Roman Exile to Islands

Exile, removal from the centre of things to the periphery, or beyond, was normal part of punishment in antiquity; the Romans didn’t invent it, but they used it a lot, especially in the later empire.

Roman Britain was a place of exile, which is why it may have become such a hotbed of trouble. Valentinian exiled Palladius, chief marshal of the court at Chalcedon (Ammianus XXII.3) to Britain, while ‘Frontinus, an adviser of the said Hymetius, was charged with having drawn up the form of prayer that was made, he was mangled with rods, and having confessed his guilt, was exiled to Britain’ (Amm. XXVIII.1.21). Similarly, Valentinus, a senior military officer, was exiled to Britain and was involved with the barbarian conspiracy of AD365, and in its suppression. The mysterious fifth province, Valentia, taken to be named for Valentinian, might actually have derived its name from Valentinus (Amm. XXVIII. 3.5).

Scilly Islands, off Cornwall, England

Followers of the supposed heretic Priscillian, including two bishops, Instantius and Tiberianus, were exiled to Sylina ‘which is beyond Britain’; believed to be today’s Scilly Islands, which many think were once a single island Scillonia Insula, of which Sylina is a variant. The action taken against Pricillian was by Magnus Maximus, who had him executed (the first major Christian figure to be murdered for heresy; figures like St Ambrose of Milan and St Martin of Tours, opponents of Priscillian, urged him not to be executed). These two prelates were too risky to be sent even to Britain; sending them to the delightful Scillies was almost like sending them to Ireland.

We shouldn’t be surprised. Being exiled to an island (deportatio in insulam) was a common Roman punishment; the future emperor Tiberius was sent to Rhodes in 6BC and spent right years there, unable to return, eventually coming back as a private citizen. Agrippa Postumus, an heir to Augustus, was sent to Planasia (Pianosa) and kept in solitary confinement for eight years until his death. Cornelius Laco, deputy emperor and head of the Praetorian Guards under Galba, was taken to an unnamed island and then killed by Otho. Apuleius in The Golden Ass tells the story of an imperial official sent to Zakynthos after losing the emperor’s favour (fictively Antoninus Pius). While this is fiction, there is nothing to suggest this was unusual.

Julia Caesaris 'the Elder'

This is distinguished from  relegatio in insulam, mainly used for women. The Italian island of Pandateria (modern Ventotene) was a frequent base for exiling imperial women. Augustus exiled his daughter Julia Caesaris there, along with his ex-wife, her mother Scribonia. Tiberius exiled Agrippina the Elder there, where she died. Gaius brought her body back to Rome, but exiled his own sister, Julia Livilla there in turn. Nero’s first wife, Claudia Octavia was exiled there in AD62, then killed. Finally Flavia Domitilla, a relative of Vespasian, was exiled to a neighbouring island by Domitian. She may have converted to either Judaism, or to an early form of Christianity.

Ventotene Island, anciently Pandateria

 Under deportatio, one lost all one’s goods and property and forfeited Roman citizenship.  By contrast, relegatio did not involve such losses, perhaps because a married woman’s property was at her husband’s disposal anyway.

Those who were to be watched were sent to Italian islands. Tiberius went to Rhodes before he could be exiled. Sending someone to Britain suggests they were not watched; one wonders why they were not simply killed. There seems to have been a tradition of sending troublemakers to Britain. It has been suggested that the satirist Juvenal was sent there for some extended period by Trajan in the second century AD.

Exile of political rivals was established punishment in the fourth century. Valens sent Phronimius, a former commander of Julian’s armies, to exile in the Crimea (Chersonesus) for having backed the usurpation of Procopius in 371. This is odd. The Crimea was a Gothic stronghold at that time, so sending someone who knew the Goths quite well into exile in Gothland reads more like an undercover mission. Two other relatives by marriage of Constantius, Eusebius and Hypatius, brothers, were exiled by Valens, but soon recalled and restored to favour (Amm. XXIX.2.11). Then again, Valens had no sons, so they may have been considered for elevation (rather than the fanatic Theodosius – how different might the end of the Roman Empire have been without him?)

Nor was Valens unusual: following the defeat of the British usurper Magnentius, Constantius II, while at Arles, ‘among other atrocities … tortured Gerontius, a count of the party of Magnentius, and visited him with the sorrow of exile’ (Amm. XIV.5.1). Constantius also enabled his urban prefect Leontius to exile to unspecified islands (Amm. XV.7.2) anyone who stood up to him in what appears at first to be a non-political matter: the arrest of a charioteer Philoromus. This may have been a disguised political cell, since there had recently been an attempted coup by the Frankish leader Silvanus; the name of the charioteer – ‘I love Rome’ – may be a secret phrase used by conspirators, a bit like shouting ‘Verdi’ in the Risorgimento of 19^th century Italy. Valentinian also exiled several senators who were said to be conspiring with Auchenius, another charioteer; some others were also tried, but acquitted (Amm. XXVIII.1.27). Julian did much the same, exiling Florentius, chief marshal of the imperial court, to an island called Boae off the coast of Dalmatia (Amm. XXII.3.6).

Banishment seems to have become normalised in the Dominate after AD284 as discussed in depth in Washburn D.A. (2012) Banishment in the Later Roman Empire 284-476 Routledge, especially p.136 . We should consider whether Britain was unusual in receiving exiles. It was an island, with the benefits of movement control. It was distant and it would be very easy for an emperor to have enemies bumped off quietly. It seems the emperor wanted political and religious troublemakers moved from the core of the empire to its periphery, but not outside it.

Given the powers of the emperor, why were these enemies not simply killed then? Since both pagan and Christian emperors, and among Christians both Catholics and Arians, exiled opponents, including those who were guilty of attempting to overthrow the state, we can’t say it was because of Christian scruples.

The difference between the early imperial processes of deportatio and relegatio, which were confined to family rivalries, including disputes over heirs and potential illegitimacies and the use of islands near Elba (these same islands had Bourbon prisons, restocked by Mussolini), the later exiles to islands were primarily aristocrats involved in coups and conspiracies. By the fourth century the old division of patrician and plebeian orders had largely disappeared, to be replaced by those of honestiores and humiliores. Honestiores couldn’t be killed without overriding reasons. It was likely that the Roman upper order would take revenge. The emperor needed them on his side.

Life in the upper orders depended very strongly on the social network individuals could command. Patrons offered endorsement to their clients. Membership of colleges of priests depended very much on people cooperating and collaborating. Exile places the individual outside the network. To be exiled was social murder.

Robben Island, where Nelson Mandela was exiled

 If we wanted a modern equivalent, the placement of South African revolutionaries like Nelson Mandela on Robben Island, intended to be permanent would be comparable. Napoleon was left on Elba and then St Helena rather than face execution in the afterglow of the Enlightenment. It need not be an island. In the 1930s, the Turinese writer Carlo Levi was sent as punishment for anti-Fascist agitation to the Italian Mezzogiorno, where he wrote his great book Christ Stopped At Eboli.

Certainly the later empire was no kinder than the earlier configuration. If the idea was to store errant officials who might have powerful connections, literally isolating them, rendering them powerless, it didn’t work. While the empire was united, exiles could be sent anywhere, and Westerners could be sent east; Bishop Lucifer of Cagliari was sent by Constantius II to Marash in south eastern Anatolia and from there to Egyptian Thebes. Constantius also exiled Bishop Athanasius of Alexandria to Trier in Gaul, and in total Athanasius was exiled five times, although it was increasingly notional – the fifth exile ordered by Valens was from Alexandria to the suburbs of the same city, which strikes me as simply a forced retirement.

The creation of twin empires in AD395 limited this, of course. An emperor could not simply exile a troublesome political or politico-religious figure to an area outside his authority. There are some instances of indulgentia, under which exiles were allowed to return and jurisprudence about whether this meant simply they were to come home, or whether property seized under deportatio and, perhaps most importantly, prior status (dignitas) were to be restored or compensated for. We can imagine that people who had legitimately bought an exile’s farming estate would resist it being given back to him. Christian emperors did use the celebration of Easter to free criminals and exiles (shades of Barabbas!), which was a way to exteriorise this: it wasn’t me who freed him, it was God.

We should note that the Romans did not have a policy of imprisonment as a punishment. If found guilty of a major crime, the alternatives were a fine, the mines, beheading, the arena or crucifixion. Prisons were generally lockups, to hold people prior to trials or pending transfer or execution. Foucault considered it a significant rise in humanity that France had moved from torturing convicts to death to holding them captive for many years, watching them perpetually in panopticon prisons (Michel Foucault: Discipline and Punish: The Birth of the Prison (1991) London: Penguin).

It’s hard to see that any Roman punishment was designed to reform criminal behaviour, except for fines.  The level of the fine would have been designed, like the medieval wergild, to be one which required support from the criminal’s family and patron to be paid, imposing a future discipline and severely limiting the resources to achieve much.

The mines were a form of slow death, where the criminal was to be worked to death, and the galleys would have been a similar fate. Forcing people into the arena to face wild beasts only works if the victim is frightened. Efforts to kill Christians failed because the Christians believed they would go to heaven. We might remember the attitudes of ISIS: you don’t love life as much as we love death.

Exile was often a sentence outside the court system. It was not a death sentence, any more than a concentration camp was, but (except for relegatio in insulam, mainly used as I said above for women) it transferred the convicts’ resources to the control of the state. Deportatio could be used to store the victim till a more politically convenient time, when they had been forgotten, so they could be killed when required with impunity.

Whether Foucault was right to see an increase in humanity and the concept of reforming the individual’s attitudes rather than damaging his body as signs of progress, the Romans never tried this.

Gildas The Roman? Gildas the Goth?

The Strange Case of Gildas

Gildas is strange, despite being known and used heavily by people discussing the transition of Britain from Roman diocese to a multiple set of political units in Late Antiquity. As Guy Halsall now describes such things, it’s messy.

It is well-known that Gildas is not a British or Roman name (Lapidge, M. and Dumville, D. (1984) (eds) Gildas: New Approaches, Woodbridge: Boydell Press). Various suggestions have been made, such as his real name being Sildag and that Gildas was a nom de guerre. If so, an anagram was a pretty thin disguise. Why would anyone wish to disguise himself? It would be fairly obvious from the level of erudition and from the political position being adopted who the author was.

There is one source of names which has not been taken into account, and that is the Goths. Many Gothic names end in —gildaz. These are usually lost when the names are rendered into English from Latin texts, where they appear as Hermanigildus or Leuvigildus. The original names would have been as Hermanigildaz or Leuvigildaz. The —az ending corresponds to —us in Latin or —os in Greek (or Gaulish). Bede interestingly calls Gildas ‘Gildus’. Ammianus mentions an officer – probably a Goth – sent from Constantinople to Julian to tell him Constantius II had died; the man was called ‘Aligildus’ (Amm. XXII.2.1)  Possibly the —az ending was pronounced something like —us anyway. The reason why modern historians have dropped the —az ending is to make them sound more like other Germans. Gothic, being separated from other Germanic languages had retained the Indo-European declension system, which most Germanic languages had dropped, which is why you don’t get it in English. I discussed this possibility with my PhD supervisor, Professor Guy Halsall, and it eventually filtered through to his book Worlds of Arthur.

But Gildas (Gildaz) does not appear to be a full name in itself. It may be a hypocoristic, that is, a familiar form of the name (like Bert for Albert/Herbert, etc.). Could Gildas have had that name because he was descended from Goths, the ones who had been in slavery, some of whom had returned to the Continental mainland? It will be objected that Gildas didn’t like Saxons; he said that the Council of the Britons had ‘sealed its doom by inviting in among them “like wolves into the sheep-fold”, the fierce and impious Saxons, a race hateful both to God and men, to repel the invasions of the northern nations’ (DEB 23). The Goths, however, didn’t consider themselves to be Germans anyway, and there is no saying that the Goths had to like the Saxons; those based in Gaul had fought Swabians and Vandals, while the Franks had fought mostly on the side of Rome against fellow Germans.

There are two vitae of Gildas, each contradicting the other; Gildas is said by the Monk of Rhuys who wrote the earlier Vita (pre-Conquest; everything after that is contaminated with Arthurian junk) to have been in Gaul during the life of King Childeric (r. 457-481). This places him fair and square in the late fifth century, rather than in the sixth. Gildas himself states that the grandsons of Ambrosius Aurelianus are his contemporaries. Taking the standard generation of 25 years and counting back, it suggests that the floruit of Gildas is only 50 years, give or take, later than Ambrosius Aurelianus. We are talking fifth century, not sixth, and somewhere in the range 457-481. The Vita can be read in English in Two Lives of Gildas (trans. H Williams, Llanerch Press, 1990).

We do however have a very similar name, Gildias. He was a vir spectabilis in Italy under Athalaric and held the role of Count of Syracuse. The king rebuked him at length in AD527 for abusing his role and robbing taxpayers blind (Cassiodorus Variae Epistolae IX.14). Gildias was a Goth. In fact, Athalaric was only eleven years old at the time and the letter was written at the behest of his mother, Amalasuntha, Theodoric’s daughter, who was regent of Italy, by Cassiodorus.

But were there any Goths in Britain anyway? We do in fact have such a reference, but a peculiar one it is. Jordanes’ book Getica is, or so he says, a rewrite of a now lost work by the Roman senator Cassiodorus in praise of the Ostrogothic king Theoderic (r. AD486-524). It has several references to Britain.

The first is a detailed description of its size, position and agriculture, its peoples and what they look like. Why should a book about the Goths include a lengthy description of Britain (571 words in an English translation) (Jordanes Getica II 1-15)?

The second is just as odd:

We read that on their first migration the Goths dwelt in the land of Scythia near Lake Maeotis. On the second migration they went to Moesia, Thrace and Dacia, and after their third they dwelt again in Scythia, above the Sea of Pontus. Nor do we find anywhere in their written records legends which tell of their subjection to slavery in Britain or in some other island, or of their redemption by a certain man at the cost of a single horse. Of course if anyone in our city [Constantinople] says that the Goths had an origin different from that I have related, let him object. For myself, I prefer to believe what I have read, rather than put trust in old wives' tales (Jordanes Getica III 38).

This is an odd comment. Quite clearly there were legends which did say Goths had been slaves in Britain and had been redeemed. Who was the ‘certain man’ and why did a horse feature in it? Could this connect with the stories of Hengest and Horsa?

It was Roman practice to reduce defeated enemies to a subordinate status and to resettle them elsewhere in the empire to work underused and under-taxed land (agri deserti). Such people were termed dediticii, and they were reduced to the level of serfs. JNL (Nowell) Myres long ago in The English Settlements proposed that defeated Germanic warriors were settled in Britain as dediticii. These were former soldiers who had unconditionally surrendered. They were excluded from the Caracalla law of AD212 which made all free persons in the Empire into citizens.

The monk of Rhuys also knew that Gildas had been born at Alaclud, which he equated as Dumbarton. Quite why a place should have not one but two British names is a mystery. Alaclud means ‘town on the Clota’, while Dumbarton just means ‘fort of the Britons’ and therefore is probably a name given by the locals to an outpost rather than an autonym. The identification of Clud or Clut with Clota is reasonable, but the second level identification with the name given by Ptolemy of Alexandria to the Clyde, is not, because there were two rivers named the Clota.

Worse, there is a second Araclud. This is the town of Auckland, now more specifically Bishop Auckland, St Helen’s Auckland and West Auckland in County Durham, within the line of Hadrian’s Wall and all on the River Gaunless, formerly the Clota; like a lot of eastern names, that river was not renamed until the Viking ninth century. Until the see moved to Durham, it was in Auckland, Araclud.

Vinovia (Binchester, Co.Durham) Roman Fort

 This makes a lot more sense than to suggest that Gildas came from a settlement a hundred miles into independent British territory. Bishop Auckland is on Dere Street, a Roman road, and is a mile away from the Roman fort at Vinovia, now Binchester, to which it may have originally been a vicus. It has been suggested that a number of Frisian units the (cuneus Frisorum Vinoviensium) were based there. There are altars at Binchester to Germanic mother goddesses the Matres Ollototae (RIB 1030, 1031, 1032) and similar at nearby locations, such as the Matres Germaniae on Hadrian’s Wall (RIB 2064), and many of the dedications involve Germanic peoples like the Tungi and Vangiones.

Principal buildings at Vinovia

 Bishop Auckland’s High Street forms part of Dere Street, suggesting there was a settlement there from Roman times. Escomb Church near the town was built in c.650 from material pillaged from Vinovia. Perhaps its association with Gildas was once known (it’s one of only three early Anglo-Saxon churches to have survived), although if he lived in this area, he either worshipped at a lost church within the fort or privately in a local house. 

Escomb Anglo-Saxon Church, Bishop Auckland, Co. Durham

For all we know, it could have been the most northerly villa in Europe mentioned by Guy de la Bedoyère (Roman Villas and the Countryside, 1993, English Heritage, ch.3 no page). The work on the survival of ladder settlements at West Heslerton and at Dorchester on Thames, Oxfordshire suggests that Auckland may have been a similar survival. In 633, the Britons and Saxons fought a major battle nearby on Dere Street at Heavenfield, near Hexham, the earliest episcopal see, recorded in the Annales Cambriae and Bede’s Historia Ecclesiastica.

Bishop Auckland High Street, part of Dere Street, running from York to the Antonine Wall

A surviving rural part of Dere Street

Who Gildas was is quite clear. He was a monk – he praises monks and nobody else. He quite clearly despises kings and judges, with the ringing phrase ‘reges habet britannia, sed tyrannos; iudices habet, sed impios’ (Britain has kings, but tyrants, it has judges, but impious ones; DBE 27). To follow Halsall, if Gildas’s Roman Britain has kings, where did they come from and if it had judges, who appointed them and why are they impious? Did Roman Britain have kings, even though they were not recognised by the imperial system? France has been a republic since 1870, but it still has claimants to both the Bourbon and Bonapartist thrones. It is far from impossible that there were local holders of thrones throughout the Roman period, who might have remained major local landowners and decurions. Some might even have been proper Romans.

Gildas mentions the Romans leaving military handbooks for the Britons, hardly the actions of an expelled overlord. These might include De Rei Militari, by P. Flavius Vegetius Renatus, who wrote sometime after AD383, because he mentions the death of the emperor Gratian in that year. However, Sabin Rosenbaum (Who Was Vegetius?, Academia.Edu, 2015) dates him to the court of Valentinian III in the 450s, which would rule this out. The supposed manuals might include Thucydides’ History of the Peloponnesian War, which the emperor Julian had used to defeat the Franks and German invaders in the 350s. He’d read the book, they hadn’t.

Gildas mentions a famine (DEB 20.2) which followed the departure of the Roman authorities. Inevitably, the removal of the tax incentive, under which landowners had to produce surplus food to sell to pay their taxes in cash, would lead to the disruption of markets. Similar famines are noted in Spain in 410, described as ‘enormous’ (Chronicle of Hydatius) and in Gaul in 411, also termed ‘enormous’ (Gallic Chronicle of 452). The cause in all three cases is connected with disruption to Roman rule. Since the Goths were not yet in Gaul or Spain in 410 to disrupt the harvest, we could relate it to the substantial incursions into Gaul and Spain by the Vandals, Alans and Swabians. Hydatius refers to a plague in Spain in 409, so it is possible that the incursions brought hitherto unknown diseases. At DEB 22.2, Gildas mentions severe plagues in Britain, which followed a period of improved trade; this may have been the time when the delegation of St Germanus of Auxerre visited Britain in the late 420s.

In fact, the late Roman state was showing itself too sclerotic to function: there had been famine and plague in Syria and Cicilia in AD333 (Chronicle of Jerome), a series of earthquakes in what is now Turkey in 341, 344 (Neocaesarea in Pontus), 358 (Nicomedia),368 (Nicea), a great famine in Phrygia in 370, and failure of the water supply in Constantinople in 373 (all Jerome). The Gallic Chronicle also mentions an earthquake at Utica in north Africa in 408. It is possible that people moving from Africa to Spain brought plague with them, rather than the incoming armies. The imperial infrastructure was being overwhelmed. Governmentality, the idea that the government has the responsibility to fix everything, was proving impossible to maintain.

Vetus Latin Bible - a page from St John's Gospel

 The text of the Bible Gildas quotes from extensively in De Excidio Britanniae is the Old Latin Septuagint (Vetus Latina), the version used before St Jerome’s Vulgate, the entirely new translation from Hebrew and Greek dating to the late fourth century. If Gildas had really been writing in the 540s, as has sometimes been suggested, he would have used at least some of the Vulgate. In those days Bibles weren’t produced as a single work, and each book was copied separately, so it would be a major resource to possess, and not the copy held by an individual monk.

If Gildas was a local lad, then we should be less surprised that Bede had access to Gildas’ book De Excidio Britanniae. There was probably a copy in the great library at Jarrow established by Benedict Biscop two generations before. Bede cites Gildas’ book De Excidio Britanniae (the title variant De Excidio et Conquestu Britanniae is modern; the work does not really discuss conquest and conquestu does not appear in the text; in Latin conquestus in a past participle, not a noun. The Saxons are only mentioned twice in the whole book).

Halsall considers the tyrannus maximus referred to by Gildas to be the usurping emperor Magnus Maximus and not Vortigern, which would certainly place Gildas in the fifth century. It would mean abandoning the letter Gildas quotes, the one to Aegidius, usually assumed to be Aetius. It would not explain many other aspects of the conundrum either.

Professor Guy Halsall, University of York

 Let’s work with Halsall’s dates for a while. If the tyrannus maximus was Maximus, his death came in AD388. The leading figure in Britain then was Ambrosius Aurelianus. Gildas calls him a dux, certainly suggesting he had a Roman role. His name invokes the Metropolitan bishop of Milan, St Ambrose, who was strongly opposed to Maximus; his full name was Aurelius Ambrosius. St Ambrose was born in Trier in AD340. This may explain the cryptic comment of Gildas that the parentes of Ambrosius Aurelianus had been killed there. This has been translated as ‘parents’ in the modern British sense, but the modern meaning in French is ‘relatives’, much closer to the Latin. Those relatives need not be older than Ambrosius Aurelianus. Gildas uses the past tense, so he is referring to a completed event, but one which might be later than those he discusses here.

However the event referred to by Gildas as tantae tempestatis collisione occisis in eadem parentibus purpura nimirumindutis superfuerat, could be a ‘storm’ in the sense that the combined armies of Theodosius and Valentinian II attacked Trier, the capital of Maximus, with the Frankish leaders Richomeres and Arbogastes, and the Romans Promotus and Timasius, associates of Theodosius.

Moreover the term used for descendants of Ambrosius in Latin is suboles, which does not mean ‘grandchildren’ as is often claimed but simply ‘offspring’, which could equally mean sons and daughters or adopted children.

Halsall leaves us with the possibility that we are looking at a badly reported Roman era event, not a medieval one, or at least a late antique one. There is nothing to stop the ‘Council of the Britons’ being established before the Council of the Gauls, which was set up in Arles after the invasions of the early fifth century. It would even make more sense if such a council had been set up after the ‘Barbarian Conspiracy’ of 367 or even after the recovery of Britain after the revolt of Magnentius (d.AD353), to address grievances such as those against ‘Paul the Chain’, executed in 361.

There is no way to place Gildas firmly in the sixth century and plenty of evidence for the fifth. The sixth century placement arises from a belief that Maglocunus was Maelgwyn of Gwyneth, who alleged died in the Justinianic Plague of the 530s-540s. Dating one dodgy document by reference to another, the Annales Cambriae, a ninth century work, which Ken Dark dismisses out of hand (Britain and the End of the Roman Empire,(2000) Tempus: Stroud) seems unsustainable. A fifth century Gildas, Goth or otherwise, seems more defensible.