's mouths his own interpretation of the meaning of "Britain being lost to the Roman name", adding a layer of emotion and a more personal sense of loss to the scene; he has changed nothing, but deepened the story, like the great Greek tragedians - unknown to him - reworked ancient myths. So here, dealing with the Saxon invasion - and with the shadow of his apparently very admired predecessor L - the use he made of the expression "the great dangers of mighty warriors" deepened what we must guess was at the back of it. We have seen that a parallel between Saxon and Roman could probably be found not only in Gildas, but in A: but by putting the comically mistaken magna discrimina in the Saxons' mouths, Gildas both underlines the parallel with another class of foreign defenders of the island, and makes the point that the Saxons really are a bad copy - with poor Latin and worse faith. It is extraordinarily sinister humour, but I am very mistaken if it is not humour - if we are not meant to laugh at the thought of bad Latin pronounced in a thick Saxon accent. This, in turn, asks the question of Gildas' actual attitude to L, one of the classic texts of that culture he was so keen to reform: was he really positive about its Latin style - or the values it embodied?

^[9]E.A.THOMPSON, Zosimus 6.10.2 and the Letters of Honorius, in Classical Quarterly 32 (1982)

^[10]Controversy about the life of St.Genevieve of Paris has shed light on the fact that late-fifth-century history of western and central north Gaul - in other words, of the province of Armorica and surrounding areas - is practically unknown. The historicity of the Vita Genouefae has been attacked on the grounds that it described a sort of prolonged siege, or at least heavy military presence, around Paris, by Clovis' father Childeric, during which the Saint negotiated with him and gained his confidence; but when scholars looked at the records, they found that there actually was no reason why Childeric should not have done exactly what the Vita says he did; and there certainly was a special relationship between Genevieve and Clovis, who had a magnificent church built for her burial and was buried there himself. Vita Genovefae, Bibliotheca Teubneriana 1888; Bibliotheca Sanctorum, Rome 1962, s.v. Genoveffa di Parigi.

History of Britain, 407-597 is copyright © 2002, Fabio P. Barbieri. Used with permission.

Comments to: Fabio P. Barbieri