The Three Romances

Three later Arthurian tales, all of which were most probably composed in thirteenth-century Wales, are to be found in the White Book of Rhydderch and the Red Book of Hergest alongside the other Mabinogion texts. For the compilers of these fourteenth-century codices, it would seem the tales belonged to the same narrative-historical tradition that produced the Four Branches, Culhwch ac Olwen and the other so-called ‘independent native tales’. Modern scholarship, on the other hand, has distinguished them from the latter on the basis of their association with the famous Arthurian works of the Old French poet Chrétien de Troyes. Historia Peredur ab Efrawg – the most complex and problematic of these texts – shares some significant material in common with Chrétien’s Perceval or Le Conte du Graal. The Welsh Romance of Owain, subtitled Chwedl Iarlles Y Fynawn, is more clearly related to Ywain, Le Chevalier du Lion and it various European counterparts. Finally, we have the Chwedl Geraint ab Erbin, which follows the plot-line of Chrétien’s Erec et Enid in almost every detail. It is because of this undeniable relationship with the Continental Arthurian texts that these chwedlau are referred to as ‘The Three Romances’ of the Mabinogion collection, the term by which they are most commonly known.i

Their resemblance to these influential twelfth-century French analogues has not been entirely advantageous to the critical appraisal of the tair rhamant, to give the Three Romances their (modern) Welsh appellation. Jones and Jones, translators of the 1949 Everyman edition of the Mabinogion, damned them with the faintest of praise:

It is probable that their charm seems less to the Welsh than to the English reader; for the former the decline from the Four Branches to Geraint son of Erbin would not be unfairly expressed in the difference between Branwen and Enid, or Manawydan and Gwalchmai. In the romances lay figures move through stock adventures in unlocalized places; they entertain but have little power to move or excite. The sorrows of Branwen touch one as close after a hundred readings, but one needs sentimentality rather than sympathy to feel pity for Enid. ii

Whether or not this is a fair assessment of the literary merits of the Romances, it neatly expresses the unconscious starting point of a number of their more recent readers. The tendency to regard the Romances as translations or adaptations rather than authentic works of native imagination has led to an inevitable down-grading of their value either as source-materials of the ‘Celtic’ tradition or independent works of literature in their own right. Even now, the rhamantau tend to be studied in terms of reception theory – as a ‘minor literature’ responding to the dominant Franco-Norman cultural hegemony represented by Chrétien’s Romances.iii All too often, the value or even the success of these tales has been gauged primarily in relation or resemblance to their French predecessors.

While Chrétien’s works almost certainly influenced the structure and ethos of the Welsh Romances, the translation model has too often led to an oversimplification of a complex reality which downplays both the formative significance of Chrétien’s own Brythonic sources as well as the literary autonomy of the rhamantau themselves. The influence of Chrétien’s tales in the increasingly cosmopolitan world of the Age of the Llywelyns does not make the rhamantau any less interesting – nor indeed any less ‘Welsh’ – than the Four Branches and the other native tales. However, the question of origins and influences – the so-called mabinogionfrage – remains an unavoidable issue in the assessment and interpretation of these Welsh Romances. Here I will attempt a reconsideration of the evidence, out of which I hope will come a conclusive resolution to the problem.

To do this we need to clarify the broader picture of the medieval Arthurian topos and its evolution: the starting point for which is the ‘native’ Arthurian materials found in the poetry, hagiography and pseudo-historical writings of pre-Norman Wales. The role of Geoffrey of Monmouth and the Caerleon school also needs to be understood, as do circumstances surrounding the spread of ‘the Arthurian meme’ throughout Franco-Norman Europe in the middle decades of the twelfth century. This brings us to the work of the great French Romancer, Master Chrétien himself, whose oeuvre played such a formative role in the development of the Arthurian tradition in Wales and on the Continent alike. An intriguing picture of mutual influence emerges when it comes to the relationship between this Continental tradition and its native Welsh counterpart. The particular way in which this synergy plays out in each of the Welsh Romances is something we will be exploring on a case-by-case basis as we come to examine the texts themselves. While it should be understood that the texts have somewhat different manuscript histories and (we might suspect) different circumstances of authorship, they do nonetheless spring from a common set of formative principles. Through further examination both of the Welsh evidence, and a literary-critical analysis of the narrative dynamics of medieval Romance, it has been possible to shed some light on the pedigree and significance of the rhamantau both in general terms, as well as gaining a better understanding of the specific circumstances of each of the three tales.

Key to the resolution of this problem has been the identification of a distinctively medieval narratological form – what we have referred to as the Magical Plot. A summary of this idea – in which the theories of literary critic Dr. Anne Wilson have been a key influence – will need to be briefly explained, but must at this stage be taken as read. The hypothesis of the Magical Plot – and what it tells us about the position of the Welsh tales in relation to the wider field of romance – will receive ample verification through the empirical material of the texts themselves. The aim of this introduction is to give some sense of our final destination, but also to familiarise the reader with some of the analytical tools that we will be bringing to bear along the way.

The Historical Arthur and the Pre-Norman Tradition

The legend of Arthur seems to have enjoyed a vigorous existence in Wales during the pre-Norman period, leaving a distinctive spoor in the vernacular and Latin literature of the ninth, tenth and eleventh centuries, prior to any influence of European Romance. We know virtually nothing about the historical Arthur – who may or may not have been a warlord active in organising the defence of the sub-Roman territories during the Brythonic resurgence of the early sixth century. It was in this way that he was remembered in one of the earliest surviving Arthurian references, found in chapter 56 of the ninth-century Historia Brittonum. Here Arthur is described not as a king or an emperor but as a specifically military figure – the dux bellorum – who led the British to victory at the Mount Baden (c.500 AD) as well as a number of other battles along the frontier with the Anglo-Saxon territories in the southern and eastern areas of the island.iv

The Historia Brittonum is composite document, with a tortuous manuscript history which as yet remains only partially understood. The ‘historical’ section, in which the battles of Arthur are described, seemed to have been shaped in an ecclesiastical context during the early ninth century, with a broadly hagiographical motivation (Arthur himself is an incidental figure in a moralising narrative in which the holy men Germanus, Patrick and Rhun ap Urien represent the nodal points in the destiny of the British peoples). However, this is appended by a list of ‘wonders’ (mirablia) which seem to have a more folkish provenance, and may have been a subsequent accretion. These list a number of mysterious or supernatural features in the landscape of Wales (the concentration of these ‘wonders’ around the south-eastern borders gives an indication of the source of this material). Included among these are two items of Arthurian interest: the supposed footprint of Arthur’s dog Cafal in a stone on a cairn, and the grave of Arthur’s son Amhar. The first of these is of particular interest, as it mentions the hunt of great boar Twrch Trwyth – which is described in vivid detail in the burlesque early twelfth-century tale Culhwch ac Olwen, described in more detail elsewhere on these pages.

So whatever the reality or otherwise of the quasi-historical accounts of Artorius the dux bellorum, there is a strong suggestion that by the end of the eleventh century a thriving folk tradition concerning Arthur and his exploits was endemic throughout the Brythonic world. The Arthur of this pre-Norman tradition was a somewhat rougher and more primitive character than he would become in the court literature of the High Medieval period. This is the Arthur we find in the early Mabinogion tale Culhwch ac Olwen: a Dark Age warlord rather than a king, surrounded by a rather motley host of warrior heroes and freakish prodigies. In other respects, this body of pre-Norman Arthurian lore has been compared with the Gaelic tradition of Finn Mac Cumail and his heroes.v As with the Fenian tales, magic and the supernatural play an important role – Arthur and his men contend with witches, giants or monstrous animals as often as they do with rival warriors or robber-chiefs. They are often involved with cattle raids, the freeing of hostages and the seizure of magical treasures. We find traces of this tradition not only in Culhwch ac Olwen but also in the early poetry of Wales (texts such as Pa Gur in the Black Book of Carmarthern) as well as in the earliest stratum of Triadic literature. This same Arthurian lore also forms an important backdrop to Welsh and Breton hagiography, in which the great British chieftain often plays a cameo role in the spiritual careers of fifth- and sixth-century holy men such as Petroc, Gildas and Cadoc. It is in this religious literature that we find some of the earliest and most unreconstructed aspects of the Arthurian tradition – Arthur portrayed as a would-be rapist in one episode, and as a bully or buffoon in many others.vi While his role in these stories is that of the secular foil to the spiritual hero, one can’t help feeling that this aspect of the Arthurian persona rings true with the sub-Roman realities which the tradition dimly recalls.

One of these Early British hagiographies, the Life of Gildas by Caradog of Llancarfan, includes what must be considered the earliest depiction of a theme which was to become a staple of the later Romance tradition – the abduction of Arthur’s queen, and her subsequent imprisonment in the faerie otherworld. Here, we are taken right back into the primitive underbelly of the Arthurian complex, a culture in which relations with the faerie otherworld were depicted in what might be described as inter-tribal terms. There is also a mythological dimension to this scenario: Gwenhwyfar’s abduction may have originally been a precursor to an otherworld conception motif, which seems to have been a traditional feature of ‘birth of the hero’ stories in the pre-Christian Celtic narrative culture (judging by what we can reconstruct of the latter from medieval sources). This episode serves as a useful reminder that the status of Christianity in sub-Roman Britain was perhaps rather more tenuous than it sometimes may appear in hindsight. A key role of the early Arthur seems to have been as a kind of quasi-judicial mediator between the diverse tribal, religious and cultural systems of the Island. One thinks, for example, of passages like this from Culhwch ac Olwen:

A short while before this Creiddylad daughter of Llud Silver-hand went with Gwythyr son of Greidawl; and before he had slept with her there came Gwyn son of Nudd and carried her off by force. Gwythyr son of Greidawl gathered a host, and he came to fight with Gwyn son of Nudd. And Gwyn prevailed, and he took prisoner Greid son of Eri, Glinneu son of Taran, and Gwrgwst Half-naked and Dynarth his son. And he took prisoner Pen son of Nethawg, and Mwython, and Cyledyr the Wild and his son, and he slew Nwython and took out his heart, and compelled Cyledyr to eat his father’s heart; and because of this Clyedyr went mad. Arthur heard tell of this, and he came into the North and summoned to him Gwyn son of Nudd and set free his noblemen from his prison, and peace was made between Gwyn son of Nudd and Gwythyr son of Greidawl. This is the peace that was made: the maiden should remain in her father’s house, unmolested by either side, and there should be a battle between Gwyn and Gwythyr each May-calends for ever and ever, from that until doomsday; and the one of them that should be victor on doomsday, let him have the maiden.vii

Frequently in the later romances (as well as the pre-Norman material) we find examples of Arthur and his men seeming to intervene and arbitrate within a strange and savage world inhabited by primitive figures similar to Gwyn ap Nudd or Gwrgwst Half-Naked. A common motif is the cessation (or sometimes the establishment) of certain kinds of ‘custom’ or ‘game’ practiced within (or inflicted upon) these backwoods communities, which often hint at various forms of pre-Christian ritual. A green knight, or a black giant or a similar figure often plays a leading role in the ‘customs’ into which the questing hero is drawn. This enemy is invariably defeated: and them either killed or forced to pay submission to the court of Arthur. A sense of liberation within the surrounding community is often subsequently expressed, and the ‘custom’ is revealed to have been an oppression to the people. We might think of supernatural enemies such as the ‘dog-heads’ or ‘witches’ described in Old Welsh Arthurian poems such as Pa Gur and the equally archaic Peniarth 16 series of Triads), but also of the Sparrow-Hawk joust and the ‘enchanted game’ of the Mist Hedge Earl in Geraint. A variant of this theme sees a deliverance from a dragon or a monster of some kind, to whom the sons or daughters of the surrounding community must be sacrificed on a regular basis. Both variants are suggestive of a lingering memory of pre-Christian custom and belief – albeit recalled through the hostile lens of medieval hagiography.

So the restless wandering of the romance heroes through the lands of giants and marvels perhaps contains a distant echo of an authentic proto-historic situation, in which Arthur and his men patrolled the often uncertain boundaries between Early British Christendom and the various non-Christian cultures that could still be found across significant areas of the sub-Roman West. Could it be that the state of cultural and social fragmentation which existed in the formative sub-Roman centuries may be responsible for the sense of disorientation often so detected by critics within the landscape of Arthurian Romance? We find in this literature not the carefully delineated countryside of the Four Branches, but rather a succession of ‘trackless wastes’ populated only by occasional outlandish communities, who may be friendly or hostile to the Arthurian court and its representatives. While a loose hegemony exists throughout certain territories of the island, with the kings and dukes of Cornwall, Lothian and Galloway (later joined by those of Anjou and Brittany) often represented as Arthur’s liegemen and allies; there is always somewhere on the Island of Britain – often little more than one day’s ride from Caerleon or Carlisle – where a ‘red knight’ or a ‘black giant’ who can be found ravishing a local community in defiance of the Arthurian writ, prepared to take up arms against the knights of the court. This sense of Britain as a fragmented and largely untamed space, in which Christian civilisation exists in contention with other cultural systems is arguably what constitutes the atmosphere of the romances. This geo-cultural backdrop is consistent with that represented in texts such as Culhwch ac Olwen, Pa Gur and the early triads, and may be ultimately derived from the particular historical situation in which these narratives were originally conceived.

The Once and Future King

As well as a mediator between tribes, and warden of bounds between Christian civilisation and the pagan otherworld, Arthur would come to occupy another important role in the native imagination: one which became increasingly important as the Brythonic peoples on both sides of the channel once again faced the prospect of territorial loss at the hands of a foreign invader. As we have seen, the Welsh had a well-developed prophetic tradition which had as its focus a restoration of their ancient rights of sovereignty over a unified Brythonic kingdom of the ‘Island of Britain’. Key to this fantasy was the idea of the return of an ancestral hero (such as Cunedda, Kynan or Cadwallader) who would unite the tribes of Wales and lead them to victory against the Anglo-Saxon oppressor. By the end of the eleventh century, in the wake of a fresh wave of colonisation – this time at the hands Franco-Norman feudal interests – the millenarian fantasy seems to have been undergoing a popular revival. The role of the prophesied deliverer was now more often than not taken by Arthur himself. The prophet would more often claim the authority (inner or outer) of Myrddion (Merlin) or the more established Welsh archetype of the vatic Taliesin.

It was in this politically-charged context that we first observe Arthur coming to the attention of the Anglo-Norman world. In a text written around 1120, we have an account of a journey that took place in 1113, in which a delegation of clerics from Laon in Normandy make a tour of Britain. While passing from Exeter to Bodmin, the clerics are shown various landscape features of Arthurian interest and were informed that they were now in ‘Arthur’s country’. On arrival at Bodmin, they narrowly avoided a violent confrontation when they failed to show appropriate sensitivity to local Arthurian beliefs:

...a certain man having a withered hand kept a vigil at the shrine [of Our Lady of Laon] to recover his health. In just the same way as the Bretons are accustomed to arguing with the French about King Arthur, the same man began to bicker with one from our community by the name of Hangello of the community of Lord Guidon, Archdeacon of Laon, saying that Arthur still lived. Then there arose a not a small tumult; many men rushed into the church with arms and if the aforementioned cleric Algardus had not prevented it, it would almost certainly have come to the spilling of blood. viii

Around the same time, William of Malmesbury makes dismissive reference in to Brythonic fabula heralding Arthur’s imminent return, but at the same he gives us some nuggets of useful information about the topos as it existed in the first quarter of the twelfth century.ix It might be plausibly suggested that such stories were ‘in the air’ during these early decades. There is no doubt that the Norman incursions into South Wales and Brittany had re-opened a number of old wounds in the Brythonic collective memory during this period. Under these conditions Arthur become rather more than simply the clod-hopping figure of folkish imagination. Instead he became the vanguard of the militant nationalistic fantasy, drawing diverse characters from the furthest reaches of the mythical and tribal-historic past into his spectral army. Here, perhaps, we see the early stages of the legend of the Once and Future King.

Geoffrey of Monmouth and the Historia Regum Britanniae

It may well have been in this form that the Arthurian complex was first encountered by Geoffrey of Monmouth, an Anglo-Norman clergyman of the early twelfth century. As his name would suggest, Geoffrey seems to have been born and raised in the Anglo-Norman enclave of Monmouth, on the south-eastern borders of Wales. This area had been in the hands of Breton vassals since the Norman conquest, and Geoffrey himself may have been of Breton extraction, or at least have been attached in some way to a Breton court community – an affiliation which would go some way to explaining his curious (and, at that time, unusual) interest in Brythonic historical tradition. One of his earliest known works was Prophetiae Merlini ‘The Prophesies of Merlin’ – which was heavily influenced by the Brythonic vaticinatory tradition described above. Geoffrey claims to have translated it from a Welsh original.

This was later incorporated into a much larger work known as Historia Regum Brittaniae, a historical prose-epic, written in medieval church Latin, relaying the fate of the Brythonic peoples from their pseudo-historical Trojan origins to the death of Cadwallader in the seventh century. As with the Prophetiae, Geoffrey made the typically medieval claim that this was a translation from ‘a very ancient book written in the British language’. No such book has ever been identified, and the general view seems to be that it is unlikely to have existed in anything other than a metaphorical sense. Instead, Geoffrey seems to have drawn on the small number of surviving Latin historical sources pertaining to pre- and post-Roman Britain (including the ninth-century Historia Brittonum, histories of Bede and the relevant sections of Gildas’s Excidio Britanniae). These fragments he leavened this with a mixture of traditional Brythonic oral material from Welsh informants – along with judicious doses of his own imagination.

The ratio of these last two components has been a question that has vexed medievalists and literary historians. Most obviously, the influence of the Brythonic tradition is to be found in the nomenclature of his kings and heroes: Cassibellanus, Heli, Dumvallo and Cadualadrus, to name but a few, all of which have clear counterparts in the Welsh tradition. But were these mere ciphers extracted from Welsh genealogy to populate narratives of Geoffrey’s own invention? To this question we can offer a qualified negative. The evidence would suggest that Geoffrey tapped into a number of authentic Brythonic narrative traditions, for which there are clear analogues in the vernacular texts found in the Mabinogion corpus and elsewhere. It has been suggested, for example, that Geoffrey’s account of Brennius and Belinus may well belong to same seam of Brythonic (ultimately Belgic) lore which finds another expression in the Second Branch of the Mabinogi. Native oral tradition also seems to have informed the legend of Maxentius (= Maxen Wledig of the Welsh tradition) that, filtered though a rich blend of Latin sources, has also contributed to Geoffrey’s complex tale of dynastic and geo-political intrigue. However, what holds these fragments together is the decidedly Welsh concept of Ynys Prydain, the unified island kingdom, which was lost due to native folly and the perfidy of the Saxons but which would one day be reclaimed. Geoffrey understood and developed this concept of pan-British sovereignty, and the integration of the Prophesy of Merlin into the very heart of the book reminds us of the importance of the Welsh vaticinatory tradition to both the content and the ideology of his magisterial opus.

Whatever the precise nature of Geoffrey’s attachment to the Breton enclave of Monmouth, on the edges of the southern Welsh marches, these geographic and cultural circumstances alone might explain how he came to be such as significant conduit of Brythonic lore into the Anglo-Norman consciousness. He mentions two significant contacts, Walter the Archdeacon and Caradog of Llancarfen, both of whom may have played a role in stimulating and informing Geoffrey’s interest in the Brythonic tradition. To the former, Geoffrey attributes the gift of the ‘ancient book written in the British language’, and later in the Historia refers to him as an oral source for his knowledge of the battle of ‘Cambglan’ (this would appear to be a genuine Old Welsh form of the more familiar Camlan). To the latter, he attributes the continuation of the story of the Brythonic people post-Cadwallader – leading to a tradition that this Welsh monk was responsible for the inception of the family of chronicles known as the Brut Y Twysogian, our main historical source for the post-Norman Welsh Middle Ages. More certainly, this Caradog is generally believed to have been the author of the Vitae of Saint Cadoc and Saint Gildas – hagiographical works which contain a significant leaven of Arthurian material of the primitive, pre-Norman type discussed in the section above. As well as these two, there may have been any number of other potential contacts and informants available in the Welsh Marches to anyone minded to seek them out.

However, there is no doubt that Geoffrey added much of his own to this native Welsh material, and it is not unfair to suggest that the process is so complete that the end product neither looks, feels, nor sounds like traditional Welsh narrative history. The influence of the Anglo-Norman cleric’s Latin education is a prominent keynote of the Historia. Virgil looms large as a presence in the medieval library, and the influence of the Aenied on the myth of Brutus and his Trojan origins needs no qualification. Ovid was another popular author in Central Middle Ages, and though his reputation was to flourish more in the generation after Geoffrey’s death, the influence of Fasti, Heroides and the Metamorphoses have also been detected in the canon’s narrative vision. Other writers such as Pliny the Elder, Cicero, Suetonius and Tacitus may have only been known to Geoffrey through anthologies and florilegia, but the style and phraseology of these classical writers is not infrequently discernable in his work. Overall, the style and aesthetics of the Historia – with its rhetorical bombast, militarism and appeal to flinty virtues – is essentially Roman rather than Celtic in conception. It might be described as Brythonic material refashioned into the mould of a narrative Latin history in the style of Virgil, Tacitus, Orosius or Livy.

Whatever the precise nature of his Brythonic affiliations and sympathies, Geoffrey was first and foremost a member of the Anglo-Norman establishment. Dedicatees of the Historia include Robert Earl of Gloucester (Henry I’s favoured but illegitimate son), Alexander Bishop of Lincoln and Waleran Count of Mellent. These names testify to Geoffrey’s powerful connections within the Anglo-Norman elite. For all his Welsh sympathies, Geoffrey at no time questioned the legitimacy of the occupation of England or Wales by the francigenae; and it has been plausibly suggested that a primary purpose of the Historia was to lend legitimacy and precedent to the expanding Anglo-Norman hegemony.x Whatever his motivation, Geoffrey’s most enduring contribution to the Arthurian mythos was to remould the image of the Brythonic folk-hero into that of a contemporary High Medieval feudal prince. Whether this was done to flatter the Welsh by endowing their ancestral heroes with a grandeur and majesty equivalent to that of the great Anglo-Norman magnates of the day; or whether it was to indulge the Anglo-Norman desire for a legendary hinterland to equal that of the Kings of France, the end result was the same. Geoffrey’s Arthur was no longer the hall-dwelling chieftain of Culhwch ac Olwen, or even the prophesied hero of the anti-Norman Brythonic resistance. Instead he became the British Charlemagne, a paradigmatic Christian king, whose sumptuous court at Winchester, London or Caerleon was a self-conscious echo of the palace communities of the great Anglo-Norman kings. It was this Arthur that was to prevail as an enduring feature of the European imagination. By the end of the twelfth century, his fame extended from Dublin to Palestine, Sicily to Sweden. It was this Arthur who we encounter in the romances of Chrétien, Béroul and Wolfram von Eschenbach – and of course in the thirteenth-century rhamantau themselves.

The Caerleon School and the Spread of the Arthurian Meme

Geoffrey’s Historia has been described as a ‘medieval best-seller’, and it is evident that its spread throughout the Franco-Norman world following its publication in 1136 was immediate and phenomenal in scale. It is not immediately obvious why the traditions of these Brythonic peoples should have appealed so much to the hard-headed military aristocracy of the francigenae. To the Anglo-Norman royalty the Historia had obvious advantages: representing, as we have seen, an alternative paradigm to the Carolingian mythos cultivated by their Capetian rivals. Geoffrey’s Arthur was also an approachably contemporary figure – a British chief flatteringly repackaged as High Medieval prince. But none of this quite explains why the Arthuriad would resonate so widely and enduringly across Medieval Europe. Some historians have suggested a growing desire on the part of the knightly class to develop traditions of their own – distinct from the classical or biblical canons, which were seen as the province of the rival clerical class. The Arthurian tradition would have been suitable for this purpose, offering as it did a generic historical canvas onto which the ideology of the feudal warrior elite could project its own ideologies. But the Arthurian complex was more than a blank slate. There must have been some quality within the matière de Bretagne – inherent or perceived – that led to its enthusiastic consumption among the Franco-Norman aristocracy, and this cannot quite be accounted for Geoffrey’s Historia alone. One suspects the role of the supernatural, which predominates in the Celtic-derived Matter of Britain (but which had been largely elided in other court genres) may have constituted an important aspect of its appeal.

For whatever reason, it is clear that Geoffrey himself was to quarry deeper into the Celtic sources in the wake of the phenomenal success of the Historia Regum Britanniae. As we have suggested, Geoffrey’s initial opus was a rather dry account, which drew on native Welsh (and possibly Breton) traditions, but presented these in the chronicler style of the classical Latin histories of Horace or Pliny combined with the no-nonsense militarism of the chansons de geste. However, Geoffrey’s later Arthurian work, the Vita Merlini (composed c.1150), was of a rather different timbre. Here, distinctly Celtic themes such as the Madness of the Bird Man dominate not only the subject matter, but also slant the entire psychological preoccupation of the work. The distinctively Welsh tradition of ecstatic prophesy takes centre stage in the Life of Merlin, with the quintessentially nativist figure of Taliesin even putting in a significant appearance. It would appear that over the decade and half since the release of his popular Historia, Geoffrey of Monmouth’s interest in the traditions of the Brythonic world had not merely been sustained but had actually deepened and intensified: with a conscious effort apparently being made to locate and capture the authentic inflexions of the native source material.

It would also seem to be the case that Geoffrey was not alone in his determined excavation of the native Welsh sources. The south-eastern area of the Welsh March remained something of a porous membrane in this process of cultural diffusion. It is possible to conceive of a ‘Caerleon school’ – the first generation of which would have included figures such as Geoffrey of Monmouth, Caradog of Llancarfen and perhaps also the mysterious ‘Bleddri’, about whom more will be said below. As in many other areas of the March, the Anglo-Norman and Native Welsh aristocracies of southeast Wales frequently intermarried, creating a distinctive Cambro-Welsh community in which Welsh and French speakers would have freely intermingled. As Arthurian scholar Constance Bullock-Davies memorably described it:

Reconstruction of daily life in those great households, in which intermarriage between French and Welsh had taken place, implies how natural the exchanging of any languages and tales must have been. It points unmistakably to a way in which Celtic story could have found its way into French… Contrary to what our history textbooks sometimes infer, because they cannot avoid over-simplification when dealing with political history on a national scale, the process of integration between Normans and Welsh and the consequent interweaving of the two cultures began much earlier than we are apt to realise. Cyfarwyddiad, latimers, and French, Welsh and English minstrels lived together in the same castles along the Welsh Marches from the time of the conquest. They could not have failed to impart to one another something of each of their native literatures.xi

It was within just such a milieu that Marie de France is thought to have gathered the native lore on which she based a number of her laïs.xii Writing in the late-twelfth century, the lays in particular betray a strong Celtic influence – analogues to which can be found in the Four Branches of the Mabinogi. Despite the nominal Breton associations of these short narrative sequences, a strong Welsh influence is arguably even more apparent. If scholars are correct in identifying ‘Marie de France’ with Marie the Abbess of Shaftesbury, daughter of William Consul, then we have further evidence of a specific connection with the Caerwent/Caerleon area of South East Wales. Marie de France would represent one of a number of senior figures within the Anglo-Norman communities on the borders of Wales consciously drawing on Brythonic narrative lore as a key ingredient of distinctive form of hybrid cultural produce which seem to have had particular cachet within the wider court circles of the francigenae.

The Welsh storytellers and historians themselves could not have been oblivious to the interest in their works from these prestigious audiences, and there is some sign that the premium attached to its produce in the Franco-Norman world may have stimulated the narrative culture of native South East Wales itself. The Fourth Branch of the Mabinogi makes reference to ‘storytellers from Morgannwg’ – suggesting that this was a district famed for its excellence in narrative performance. It has been noted that the characteristic word order and syntax of prose Middle Welsh seems to have been influenced by the speech-patterns of Gwent and Glamorgan xiii – suggesting that the impetus for recording traditional narration in vernacular literary form may have begun in Wales within this particular context. Whether or not this was the case, texts such as Culhwch ac Olwen themselves prove that a purely native tradition of Arthurian narrative prose was already well underway in South Wales by the beginning of the twelfth century. An early version of Geraint ac Enid might have emerged from this same context (the two texts are closely associated in the late manuscript tradition as we will see). The key point is that there were a number of attested instances of stories about Arthur in South Wales being told (and written) in the early twelfth century; and no shortage of potential informants for interested members of the Anglo-Norman community to consult should they wish to deepen their understanding of the fashionably exotic Brythonic tradition.

One native Welsh authority from the southern marches is worth mentioning at this juncture. A number of continental writers were to identify a certain Breri or Bleheris as the original authority on the twelfth-century Romance tradition and the Grail legend in particular. This figure has been plausibly identified with a Welsh interpreter known as Bleddri ap Kadifor, who received a grant of land in the Carmarthen area around 1130 – precisely the time when we can expect that this Cambro-Norman tradition of Arthurian Romance might have been undergoing its prototypical phase of literary development. It would have been through intermediaries such as Bleddri ap Kadifor and, not to mention their influential sponsorsxiv, that the fashionable Matière de Bretagne became a popular feature of courtly entertainment in the French-speaking world. Geoffrey might have been an important figure in this process of cultural transmission, but he was by no means the only point of contact. This will need to be borne in mind when we come to consider the Brythonic influences that in some way seem to have influenced the influential romances of Chrétien de Troyes.

Chrétien de Troyes: Context and Sources

What place is there within the bounds of the empire of Christendom to which the winged praise of Arthur the Briton has not extended? Who is there, I ask, who does not speak of Arthur the Briton, since he is but little less known to the peoples of Asia than to the Bretons, as we are informed by our palmers who return from the countries of the East? The Eastern peoples speak of him as do the Western, though separated by the breadth of the whole earth. Egypt speaks of him, and the Bospherous is not silent. Rome, queen of cities, sings of his deeds, and his wars are not unknown to her former rival Carthage. Antioch, Armenia and Palestine celebrate his feats.

These words, from an anonymous French scribe in the 1170s, illustrate the extent to which ‘the Arthurian meme’ had spread throughout the Christian world by the end of the third quarter of the twelfth century. This was the age when the Duchy of Anjou and its Atlantic territories (including the Kingdom of England) represented the military, economic and cultural centre of gravity of French-speaking European civilisation. Under Angevin leadership, the Arthurian legend spread throughout Western Europe and been carried by the crusading armies into their colonies in the Middle East, as the above quote would tend to suggest. The popularity of the Arthurian meme – established by Geoffrey and others in the previous generation – was now a totemic mainstay of the Franco-Norman secular elite, the internationally-acknowledged leit motif of the Christian feudal warrior of the High Middle Ages.

It was within this context that we need to situate the twelve-year project of Chrétien de Troyes, who began the composition of his first Arthurian Romance, Erec et Enide, around 1170. By this time, not only was the popularity of the Arthurian complex already a pan-European phenomenon, but the lines of communication between its courtly Franco-Norman audiences and its native Welsh sources were already well-established. This last point is all too often overlooked by scholars who are inclined to take the view that Chrétien produced his Arthurian Romances ex nihilo. This last misconception has led to a number of misunderstandings about both the works of Chrétien himself and their Welsh equivalents.

Of Chrétien himself we know next to nothing. He describes himself as a poet, but some sort of clerical background may also be inferred (evidence of a classical-liberal education is apparent in his allusions to Macrobius’s Psychomachia, Ovid Ars Amatoria and other staples of the medieval schoolroom). Beyond that, the most specific information we have relates to his patrons Marie de Champagne and Count Phillip of Flanders, dedicatees of Lancelot (composed around 1177) and the Conte du Graal (1182) respectively. This would indicate a career centred on the powerful ducal courts of the northern French plain. While nominal vassals of the Capetian kings of France, the Houses of Champagne and Flanders both fostered alliances with the rising Angevin power to the West. The sponsorship of Chrétien’s Arthurian projects by these aristocratic patrons may be seen in part as cultural nod to this cross-channel connection. But also (no less importantly) the Romances represent a literary articulation of the complex and often conflicting social, political and moral imperatives which were endemic within High Medieval feudal society in general, and were perhaps particularly characteristic of these semi-independent dukedoms. We will have more to say about the social (and psychological) bases of the Arthurian Romance in due course. But first we must return to the other much-debated question of Chrétien’s source material, in particular that which pertains to the matière de Bretagne.

That the basic setting and nomenclature of Chrétien’s Arthurian romances owes its influence (directly or otherwise) to native British sources is not controversial. Rather more problematic is the degree to which this Brythonic material also informed the underlying plot-lines or narrative structures themselves. As with Geoffrey, we also need to reckon also with a generous leaven of classical influence as well as the artistic input of the Champagnoise poet himself. A case in point is the famous ‘love triangle’ complex, which is central to the plotlines of Romances such as Cligés and Lancelot. On one hand, the influence of Ovid as well as the Provençal troubadour tradition was clearly an influence in Chrétien’s presentation and interpretation of these erotic configurations. However, close examination reveals the hallmark of an insular Celtic provenance running through the grain of the narrative itself. The romance of Lancelot begins with what must be considered the atavistic prototype of the royal adultery scenario: that is the abduction of the queen by a demonic otherworld lover. Known as Meleagant in Chrétien’s tale, this mysterious figure arrives at the court of Arthur unannounced and (like Gwawl fab Clud in the First Branch of the Mabinogi, or Gawain’s Green Knight) makes a challenge and a demand which the hosts are foolish enough to accept. It is almost certainly the case that this scenario represents a retelling of a traditional Brythonic tale, an earlier version of which is found in Caradoc’s Life of Gildas, in which a certain Melwas ‘the king of the summer country’ is described as kidnapping Gwenhwyfar and imprisoning her in ‘the city of Glass’.xv The subsequent affair between the queen and her rescuer Lancelot has its own clear resemblance to that of those famous Brythonic lovers Tristan and Yseult,xvi which also finds its echo in the central plotline of the tale of Cligés.

That the debt to the British tradition evident in both of these romances is particularly interesting as neither Cligés nor Lancelot is considered among the more ‘Celtic’ of Chrétien’s works. In his preamble to the latter, the poet describes how the matière and sens (substance and meaning) of the story have been passed onto him by none other than his patroness Marie de Champagne herself. This attribution is open to a variety of interpretations, but at the very least we can draw the conclusion that the countess was effectively sponsoring this representation of herself as an authority on the tradition and its inner significance (a posture that was not untypical this self-assured daughter of Eleanor of Aquitaine). But what is interesting to consider is to what extent the matière de Bretagne had become gained such a foothold in the discourse of the Franco-Norman elite that royal patrons such as Marie de Champagne were boasting of their own expertise in this obscure mythical system which, little more than a couple of generations before, was barely known beyond the hills and farmyards of Brythonic Wales. The particular Welsh traditions represented here were for the most part not derived from Geoffrey’s Historia or its Anglo-Norman translations. Chrétien, Marie, or perhaps a number of their informants, had somehow gained access to their own source of native Brythonic lore, clear analogues for which can be found in Culhwch ac Olwen, the Old Welsh poetry, the Saint’s Lives, the Triads and the other representatives of the ‘pre-Norman’ Arthurian tradition.

We have already seen how, in a previous generation, Geoffrey, Wace and Marie de France were immersing themselves in the Brythonic historical and narrative culture in an effort to extract the distinctive essence required for their literary projects. Their informants were men such as Caradog of Llancarfan – sympathetic to the Anglo-Norman administration but also steeped in the native Welsh cyfarwyddyd. One might expect that such men should have been sought out and consulted in greater numbers as the popularity of the Arthurian meme achieved its fashionable status amongst the European elite. As the twelfth-century progresses this process of cultural transmission becomes ever more diffuse: not least because this matière de Bretagne was evidently in free circulation within French-speaking oral-narrative culture itself. Many of the personal names used by Chrétien and his continuators are traceable to Welsh equivalents, but show traces of Old French phonology: Yder fils Nut (>Edern fab Nudd), Giflet fils Do (Gilfaethwy fab Dôn), Maheloes (Melwas) and Carados Briebas (Caradog Briechfras). Other names have mutated beyond all recognition. The charge that must be answered is whether the matière de Bretagne as a genre of medieval French narrative can really be considered 'Brythonic' in any meaningful sense.

Such a charge would be justifiable if it could be proved that the Brythonic narrative frameworks themselves broke down in the process of transmission, leaving nothing but a vague Arthurian setting and some exotic-sounding nomenclature for the French Romancers to refashion as they choose. This is more or less the view of many critics and literary historians even to this day. Such a view, however, is demonstrably false. Chrétien did not simply take names and locations from the Brythonic tradition – he imported from these sources complex motifs and fully-formed narrative structures. Sometimes this involved a scenario like that of the kidnapping of Guinevere, as suggested above, onto which he built a complex superstructure of his own creation. In the case of Lancelot, Chrétien developed the poignant scenario of the doomed adultery between queen and hero, and infused it with a potent fusion of erotic and religious sensibility. In Erec et Enide, he takes what seems to have been Breton or Dumnonian tale of wandering and giant-killing, woven this into a story of marital disharmony, apparently linking this with the late-antique allegory Mercurii et Philologie.xvii These later accretions were not inherent to the original source, but they cannot be described as much more than a decorative overlay to a narrative infrastructure which was common to more or less all of the romances. Of equal significance, however, is the semantic power of this infrastructure itself, which is inextricably hard-wired into the narratives themselves as we shall see. The unresolved question is whether this infrastructure, which is best explained in terms of the magical plot, is a Brythonic or Franco-Norman innovation, or something of a hybrid innovation: conceived on native Welsh soil, but brought to fruition by the Champagnoise poet under the high noon of feudal Christianity.

The Arthurian Romances of Chrétien de Troyes

Chrétien adopted the basic framework of Geoffrey’s High Feudal Arthuriad, but with a significant shift in perspective. While Geoffrey’s focus is on the king himself and the issues of kingship, the centre of interest in Chrétien’s romances is on the actions and aspirations of specific individuals in the court – primarily the young landless knights of Arthur’s household retinue. Each of the romances focuses on the fortunes of one of these individuals, with the wider sphere of the Arthurian court and its surrounding countryside serving as a backdrop to the hero’s development. The foregrounding of these young retainers signals an important and permanent change within the Arthurian narrative system, a change which would persist into the later tradition. The pre-Norman Arthur of Native Wales is a hero in his right: a dynamic figure who leads his host from the front, whether in battle or in pursuit of supernatural adventure. Geoffrey’s Arthur is likewise a military hero, a feudal warlord, a man of action after the mould of the Anglo-Norman Henry I. In the stories of Chrétien, however, Arthur the King takes a backseat role, with the focus of action in the court almost entirely delegated to the young knight-errantry. This regal senescence is a conscious development, and the unmistakeable generational tension it implies forms a poignant undercurrent of the High Medieval Arthurian Romance, as we will be considering in more detail below.

Each of Chrétien’s tales, then, focuses on the adventures of one of Arthur’s knights. Often there is a developmental theme, a sense of learning and progression, as the young hero moves from a position of impetuous ignorance to worldly wisdom. This process is represented in the form of a journey, or a series of journeys, typically involving a sequence of violent, amorous or supernatural encounters. It is thus in the romances of Chrétien that we first find that most characteristic of Arthurian motifs: the chivalric quest. This represents another important shift from the perspective of Geoffrey’s Historia, with its focus on royalty, statecraft and large-scale military encounters. Chrétien offers an altogether more intimate perspective, focussing specifically on the individual hero: his behaviour and its contradictions.

So the standard form of these romance quest-narratives sees the hero (either as a young knight or an aspirant squire) faced with a bewildering series of physical challenges and magical dilemmas. The hero will typically (but not always) begin the tale as a landless bel inconnu, before eventually progressing into a state of landed, propertied matrimony. In contrast to the grim but unequivocal moral universe of the heroic-age saga or chanson de geste, the world of the romance is characterised by a dream-like ambiguity, in which the hero’s involvement is often dogged by confusion and indecision. At times the Romance touches on the comedic or the burlesque in its treatment of its bewildered protagonists. Literary critic Frederic Jameson points out that in medieval romance literature ‘the hero’s dominant trait is naivité or inexperience’. He emerges as ‘something closer to an observer, a mortal spectator surprised by supernatural conflict, who then himself is gradually drawn in to reap the rewards of victory, without ever quite being aware of what was at stake in the first place’.xviii

This representation of the protagonists involved, which at times approaches the subversively anti-heroic, might be seen in the context of the proto-humanist orientation of this late-twelfth century literature, a sympathy which is largely absent from earlier saga traditions.xix While the leading protagonist of Chanson Roland, for example, wrestles with heroic predicaments in which he must choose between honour and military expedience, the gauche anti-hero Peredur finds himself ensnared in situations which explore more intriguing human dilemmas, the tension between instinctual desire and social propriety being a characteristic example. Chrétien is has sometimes been regarded as the ‘inventor of the modern novel’, and the voyeuristic scrutiny found in his romances of these private but recognisable psychosocial situations, for which the hero is both ‘a vehicle and a registering apparatus’,xx lends some substance to the claim.

However, there are also some importance differences between the modern and medieval forms. We might understand the conscious ego of the protagonist to be the focal point of the modern novel. The Romance, on the other hand, makes most sense if it is recognised that represents things from the point of view of the unconscious instinctual mind. It is this perspective that givr the narrative texture of these ‘magical’ plot-lines often their distinctive primitive, dreamlike quality. Moral contradictions abound. Characters behave in ways that have no immediately discernable explanation. Mysterious emphasis is given to actions or objects which to us would appear inconsequential; while apparently key narrative developments are bypassed, ignored or left resolved. This quality has sometimes confused literary critics of the past, who saw this disjointed narrative constituency as evidence of textual corruption. A more satisfactory reading is derived when we allow the Romance to speak for itself, to project its theatre of desires on its own terms. A psychoanalytic perspective can reveal much of the underlying coherence of the medieval Romance, as we shall see. But a degree of historical understanding is also required: Romance is very much a product of the social, instinctual and spiritual tensions of the medieval Christian psyche in particular. Chrétien’s genius was his ability to rework his inherited Brythonic source materials in a way that underscored these inner tensions, shedding new light on troubled depths, and opening up the medieval mind to new horizons of self-awareness.

We will be examining the characteristic modalities of Romance in greater detail below. But first we must conclude this section with a brief overview of Chrétien’s Arthurian works. These begin with Erec et Enide, written in or around 1170, possibly to celebrate the wedding of Henry II’s son Geoffrey to the daughter of Conan IV of Brittany on Christmas Day of 1169.xxi It is based on a tale some believe to have been Breton rather than Welsh in origin, although this question is by no means resolved.xxii The Romance, which we will be examining in more detail in the course of our study on Geraint, takes as its central theme the jealous rage of hero, a fit of distemper that leads to a prolonged peregrination through a typical Arthurian landscape of hostile knights, giants and rapacious earls. The long-suffering Enide uncomplainingly bears her husband’s churlish treatment as he repeatedly tests her patience and loyalty by exposing her to all manner of hardship and privation. On its surface, this tale develops the characteristically medieval theme of the Faithful Wife, and explores the tension between a knight’s responsibilities as a husband and lover on one hand, and as a feudal warrior on the other. However, as is the case with most Arthurian romances, there is more to this narrative than initially meets the eye. In all of Chrétien’s works there are learned allusions to the scholarly literature of late antiquity. In the case of Erec (as we have seen) it was Martianus Capella’s allegory of the seven liberal arts, De Nuptiis Mercurii et Philologie, which offers this scholastic point of reference. This parallel is made explicit during the coronation feast at the end of the tale, when Erec is clad in a robe woven by fairies with figures representing the four liberal arts of the quadrivium (Geometry, Arithmatic, Music and Astronomy). Some critics have gone further still with this analogy, seeing the adventures of Erec and Enide themselves as an allegory of the process of translatio: with the various enemies encountered along the way echoing the syntactical figures discussed in the Nuptiis.

Needless to say, such rarefied discourse would have sailed over the heads of all but the learned of Chrétien’s peers. For others in the audience, the tale could be regarded on a more human level as a study of morbid jealousy and marital discord. Others may have simply enjoyed it as a fast-moving adventure, following the progress of its hero from one act of military valour to another. Such breadth of appeal was entirely typical of the multifaceted genre Chrétien did so much to develop. Even the least sympathetic critics have acknowledged the ability of the Arthurian Romance to offer so much to so such a wide constituency:

The professional men of letters of the twelfth century knew their business well; they developed their material and conventionalized it at one and the same time: chivalry and courtly love, knight-errantry and faerie, religion, society and morality, mysticism and poetry – place for all of them was to be found in the Matter of Britain, whose mighty accumulations met the diverse requirements of feather-brained page, love-sick courtier, gallant warrior, gentle lady, reverend senior, the dreamer and the man of action, the artist and the student of affairs. It displayed the patter of society and the web and woof of human behaviour; it was as remote as Brocelainde and as close as the nearest tilt-yard. It held treasure for every seeker.xxiii

Erec begins with a classic Celtic narrative sequence in which a hunting expedition leads to an encounter with unearthly stranger: in this case ‘the Sparrow Hawk knight’ who leads the hero into an unknown realm, where Erec initially finds and wins his wife after beating the Sparrow Hawk knight in a feat of ritualised combat. The true name of this character, ‘Yder fils Nutt’ (Edern fab Nudd in the Welsh text), points to an origin in the most archaic stratum of the Arthurian tradition (this Edern was known to hagiographic legend cited by William of Malmesbury in the early 12th century. We might at least entertain the possibility that this individual was brother of Gwyn ap Nudd, a figure from the British pagan tradition). Emerging from the same background is Mabonagrain, whom the hero encounters and defeats towards the end of the tale. Alongside their menacing pagan associations each of these figures appears to echo an aspect of the hero’s own behaviour – embodying an outward projection of the hero’s own anti-social or covetous instincts. As with all the other romances, anxieties surrounding what we will refer to as ‘the sovereignty complex’ seem to drive the plot at the infrastructural level. The traumwerk of the Romance (to use the Freudian term) seems to be to neutralise these anxieties through a process of controlled fantasy, the workings of which we will be exploring in due course.

Between 1176 the 1178, Chrétien had a particularly fruitful period he completed three more Arthurian Romances: Cligés, Lancelot and Yvain. In the first of these, Chrétien makes the connection between the Arthurian world and the antique world of the earliest medieval romances, which tended to revolve around classical heroes like the Ovidian Piramus and Thisbe, the Trojan Dares or Aeneas of Virgilean fame. It was this connection to classical antiquity, incidentally, which gave the genre of Romance its name (> Old French roman – literally a ‘Roman’ tale). The eponymous hero of Cligés is described as the son of Alexander, heir to the empire of ‘Greece and Constantinople’. This Alexander hears of Arthur’s greatness as a youth and travels to his court, where he aids the British king in his campaign against a treacherous usurper. After winning acclaim in Arthur’s war-host, the young Byzantine prince is made a knight and married to Arthur’s niece. Their half-British son Cligés is born soon after.

The main part of the tale hinges on the situation in Greece after the death of Alexander. The sovereignty of the Empire was meant to have passed to Cligés, but was instead usurped by his uncle Alis. Alis then decides he will marry Fenice, the daughter of the Emperor of Germany, and sends his nephew to collect her. Replaying the saga of Tristan and Iseult, the bride-to-be and her royal chaperone fall in love on the way back to Greece. The characteristic love-triangle motif is established, underpinned by a theme of thwarted sovereignty. We also encounter in this tale an early example of the feigned death – another popular medieval device which recurs most famously at the end of Romeo and Juliet.

Throughout the remainder of Cligés, the action alternates between Greece and Arthurian Britain – an apt metaphor, it has been suggested, for the blending of classical and Celtic influences in the poetry of Romance.xxiv Certainly, the tale of Cligés can be seen as an interesting transitional work which makes links not only with the world of classical Romance, but looks back to a representation of Arthur that is more characteristic of Geoffrey of Monmouth’s vision: the British king projecting power in a European sphere, his stature equal to that of the Emperor of Rome. A further affiliation is apparent with Apollonius of Tyre on one hand, and the family of Anglo-Norman romances relating to generic northern figures such as Beaves of Hamtoun, King Horn and Guy of Warwick. If a case is to be made for a wider medieval (non-Celtic) development of the magical plot, then Cligés would have to be considered a key staging post in this process. This is an argument to which we will return.

Lancelot, also known as the Le Chevalier de la Charette (‘The Knight of the Cart’) is another important text from this period which is sometimes considered to lack substantial Celtic input, and on this basis thought to be largely a work of Chrétien’s own imagination, albeit with significant inspiration from classical works such as Ovid’s Pyramus and Thisbe or Echo and Narcissus. However, as we have already seen, key elements of this narrative also belong to the oldest stratum of Arthurian lore – representing as it does the abduction of the queen by an otherworldly prince. The magical realm in which the queen is imprisoned is known in Chrétien’s tale as the kingdom of Gorre, which (it has been suggested) might be linked to the Brythonic motif of the ‘Island of Glass’ (Gorre < Old French Voirre ‘glass’). The ‘Sword Bridge’ which Lancelot must cross has a close parallel in the Sword Bridge crossed by the Irish Cú Chulainn during his wooing of Emer. More certainly, the abductor Meleagant is related to the Brythonic faery king Melwas, who is also recalled elsewhere by Chrétien under the name of Maheloas ‘lord of the Isle of Glass’. Thus the provenance of this otherworld abductor, as well as the manner of his entrance in the opening episode, bears the distinct trace of Brythonic story-telling – whether directly from a Welsh informant, or (more likely perhaps) via some kind of intermediary source from the Old French or Breton narrative cultures.

Beyond this, there is no doubt that Chrétien added much of his own, and some of this evident in his chatty learned references to Ovid and the other schoolroom favourites. The Provençal troubadour tradition has also been identified as the inspiration for the episode involving Lancelot’s seduction of the tower-imprisoned queen. The complex sub-plot involving a community of Britons imprisoned under the rule of the Otherworld king may derive ultimately from a traditional Welsh narrative, similar to that found in the Third Branch of the Mabinogi – but one might also suspect that it had a more specific resonance for the Breton communities of northwest France in the mid-twelth century.xxv Once again, a Breton influence in Chrétien’s early work might be suspected, even if conclusive evidence for its proof has yet to be established.

A notable feature of the Romance of Lancelot is the weakness and passivity of the king, who remains a fairly marginal figure in the Arthurian tradition post-Chrétien. This diminution of status is perhaps confirmed by his conduct in Lancelot, where his impotence and lack of authority are ruthlessly exposed. The king makes no attempt to challenge or avenge Meleagant when the latter menaces his wife and boasts of the oppression of the captive Britons. Soon after, Kay resigns from Arthur’s service in disgust. The king is unable to influence his decision, and ends up sending the queen to prostrate herself on his behalf. Kay eventually agrees to return to the court only on condition that he is given permission to take up Meleagant’s challenge, which involves imperilling the safety of the queen. Arthur is unwilling and perhaps unable to refuse. It is against this background that the adultery of the queen and her rescuer Lancelot is acquires a kind of moral legitimacy.xxvi

While the Champagnoise poet was the first to fully develop its psycho-dramatic potential, the notion of Arthur as the roi fainéant ‘the do-nothing king’ was not Chrétien’s own invention. The Anglo-Norman poet Wace, who produced a vernacular interpretation of Geoffrey’s Historia, was already hinting at a state of affairs in which the king was being eclipsed by his younger and more active household knights. The adultery of the queen was also a theme at the end of the Arthurian section in the Historia itself, and we have little reason to doubt something similar was present in Geoffrey’s traditional source material. The same fate befell the Irish Finn MacCumail, reminding us that the motif of the old grey-head cuckolded by a younger protégée/rival was by no means incongruous to native Celtic narrative system. However, it is only within the feudal, Christian social setting that the moral and psychological tensions within this Oedipal situation could be exploited to the full. It was the achievement of Chrétien to realise this most completely – creating in the process a powerful literary formula which would endure throughout the Middle Ages and beyond.

The third of Chrétien’s romances from this period is also widely considered to be his finest. Yvain or Le Chevalier au Lion has been variously described as the poet’s chef d’oeuvre, or the most ‘characteristic’ of all the romances. It marks a departure from Chrétien’s Breton influence, and (for reasons that will be made clear in our analysis of Owain) it would appear instead to be drawing on material of North British origin, most likely via a Welsh informant. The original tale would appear to relate to the sixth-century Rheged king, Owain son of Urien. At some stage in the story’s development, the memory of this sixth-century warlord seems to have been drawn into the Arthurian sphere. Perhaps as a result of the integration of the Rheged prince into the Arthurian hegemony, the tale acquires a new set of dichotomies, tensions between the personal territorial ambition of the original narrativexxvii on the one hand and the social community of the feudal retainer on the other. It was perhaps here that we have the clearest sense of an emergence of what we will refer to as the magical plot. Whether Chrétien was the first to re-engineer the North British myth of sovereignty in this way, or whether he merely gave a fuller literary expression to the ambivalence of Owain’s integration into the Arthurian system, it is this dichotomy that lies at the heart of both the Welsh and the French versions of the tale.

Literary historian Stephen Knight offers the view that the conflict between proesce (‘bravery’) and cortois (‘civilisation’) characterises of the Arthurian Romance in general, and the romance of Yvain in particular. In the same essay, Knight also refers to the works of Bloch, Koehler, Duby and others who identify a socio-political dialectic within the classical Arthurian Romance – an attempt to present a coherent ideological vision of ‘chivalry’ to unify a fragmented warrior aristocracy, a social class that was facing an erosion of its ancient privileges from an increasingly wealthy and centralising royal power.xxviii Much is explained about the ambivalent attitude to Arthur himself, and Kay his seneschal in particular, by reference to this particular social context.

However, as we will see, the roots of the dichotomy at the heart the Romance run deeper than it might initially appear, and its fundamental significance lies way beyond the specific constellations of French feudal politics in the latter half of the twelfth century. This becomes clear when we consider the central binary opposition around which Chrétien constructed this romance. The first part of the tale sees the hero journeying to an unknown land (that we may refer to as ‘the Fountain Realm’). Here, he encounters and kills a knight who turns out to be married to Laudine, the countess of the Fountain Realm. He then marries Laudine, the widow of his erstwhile enemy, and rules the kingdom alongside her in his stead.

What is being presented here is a primal fantasy of sexual and territorial acquisition following an act of homicidal violence. This basic sequence is what we will refer to as the ‘sovereignty complex’, and can be seen to represent the central problem both in Chrétien’s Arthurian romances and their Welsh equivalents. This fantasy structure used to be associated by an earlier generation of critics with medieval Irish concepts of kingship, and through that to the mythology of a notional pre-Christian pan-Celtic sovereignty cult. Be this as it may, the sovereignty complex at work in the Arthurian romances was if anything even more universal – its dynamics are as easily explained in terms of depth psychology or sociobiology as by a Celtic or Indo-European traditional residuum. What is being touched on here is the ‘savage heart’ of Romance, representing as does the impulses of the instinctual id – the relationship with which has presented the central challenge for civilisation from its earliest beginnings.

We will return to this question in due course, but first we must consider the other side of the dichotomy, which is to be seen in the hero’s curiously complete re-assimilation into the Arthurian milieu during a friendly visit to the court of his erstwhile liege-lord. This occurs after a visit by the king to the fountain realm, which includes an episode of mistaken identity leading to a near fatal encounter with Kay, Arthur’s unpopular court steward. Once these confusions are resolved, Yvain agrees to pay a visit to the court of the king, after promising his newly-wed wife that he will be gone no more than a year (!). However, so involved does he become in the Arthurian social world of jousting and courtly camaraderie that he manages to overstay even this generous leave of absence. It is this Arthurian court – the archetype of feudal civilisation – which is consistently represented as the polar opposite of the anti-social impulses of the animal id in Ywain as in Chrétien’s other Romances. The tension, acknowledged or otherwise, is between the social consciousness of feudal service and integration into the Chrisitian civilisation of Arthur’s court, and the counter-social unconsciousness of instinctual self-gratification – as represented by Ywain’s initial journey to the Fountain Realm.

It was characteristic of Chrétien’s art (though less so of that of the composers of the Welsh rhamantau) that this dichotomy should be represented allegorically, using an established symbolic vocabulary. At one stage in his wanderings, Ywain discovers a serpent and a lion locked in combat. The hero demonstrates an instinctive sympathy with the lion and sets about freeing it from the serpent’s grip. After this, the lion remains with him as a faithful retainer. It accompanies him on his subsequent adventures, which eventually lead him back to the Fountain Realm, and to a state of marital reconciliation with the Countess Laudine. It is the lion which gives the poem its subtitle – Li Chevaler au Lion – and would have been rich in symbolism to the medieval European mind. The lion was considered a king among animals and had pride of place in medieval bestiaries. It was believed to have an honourable nature, and the bestiaries attribute to it ‘noble’ habits, such as an unwillingness to attack a prostrate man. In some contexts it has a Christological significance – following a typology that goes back at least as far as the Physiologus. At the other extreme, the instinctual ferocity of the lion makes it a potential symbol for supernatural evil, as in 1 Peter 5:8 “Your adversary the devil walks about like a roaring lion, seeking whom he may devour”. Elsewhere in the Bible, the Lion stands variously for the wrath of God, the Messiah (‘Lion of Judah’), the pride of the tyrant and the martial power of the victorious army. It was first and foremost as a symbol of secular power, and remains one of the most widespread heraldic motifs adopted by kings, emperors and nation states. As with the dual nature of the lion in the scriptural tradition, this power itself is morally ambivalent and can be used either for good or ill. It perfectly embodies the dual aspect of the knightly warrior aristocracy: a guardian of civil society when bound by the oaths of feudal obligation and Christian virtue; a destructive force of nature when unleashed from these bonds.

The role of the lion in Yvain corresponds most closely to that of the “lion of Androcles” motif, most commonly known to medieval readers through the Vita of St Jerome. In this hagiographic legend, the saint meets a lion on a desert road and removes a thorn from its paw. Thereafter the beast becomes a faithful retainer to the holy man, representing perhaps the submission of the lower animal nature to the higher spiritual consciousness of the Saint. The lion in Yvain might be thought of as having similar significance: representing the virtue of fides or loyalty, through which it attains a kind of moral sublimation. Its significance here is further qualified through its binary opposition with the serpent. To the medieval mind the reptile (sometimes described as dragon) would have represented the diabolic aspect of the animal self, with the Lion standing in for its propensity for these energies to evolve towards a higher purpose.xxix The Lion, then, stands for the animal self redeemed through submission to a higher force. Likewise, Yvain’s acquisition of this feline retainer represents the conquest of his own animal self. It is probably no coincidence that for the remainder of the tale, Yvain’s actions are directed almost entirely to the service of others, and he exhibits very little of the hot-headed instinctuality seen in the earlier part of the Romance.

Yvain Li Chevalier au Lion thus represents the signal narrative of the chivalric myth, lending an almost spiritual quality to the state of feudal service while taking nothing away from the heroism and romance its knightly adherents. It was perhaps the powerful socio-political significance of this narrative which gave it its widespread appeal: it seems to have been rendered into a number of medieval European vernaculars (Middle English, German and Swedish translations survive, as well as the Welsh analogue discussed below). Its combination of fast-moving adventure, erotic fantasy, feudal morality and scholarly allusion lent an appeal for a wide variety of audiences, as suggested above. It remains the most paradigmatic of Chrétien’s tales, and the one which is most often considered by scholars attempting to define the essence of Arthurian Romance.

Chrétien’s last Arthurian Romance was no less influential, and was also translated into a variety of European vernaculars. Perceval, also known as the Conte du Graal, seems to have been left unfinished by Chrétien, leading speculation that the project was interrupted by the author’s death. Whether or not this is the case, such was the interest in the tale that Chrétien had begun, the work saw no less than four subsequent ‘continuations’, leading to the expansion of poet’s original 10,000 lines into a gargantuan verse-epic of over 50,000 lines. These developments parallel the great prose romances of thirteenth-century France in scope and ambition, and signal the way towards late medieval classics such as Malory’s encyclopaedic Morte D’Arthur.

In Perceval, Chrétien develops theme of the Bel Inconnu – the rustic provincial whose arrival at Arthur’s court fascinates and horrifies the jaded habitueés of the royal circle. This sequence is based on a well-worn Celtic theme, which can be seen most clearly in the arrival of the youthful Culhwch, who boorishly rides his horse into Arthur’s hall at the start of the native Welsh tale bearing his name. However, Chrétien succeeds in adopting this sequence for his own ends, using it to comment ironically on the effete sophistications of court community, contrasting these with the rough-hewn virtues of the provincial arriviste. Such observations would have been immediately recognisable to the Chrétien’s audience, and undoubtedly contributed significantly to the chord this narrative vision seems to have struck throughout the courts and palaces of his High Medieval audience.

A related theme is the learning and development of the hero – whose virile proece needs to be tempered with the civilising refinements of corteois. His initial lack of governing graces creates all manner of difficulties for the hero, and towards the beginning of the tale he is seen committing what comes close to being an act of robbery and sexual assault. The story sees him gradually overcome these intemperate qualities, acquiring valuable lessons in wisdom, morality and courtly love which Chrétien represents as the indispensible virtue of the chivalric class

A third theme which was to persist through the subsequent Arthurian literature of Medieval Europe appears in a critical scene in the middle of Chrétien’s tale. Having undergone a series of initiatory tests and revelations, the hero arrives at the court of the mysterious Fisher King where he is presented with a baffling spectacle while dining with his host. A youth parades through the hall carrying a bleeding spear. He is followed by two boys carrying candelabras and a maiden bearing a jewel encrusted golden platter of “so great a radiance that the candles lost their brilliance just as stars do at the rising of the sun or moon.” Here we have the first sighting of the legendry Holy Grail – a motif of cardinal significance in the subsequent literature of medieval Europe. It would later become inextricably conflated with the Christian symbolism of the Eucharist, and through the metrical romance of Robert de Boron it would be specifically identified with the chalice of the Last Supper brought to Glastonbury by Joseph of Arimathea.

Hints of this complex are present to some degree in Chrétien’s tale. The grail, it transpires towards the end of the tale, contains a single wafer of the Eucharistic host which miraculously sustains the ailing father of the Fisher King (who also suffers from a debilitating wound). Beyond that, the significance of the grail procession remains elusive, not least to the hero himself. At several points during the procession described above, Chrétien notes that his young hero refrains from asking for an explanation, fearing that ‘it would be thought impolite’. Later on, he is told that he committed a grave error by not asking about the grail, and that has he done so the Fisher King would have been healed. This compounds the strange sense of guilt and bewilderment, the ‘edgy ambiguity’ that accumulates throughout this tale. Crucially, this apprehension seems to be tied up with a suspicion of certain unspeakable crimes associated with the hero’s family – but the precise nature of this ancestral wrong is never fully revealed. A kind of omerta seems to hang over the significance of the grail, the bleeding lance and the debility of the Fisher King and his father – a lingering sense of some kind of unspeakable family secret which permeates the atmosphere of the romance as a whole.

As I have argued elsewhere, there are some grounds for believing that this peculiar plotline emerged fully-formed from native Wales in the early-twelfth century, and reflects an allusion to the regicide of Gruffydd ap Llywelyn who was killed by his own men in c.1063 – an act of dynastic violence in which many of the most powerful families in the land were to some degree complicit. If this theory is correct, then it provides further evidence of a specific Welsh origin for both the form and the atmosphere of the Arthurian Romance, as well as its more incidental details of setting and nomenclature. As such, it has some bearing on the answer to the time-worn mabinogionfrage, as we will be considering below. However, detached from this specific dynastic-political context, the Grail story seems to have become a useful repository for a more generalised sense of guilt, religious awe and mystification: all of which resonated powerfully with the central emotions of Medieval Christianity.

It remains the case that we know next to nothing about the man who wrote under the name of ‘Chrétien de Troyes’, or how he came to be so well versed in the matière de Bretagne. What we can say, with some degree of certainty, is that he was neither the first nor the only individual in the twelfth-century French-speaking world to deploy the Arthurian topos as his vehicle of inspiration. Wolfram von Eschenbach accredited a certain provençal poet named Kyot as the chief authority for the story he translated under the name of Parsifal. Various forms of the name of ‘Bleheris’ are cited by continental Romancers as the source of the Grail tradition. Chrétien, perhaps rather obsequiously, names his patroness Marie de Champagne as chief his informant of the tradition relayed in Lancelot. From the previous generation we have Geoffrey of Monmouth, Marie de France and Wace of Jersey, all of whom (directly or otherwise) seem to have had independent access with the traditional material of Brythonic derivation. There can be no doubt that a number of individuals – working over two or even three generations – were involved in the process of cultural transfusion that created the genre of Arthurian Romance. So, while it is Chrétien who is so often remembered as the premier romancer, he would be better regarded as a link in a chain, or (more accurately perhaps) as a well-placed node within an extensive inter-cultural network. Chrétien was not the sole mediator of the narrative traditions of native Wales into the thought-world of High Medieval France. However, it is with some justification that he is known as The Master. From his works themselves we can sense a comfortable genius for capturing the structure and atmosphere of the native Welsh tales – born as the latter were out of a highly specific and localised mythological and dynastic-political vision. It was in his hands, as much of those of his sources or redactors, that this narrative essence would be expanded into something altogether more accessible and universal. The story of the Grail illustrates this process as well as any other: how its youthful protagonist could take on the role of the feudal everyman, while its incidental marvels could be eventually carry the weight of the most profound mysteries of medieval Christendom.

The Welsh Romances and the Mabinogionfrage

The rediscovery of a group Middle Welsh texts in the early nineteenth century immediately raised a question which is yet to be conclusively answered after two hundred years of academic debate. That Owain, Geraint and Peredur were related to in some way to Chrétien’s romances of Yvain, Erec et Enide and Le Conte du Graal was beyond doubt, but exactly what was the nature of this relationship? Were the French tales derived from the Welsh, or vice versa? As we have seen, the question of derivation is one that continues to preoccupy Arthurian scholarship even to this day. In the nineteenth century, with its obsessive interest in national characteristics, folklore and mythology, the question was even more sharply defined. Following a methodology inherited from classical and biblical studies, nineteenth century philologists started from the assumption that a single written source – an urtext – must necessarily underlie both the French and Welsh versions. The question was (and, to some extent, continues to be) whether the ‘French artistic genius’ or ‘Native Welsh tradition’ could lay claim to the mythological (or poetic) material on display in this family of Arthurian Romances. While most modern scholars would insist that the reality is altogether more complex, it remains the case that nearly every analysis of the rhamantau is informed by one view or another of so-called Mabinogionfrage or ‘Mabinogion Problem’. The question has not gone away, and continues to be answered in a variety of ways. Before we can attempt to untangle our own interpretation of the evidence involved, a brief review of the main developments in this bicentennial debate is required.xxx

The Welsh tales were first translated (and thus ‘rediscovered’) by the antiquarian William Owen Pughe, who had begun publishing his translations of the Mabinogion tales in the Cambrian Register of 1795. His work was continued by Lady Charlotte Guest, who released her translated facsimile of the Three Romances in 1838 (it is perhaps significant that these were the first of the Mabinogion tales to be published in full to the English-speaking public). Pughe was convinced not only that the Welsh tales were the direct source of Chrétien’s Arthurian poems, but also that what lay before him on the pages of the Red Book of Hergest represented the very source and fountainhead of European Romance. This view was broadly endorsed by Lady Guest, and also informed the approaches of subsequent scholars including Matthew Arnold, Sir John Rhys and (in a later generation) W. J. Gruffydd himself.

However, by the late 1880s, two alternative explanations were being advanced by Continental scholars. The first of these, argued most strongly by philologist Wendoline Förster and his ‘Romanist’ school, simply reversed the initial proposition by identifying Chrétien as the source and the Welsh tales mere derivations. A third possibility, known as ‘the common source theory’ had also been proposed and adopted by (among others) the great medievalist Gaston Paris. According to this view, an Anglo-Norman urtext (now lost) underlay both the Welsh and French versions of the tale. Recent advocates of the theory have tended to locate this Anglo-Norman source to the southeastern borders of Wales, thus placing it in the same local context as Geoffrey of Monmouth, Caradoc of Llancarfan and (possibly) Marie de France.

Finally, in the later twentieth century, various scholars began to venture a fourth possible solution to the Mabinogionfrage. This was essentially a modified version of the ‘Romanist’ view, regarding the Welsh tales as deriving their structure and basic content from Chrétien’s works, but with the qualifier that much was also added by the Welsh redactors in the way of style and supplementary detail from the local narrative tradition. Over the last twenty years this latter view, which we might call the Welsh Adaptation theory, has become the most widely accepted, though there are also a number of scholars (particularly within the field of Celtic studies) who would subscribe to some variant of the Common Source hypothesis.

So what are the merits of these various interpretations? As we will see, linguistic, literary and paleographic analysis has established that the extant Welsh Romances are highly unlikely to predate the late-twelfth century at the very earliest – thus ruling out the old interpretation that the Welsh Romances in their extant form could be the direct sources used by Chrétien in his Arthurian compositions. Nonetheless, the sense that these Welsh prose texts have some kind of autonomy from (if not priority to) Chrétien’s composition has proved remarkably tenacious. Such a view is not without foundation. First of all there is the simple fact that the Arthurian tradition in a more general sense has its indisputable source in insular Brythonic culture, whatever the subsequent role of French or Breton conteurs. That the source-traditions used by Chrétien himself might have remained extant in Wales (or in the Anglo-Norman marches), finally ending up in the rhamantau as we know them today, is not beyond the realms of possibility. Indeed, there are specific elements to found within all three of the Welsh Romances which suggest some kind of independent knowledge of an underlying source-tradition – details relevant to the North British background Owain for example; or the presence of the severed head in place of the grail in the relevant section of Peredur (the significance of this will be discussed in due course). On top of this, the complete assimilation of the rhamantau within the prose-narrative system of medieval Wales – the same tradition that produced Culhwch ac Olwen and the Four Branches of the Mabinogi – contributes to the natural sense of a native origin for this material, which seems to have sprung as readily to the lips of the oral cyfarwydd as it did to the book-learned scribe.

Against this view, it is hard to deny that the Welsh Romances are so similar to their French counterparts that it seems unnecessarily torturous to exclude the possibility of some kind of influence (direct or otherwise) of Chrétien’s oeuvre on Welsh narrative culture. The reason for suspecting that the direction of influence is from France to Wales is the almost certain fact that Chrétien’s works had been written at least a generation before the extant versions of the Welsh Romances were composed. This is not simply a matter of late dating of the manuscripts in which the rhamantau are preserved. A thirteenth-century provenance is written into the very fabric of the Welsh Romances. The social world and material culture depicted in the Romances (Welsh and French alike) is clearly that of the High Medieval feudalism. While such a world could have been conceived (or assimilated) in the Age of the Llywelyns, it harder to imagine this process taking place within the clannish, isolationalist milieu of Owain Gwynedd (c.1100-1170) or Lord Rhys (1132-1198). The culture and political geography of twelfth-century native Wales is more clearly reflected to in the Four Branches or (even more so) in Culhwch ac Olwen. The language of the Romances seems to post-date these more obviously ‘native’ textsxxxi , and the manuscript tradition gives little support to provenance much earlier than the late-twelfth century. More suggestively still, we find allusions to characters and situations from the rhamantau in the poetry of the thirteenth-century gogynfeirdd, whereas such references are entirely absent in the works the late-twelfth century poets.xxxii In short, there is very little evidence to suggest any date much earlier than c.1200 for the composition of any of the rhamantau. A thirteenth-century provenance, on the hand, has much to recommend it, and it is perhaps within such a historical context that we can best understand the relationship with the French Romances.

For the native Welsh, the first three quarters of the thirteenth century seems to have been something of a cultural renaissance, during which time the combination of strong leadership within and the relative disarray of its enemies without gave the nation a rare opportunity to assert its own identity and claim a place among the nations of Europe. Under the hegemony of the princes of Gwynedd – Llywelyn Fawr and then his grandson (also called Llywelyn) Wales was starting to assume many of the characteristics of an autonomous, centralised medieval state – with a coin-based currency, a developed fiscal system and ambassadorial embassies to the Vatican and elsewhere. During the so-called ‘Age of the Llywelyns’ the Welsh elites intermarried more frequently with the neighbouring Anglo-Norman Marcher families, and showed a greater inclination to integrate in a more general sense with the cultural mainstream of Europe than had been the case in the previous century. Literature was an important expression of this outward-facing mood of national self-confidence. Not only did the Welsh in the Age of the Llywelyns show a greater willingness to commit their own narrative traditions to writing, there appears to have been a growing appetite for the ‘classics’ of European courtly literature (both religious and secular) amongst the literate elites of Wales. Hence we find a number of well-known texts from the French-speaking world translated into Middle Welsh and preserved alongside materials of more local provenance in thirteenth-century manuscripts, a number of which are still extant. These manuscripts, in turn, were to become the direct sources of the great codices of the fourteenth century: including of course the White and Red books from which the Mabinogion was abstracted).

Religious and philosophical works found in Welsh libraries during the Age of the Llywelyns included the apocryphal account of the assumption of the Virgin (Transitus Mariae), the Vita of St Catherine along with numerous other works of classical hagiography, and a translation of Honorius of Autun’s Imago Mundi. The secular works translated into Middle Welsh by thirteenth-century scribes included The Song of Roland and various other texts from Charlemagne cycle; The Seven Sages of Rome (a perennial medieval favourite); the Anglo-Norman Romance Beaves of Hamptoun; and the legend of the fall of Troy known in Latin as the Historia Daretis. Unsurprisingly, a particularly popular subject for translation was Geoffrey of Monmouth’s Historia Regum Britanniae – the Welsh version of which survives in over sixty manuscript editions. The Welsh elite, evidently, enjoyed this Anglo-Norman rendition of their own native lore, and the so-called Brut Y Brenhined (‘Chronicle of the Kings’) seems to have become accepted as the authoritative national history, more in tune perhaps with the modernising spirit of the day than the more nativist idiom of the early Triadic tradition or the Four Branches of the Mabinogi. Within such a context, it would be entirely unsurprising to find Welsh translations of at least some of Chrétien’s works taking their place among the accepted Arthurian canon. There seems little reason to doubt that such a process explains the presence of the Tair Rhamant within these thirteenth century manuscripts, alongside works of a similar provenance.

As we have suggested above, to define Owain, Peredur and Geraint as ‘translations’ of European exemplars is overly reductive, and such an understanding has done considerable damage to the understanding and appreciation of the Welsh Romances as works of literature in their own right. And neither is the ‘translation’ model entirely accurate, at least not in the modern sense of the word. Professor Brynley Robert has suggested the broader medieval concept of translatio might offer a more useful way the process by which a narrative such as that of Yvain/Owain might be transplanted from one culture and linguistic environment into another.xxxiii Translatio does not simply involve the rendering of a text in another language, it also describes the process of adaptation into the narrative and idiomatic norms of another culture. Part of this process involves the stripping out of all narrative elements which would be incongruous within the host culture, and the assimilation of the narrative into an acceptable local format. In the case of Yvain and its Welsh translatio, we would expect to see the removal of the exegetical longeurs (there is never any authorial commentary in Middle Welsh tales), the alteration or removal of any culturally dissonant elements from the narrative itself, the addition of various local topographical and antiquarian details, and the rendering of the verse original into Middle Welsh prose. Such a process, in other words, would not involve a word-for-word translation – but rather the adaptation of a narrative theme into an alien linguistic and cultural milieu. There would be a recognisable resemblance between the product of this process and the original narrative, but one would also expect to find substantial differences in terms of narrative content as well as the surface detail of style and idiom. The operation of such a process in thirteenth-century Wales certainly offers a convincing explanation of both the similarities and the differences between the rhamantau and the corresponding French Romances. As Roberts suggested ‘I think it very likely that if a Welsh writer were trying to create or imitate or retell a “French romance” in the Welsh prose tradition the result would something very like Owain.’ xxxiv

Brynley Roberts was one of the key authors of the ‘Welsh Adaptation’ theory, which now represents the dominant consensus in medieval scholarship when it comes to the understanding of the rhamantau and their relationship to their continental equivalents. Recent work on the Welsh Romances has looked in more detail at this active process of translatio and considered precisely what kinds of alterations have been made, attempting in the process to ascertain the transcultural motivations implied by these changes. An example of this kind of work can be seen in a recent paper by Kristen Lee Over which represented the rhamantau in post-colonial terms, as works of ‘cultural mimicry’ which might nonetheless serve to ‘convert romance into a new Welsh genre, restyle Arthur’s court and his knights as a warrior teulu, infuse French cultural influence with warrior themes of the insular literary tradition, and turn against itself the complex of cultural and ethnic superiority inherent to medieval French romance.’xxxv Such a thesis depends on a very specific interpretation of the intercultural relations between Wales and the French-speaking world in the thirteenth century – perhaps one that would be more appropriate to the post-conquest era than the Age of the Llywelyns. In other words, as with so many contemporary studies of the rhamantau, it is contingent on a particular answer to the Mabinogionfrage, and indeed appears to be predicated on the assumption that this problem has been conclusively resolved in favour of the ‘Welsh Adaptation’ theory.

But is such an assumption really justified? There are still a number of scholars who find the common source theory the more persuasive, and in many respects the case in favour of such an interpretation has yet to be conclusively disproved. As we shall see in the articles about the individual rhamantau themselves there are a number of minor but significant details which would suggest that the Welsh Romancers had independent access to a similar tradition to that which informed Chrétien’s narratives in the first instance. Given that this material rests on Brythonic foundations (and in many cases reached France through Welsh intermediaries) it would not be entirely surprising if such analogues were still current in Wales in the twelfth and thirteenth centuries. While it seems unlikely to me that any of the rhamantau were drafted in total ignorance of Chrétien’s influential oeuvre, it is equally possible that local Welsh material was incorporated into the end product both during and after the process of translatio. As part of this process it would be perfectly natural that the memory or knowledge of some pre-existing local variant – related indirectly perhaps to Chrétien’s own source – may have influenced, to varying degrees, the extant content of the Welsh Romances. Such a possibility would alter our understanding of the nature of the inter-cultural exchange, and also point back to some degree to the common source theory.

It is worth underlining the fact that there is no evidence to suggest that the so-called ‘Three Romances’ were ever considered as a textual unit in the thirteenth century, and neither is there any specific evidence of common authorship. As Ceridwen Lloyd-Morgan points out, they do not appear as a continuous sequence in the fourteenth century codices, and appear to have had quite distinct manuscript histories.xxxvi This of course doesn’t rule out the possibility of some kind of textual unity or common underlying literary purpose, but neither can such a possibility be taken for granted. Thus there may be more than one possible answer to the Mabinogionfrage – depending on which rhamant the question is applied to. One of the reasons that this rather tortuous debate has persisted for so long is that, as one critic has noted, ‘there is enough evidence to give a variety of answers … but not enough to give a decisive one’xxxvii. Perhaps, more problematically, there may not be a simple answer which can be applied to all three of the rhamantau and their relationship to the French romances, as there may have been somewhat different evolutionary histories involved in each case. What I will offer here is a brief summary of my own conclusions on the formative circumstances of each of the Welsh Romances, taking into account stylistic, linguistic and orthographical features, as well as the evidence of the structure and content discussed in the articles below.

The editor of the Dublin Institute of Advanced Studies edition of Owain suggests that the style, language and orthography of the text appears to be slightly closer to that of The Four Branches or even Culhwch ac Olwen than that of the other two romances. On this basis, it might be concluded that Owain is both older and less influenced by French literary models than the other two romances. R. L. Thompson even suggests a twelfth-century dating.xxxviii Interestingly, of the three Welsh romances, Owain is also the one for which ‘a common source’ seems the most applicable, as we will see. That the lead protagonist himself, Owain son of Urien, was a historical figure drawn from the Hen Gogledd milieu is uncontroversial. It is also clear that a primitive version of the legend of Owain’s seduction of the Laudine would have existed in insular sources before the advent of Chrétien’s work (a variant of it is preserved in the hagiography of Glasgow’s St Kentigern – who himself was represented as the offspring of this union!) That there are other stories relating to Owain and Urien preserved in Wales along with other elements of the Hen Gogledd tradition is also well-recognised. It therefore seems more likely than not that Chrétien’s Yvain was informed, directly or otherwise, by a Welsh narrative tradition, which probably took the form of a kind of modified territorial sovereignty myth, with roots dating back to geo-political ambitions of Rheged in the late sixth century.xxxix What is striking about the rhamant of Owain is that we find clear evidence of knowledge of this original northern context (with the definitively Lothian figure of Cynan fab Clydno in place of the Calogrenant, for example) which seems to suggest independent access to a source tradition that included details omitted by or not known to Chrétien. None of this, of course, rules out the possibility that the author of Owain was at the same time aware of and influenced by Chrétien’s Yvain. But more so perhaps than with either of the other rhamantau, there is a clear trace of an extant Welsh source tradition – similar perhaps to which was also known to the author of Yvain – which in some senses gives Owain a kind of autonomous authority its own.

At the other end of the spectrum, we have the tale of Geraint. Of all the three Romances it is this one that seems to fit the derivative paradigm of ‘Welsh Adaptation’ most comfortably. It follows Chrétien’s tale episode by episode, and sometimes even phrase by phrase,xl although it has been consummately restyled in the native chwedl idiom, and incorporates additional details native Welsh legal and narrative tradition. There has been some suggestion of a Breton rather than Welsh origin to the source material involved – although this theory is dependent on a particular interpretation of a possible etymology of the names of the hero and heroine. Rather more convincing is the suggestion from the medievalist Beate Schmolke-Hasselman that Chrétien’s Romance may have been composed to celebrate the wedding of Geoffrey Plantagenet in 1169. While this would suggest a Breton context for Chrétien’s first romance, which perhaps have informed its nuptial theme, it does not prove that the material of the story is necessarily of specifically Breton origin. A number of episodes and motifs within Erec have analogues within insular Celtic sources (one thinks of the ‘Chase of the White Stag’ episode, and its relationship to opening scene of the First Branch). However, the closest thing we have to a prototype comes, interestingly, not from Wales but from the West Country of England – the traditional patria of Geraint himself. This takes the form of a fragment of ecclesiastical legend, mentioned in William of Malmesbury's De Antiquitate Glastoniensis Ecclesiæ, which was first written in 1135 (although the Arthurian notes may represent later interpolations) . It features Edern fab Nudd as the hero (who appears as the Sparrow Hawk knight in both the French and the Welsh versions of the romance). Like Geraint/Erec, he is depicted in a fight with three giants, and also like the romance hero is depicted as being carried away from this fight in a dead faint, to be healed at the court of Arthur. It may or may not be significant that this anecdote situates the action immediately after the Christmas feast day (both Geraint and Erec associate their key events with seasonal feast-days – although such calendrical references are not uncommon in other romances). We might speculate that the basic tale, which Chrétien claims “storytellers habitually fragment and corrupt in the presence of kings and counts” owes its origin to some kind of traditional pageant recited or performed at these seasonal feast-days throughout the Brythonic world. (Comparisons might be made here with the calendrical storytelling that Eric Hamp discerns underlying the term mabinogi).

Whether or not this was the case, it would appear more likely than not that some version of this popular tale was extant in Wales (or the West Country?) when Geraint was in its formative phase. One might note that there are one or two details that hint (but do no more than this) at an independent knowledge of an older tradition. One thinks, for example, of the location in Geraint of the first meeting with Sparrow Hawk knight (Edern map Nudd) in the Forest of Dean – the very location of the cult-centre the Romano-British diety Nodens (>M.W. Nudd). Details such as these are less conclusive that the traces of an independent tradition in found Owain, for example, but they do remind us that the complex picture of ongoing mutual influence between the Brythonic and French-speaking worlds, the breadth and complexity of which makes any assessment of prior authority almost impossible to ascertain.

An even more complex picture is presented by Peredur – not least because of the extreme variance between the different manuscript versions of the text extant. Most scholars agree, however, that all versions of the text contain something very similar to the first part of the Le Conte du Graal in its opening episodes. The longer version as found in the White and the Red Books also shows the influence of Chrétien and his continuators in its final section. The intermediate sections contain elements which could well have come from the chansons de geste and Anglo-Norman romance. The hero’s ultimate love interest – the exotic ‘Empress of Constantinople’ – resembles nothing so much as the oriental beauties wooed by characters such as Beeves of Hamptoun or Guillaume d’Orange. At first glance at least, Peredur looks like a straightforward case of creative translatio, which (as both its linguistic profile and manuscript context would suggest) probably took place at some time in the mid-thirteenth century. (The influence of Owain is also possible to detect in Peredur – with the ogrish 'Du Traws' appearing in both tales).

However, a closer examination of both this text and its Old French exemplars raises a number of questions. We may wonder about the nature of Chrétien’s source material, which may or may not have been a Welsh or Northern British tale similar to that which informed Yvain. We have noted Celtic analogues to the Bel Inconnu motif and the ‘Loathly Lady’ (both of which were to become Arthurian staples). We might also add that a basic ‘Exile and Return’ and ‘Prophesied Death’ framework can be seen beneath the surface of this tale which is typical of (though by no means exclusive to) the Irish and Welsh tradition. However, most troubling of all, is the severed head which replaces the miraculous cup in the Welsh version of the Grail Castle episode. There are compelling reasons (dealt with in more detail elsewhere) for believing that this refers obliquely to the killing of Gruffydd ap Llywelyn in c.1063. The role of the witches of Gloucester in Peredur, and their possible origin in a pagan cult from the Gwent/Forest of Dean area (an equation backed by older Welsh and Breton sources) may also be significant. If my interpretation of this evidence is correct, it would imply that Chrétien’s source for the Perceval romance was a fully-formed Welsh legend of exile and return which was current in the early twelfth-century, where its typological references to the killing of Gruffydd ap Llywelyn would have carried a potent resonance (this dynastic crime which seems to have cast a considerable shadow over the succeeding generation of the Welsh royal kin). The location of such a source in a Welsh, twelfth-century context has the potential to redefine, in no small measure, our understanding of the early history of Arthurian Romance.

My purpose here has not been to reanimate the corpse of the interminable mabinogionfrage, but rather to lay it gracefully to rest – by bringing it to what I would regard as its inevitable conclusion. The earliest answers to question – in keeping with the nationalistic preoccupations of the day – would accept nothing less than a total victory for either ‘Celtic-Welsh mythological tradition’ or ‘French poetic genius’ for the honour of the riches of Romance. Subsequent theories modified these initial counterpoints into a more nuanced debate between the ‘common source’ hypothesis vs. a translatio model of Welsh adaptation, but the dividing lines remains essentially the same. The fifth and final stage of this debate might involve accepting a model that was more fluid and more complex still: a recognition that the developmental circumstances of each of the three rhamantau may not have been identical, but in each case a process that involved both an adaptation of Chrétien’s originals and the use of pre-existing local source traditions is not inconsistent with the available evidence. Given that it is clear that Arthurian Romance evolved out of an ongoing dialogue between Wales and the French-speaking world, is it not sufficient simply to regard the rhamantau as products of this fertile hybridity? Once we remove the rather unnecessary concept of the urtext, it is possible to conceive of a situation in which a several versions of each of the romance narratives were current throughout the Brythonic world and also (from the post-Galfredian period onwards) throughout French-speaking Europe as well. By the late twelfth century, a derivative text could be influenced by one or more of these versions. What is more interesting than the question of origins is the persistence of particular motifs, but also their specific use within particular medieval contexts. This is what we will be attempting to understand in the following section.

Such a perspective allows for a variety of interpretations. The adoption of the romance within the Welsh narrative system could be seen as ironic mimicry of Anglo-Norman mores; an aspirational assimilation into the High Medieval mainstream; or a defiant reclaiming of a native tradition. Maybe it was all of these things, at various times. Equally likely (or so it seems to me) these transcultural borrowings (in both directions) were a fairly unselfconscious process – the stories were reproduced simply because they were good stories. What made them enjoyable for the audiences of thirteenth-century Wales and twelfth-century France is precisely the same quality that makes them so often baffling for readers today – that is they resonated specifically with the medieval mind, and were designed to disclose and assuage intra-psychic fault-lines that were particular to the world of feudal Christianity. Chrétien’s input seems to have been to bring sharp focus to his inherited source materials, foregrounding the experience of its individual heroes and giving emphasis to the narrative peaks and troughs that would be felt most acutely by his courtly audience. Yet the basic contours were already present within his source material, and (I suspect) already well understood – if differently expressed – within the narrative culture of medieval Wales.

The Magical Plot: Exploring and Defining the Modalities of Romance

The investigations of Sigmund Freud and the Vienna School of psychoanalysis revealed that human motivation habitually operates at a number of different and often contradictory levels. The truth and value of this observation is not diminished by the questionable benefits of psychoanalysis as a therapeutic tool, nor by the declining fashionability of post-structural theory. It remains the case that depth psychology has its uses in the study of Arthurian romance – where the contradictory forces of human motivation are evident in both the formative structure of the tales themselves, as well as the behaviour of the characters depicted within them.

The application of psychoanalytic perspectives to traditional narrative discourse is of course nothing new. Freud himself on more than one occasion equated creative activity of dreaming mind with that of the folktale, and explored this hypothesis more fully in essays such as The Theme of the Three Caskets (1913). His followers such as Silberer and Rank continued this work in various explorations of their own, while C. G. Jung pioneered his own approaches in essays such as The Phenomenology of Spirit in Fairytales (1948). Similar lines of interest were pursued by the Gestalt school of dream psychology, and by literary-critics who also informed by structuralist approaches such as Vladimir Propp’s Morphology of the Fairytale (1928).

It is against such a background that medievalist Dr. Anne Wilson seems to have formed her ideas of the ‘magical plot’ in traditional tales: using the Arthurian romances of Yvain and Perceval as two of her primary case studies. Wilson proposes that traditional tales (a broad category that seems to encompass folktales and medieval romances, as well as certain modern works) operate on at least two different levels: a ‘rational’ level, in which events follow an explicable cause-and-effect sequence we typical of a contemporary novel; and an underlying ‘irrational’ level, where the narrative is structured by a more dream-like consistency, in which a type of logic Wilson refers to as ‘magic’ is used to manage the fears and desires of the unconscious mind. At this level, Wilson contends, the text behaves exactly like a dream in that all the events and contents relate, in a solipsistic way, to the desires of the dreamer (i.e. the author and the audience) as represented by the hero.

Wilson’s use of the word ‘magic’ is entirely typical of her approach – which (unusually for an academic writer) seems to be predominantly tacit and intuitive, rather than theoretical and didactic. For Wilson, ‘magic’ has a very specific meaning, describing the activity of a certain level of thought: a primal process of defensive ideation thrown up by the ego as a last resort against the state of anxiety and helplessness, a shoring-up of psychic resources against the collapse of the human self into a state of animal panic.xli Beyond this, magic can be used to control, defend, manage or indulge other overwhelming instinctual forces: notably those relating to guilt and desire, where we find the most characteristic tension between the superego and the id. ‘Magical narratives’ or ‘fantasy structures’ are created in response to this situation, and it is in this form that they appear embedded in traditional tales such as those we are considering:

Magical plots have urgent feelings to satisfy or dispel, in situations that seem extraordinary to the rational mind but are nonetheless real in their own area of experience. The thought concerned belongs to a world consisting of the protagonist and parental figures, and is therefore haunted by prohibitions: the desire for sovereignty (which is to be understood in its widest sense) is haunted by a sense of treachery, and sexual love by ideas of theft and incest.xlii

It could be argued that this Oedipal focus offers a rather narrow definition of the scope and content of magical ideation. Indeed, I have found what would appear to be a similar type of process at work in the Four Branches of the Mabinogi – but here the desires and anxieties are communal rather than individual. Thus, on the ‘magical’ level of the Four Branches, the narrative often represents complex ritual sequences for the geo-political purposes such as apotrophaic territorial defence (by means of the burial of the head of Bendigeidfran); or the exorcism of threatening otherworld forces through a of series of moves culminating in the ‘eagle in the oak’ sequence towards the end of the Fourth Branch. All of these would fit Wilson’s understanding of magical thinking as ‘active and creative – pursuing a purpose, finding means, investing power in things, believing in the power, and making use of it.’xliii

However, in the context of the Arthurian Romance I believe Wilson is essentially correct to identify ‘the desire for sovereignty…in the widest sense’ (my italics) as the main driving force behind the magical plot. We do not have to fall back on reconstructions of Celtic pagan religion to see that this is the case – the problem of the ‘sovereignty complex’ is on one hand more universal than this, and on the other specifically congruous with the social and economic conditions of the feudal middle ages. By the sovereignty complex I refer to conflation of the concepts of regicide, sexual gratification (copulating with the spouse of the murdered king) and territorial acquisition – a primitive, instinctual conflation of ideas that we often find in association at the ‘magical’ level of Arthurian Romance.

It would not be difficult (should one feel the need to do so) to identify a variety of sociobiological drivers behind this fantasy of the violent seizure of territory and its association with sexual and reproductive gratification. But of considerably more interest here is the specific connection with the feudal Middle Ages, and in particular with the socio-economic position of that class of landless retainers with which the Arthurian Romances are most directly concerned. Literary historian Jacques Le Goff (here quoting Georges Duby) calls attention to a particular group within twelfth century aristocratic society known simply as ‘youths’ (juvens) in Old French, cnihtes (>‘knights’) in Middle English:

Duby writes: “A youth was in fact a grown man, an adult. He was admitted to a group of warriors, given arms, and dubbed; in other words he became a knight…Youth can therefore be defined as the part of life comprised between dubbing and fatherhood” which could be very long time indeed. Youths were footloose, vagabond and violent. They were the “leading element in feudal aggressiveness.” And their long, adventurous quest – “a long sojourn shames a young man” – had a purpose: to find a rich mate. “The intention of marrying seems to have governed all of a young man’s actions, impelling him to cut a brilliant figure in combat and to show his prowess in athletic matches”. Marriage was made more difficult by the proscriptions of the Church, which often made it impossible to find a bride close to home. Duby himself has noted the inescapable parallel between this situation and that described in the courtly literature.xliv

That the heroes of Arthurian Romance (both French and Welsh) can be described in these terms is beyond doubt. They are almost all represented in the first instance as landless youths, drawn into the ‘warrior society’ at the court of Arthur, but subsequently driven to wander further afield. They are driven by a kind of compulsive restlessness which characterises feral youth in its quest for material and sexual satisfaction. The life of these young men, in historical fact as much as in the Romantic imagination, was a complex field in which bravery and aggression were required to win a place among the settled world of married, property-holding adults; while at the same time these very same instincts were repeatedly (and necessarily) curbed by the injunctions of church and feudal bond. This tension between desire, aggression, guilt and vindication – the classic elements of the magical plot – are the chief drivers of the narrative engine of Romance.

It may well be no coincidence that this particular narrative genre became popular in Wales during the thirteenth century. We might recall the changes that were of such interest to the writer of the Four Branches in the previous generation: namely the tightening up of incest laws and the devaluation of the clan network in favour of the more exclusive system of primogeniture. These changes may well have lead to a more sharply defined knightly caste in Wales during the Age of the Llywelyns, whose local restrictions in terms of sexual or economic opportunity seem to have forced them abroad into circulation among the feudal aristocracies of the Anglo-Norman or Frankish worlds (the involvement in young males from the Welsh aristocracy in trans-national enterprises such as the Crusades being an example of this trend). The rising popularity of the Romance in the courts of Wales is both a cause and an effect of this adoption of Continental norms amongst the native warrior elite.

The validity of the label tair rhamant for these three Arthurian tales has been called into question by Ceridwen Lloyd-Morgan, who points out that the term is not used within the texts themselves in the way that (for instance) mabinogi was used at the end of each of the Four Branches. As we have seen, the texts do not necessarily constitute a thematically-linked group and appear to have had quite diverse text histories. Lloyd-Morgan even questions whether the three should be thought of as ‘romances’ at all, pointing to the lack of certain features (for example, the link between sens and matière) which she regards as essentially characteristic of the genre.xlv At this point, however, the line of reasoning starts to look suspiciously circular. It becomes possible to define what is and is not a romance and, on that basis, either include or exclude the three Welsh tales from according to preference. I shall not therefore be pursuing the question as to the appropriateness or otherwise of these labels for the texts in question, but will continue to to use the terms rhamant(au) and Welsh Romance(s), simply because these are convenient and recognisable designations. (Should a more appropriate term gain acceptance, I would be happy to adopt it.)

Of greater interest, in my view, is the functional operation of the Welsh Romance: the workings of which I have sought to disclose in more detail in the chapters below. I will conclude this introduction by reference to the problem of the historical origin of the magical plot, a problem that is yet to be resolved. While it is not difficult to detect and understand the function of the magical plot within the context of High Medieval romance, it becomes harder to do so in other genres (both ancient and modern) as Wilson has sometimes sought to do. Among her canon of magical narratives is included certain popular works of nineteenth- and twentieth- century fiction, such as Daphne Du Maurier’s Rebecca, Elizabeth Wetherell’s Wide, Wide World and Christoph von Schmid’s Basket of Flowers. These are analysed as magical plots, or plots containing magical elements, alongside more obviously traditional narratives such as Cinderella, Cap O’Rushes, The Goose Girl in their numerous variants (found in sources from the late-medieval to the relatively recent). Her logic in defining what is and is not a magical plot in texts such as these is not always clear. A more workable definition is probably required before a comprehensive historical analysis of this intriguing literary mode can be fully understood.

It is tempting, of course, to trace the mode back to Arthurian Romance, and even wonder if Chrétien developed it after picking up on the edgy allusions that apparently characterised even quite informal social discourse in native Wales. xlvi A specifically Welsh pedigree for this aspect of the Arthurian Romance is suggested at the end of our analysis of Peredur. But it should be understood that this question remains far from resolved. Wilson herself makes a convincing case for the late classical Apollonius of Tyre as a prototypical magical plot. My own suspicion is that once it is better understood, the magical plot will be detected at work in a surprisingly wide variety of narrative structures, even opening up the possibility of an ahistoric, polygenetic interpretation: where universal human psychological condition rather than any specific lineage of influence might be seen as the key to this particular form of narrative discourse.

Nonetheless, the Arthurian Romance has a special significance when it comes to the understanding of the magical plot. And for the social-historical reasons suggested above, it is perhaps not surprising to find that it was within this particular context that we find the imperatives of magical plot is at their most insistent. It may well have been Chrétien who initially developed the Romance as a vehicle for the magical plot, but are signs that the same type of discourse was being used by other storytellers, apparently quite independently of his influence. One might mention in passing the group of Anglo-Norman ‘family romances’ that includes King Horn, Beeves of Hamtoun and Guy of Warwick – all of which, to varying degrees, feature devices characteristic of the magical plot. The influence here does not appear to be Chrétien or the matière de Bretagne, but more probably from the aforementioned Apollonius of Tyre, which retained an extraordinary popularity throughout the medieval world.

So if we are to seek out a historical origin for this narrative mode, it would be unwise to focus too exclusively on Arthurian Romance or its Celtic background. Nonetheless, as I hope to demonstrate, the Welsh Romances show a particularly adroit handling of the magical plot, and we might well wonder (once again) whether some understanding of this nature was already present in the native storytelling culture prior to the importation of Chrétien’s romances. xlvii


iThe issue of the date and provenance of these texts will be discussed below
iiMab.p.xxviii
iiiKristen Lee Over ‘Transcultural change: romance to rhamant’ in Medieval Celtic Literature and Society ed. Helen Fulton (Dublin: 2005)
ivMany of these battle sites are located in southern Scotland, leading some critics to suggest that this earliest stratum of Arthurian tradition was northern in origin.
vVan Hamel A. G. ‘Aspects of Celtic Mythology’ Proceedings of the British Academy xx (1934) 207-248
viThe Arthur of the Welsh (ed. Bromwich, Jarman and Roberts; Cardiff: 1991) pp. 82-83
viiMab. p. 128-129
viiiJ.B. Coe and S. Young (ed. and trans.), The Celtic Sources for the Arthurian Legend (Felinfach, 1995), p. 47)
ixOne thinks of the proto-Geraint Edern fab Nudd episode discussed below (see n.#). This appears ‘"On the Antiquity of the Church of Glastonbury” (c.1130) although it may be a later interpolation
xSee Knight, Stephen Arthurian Literature and Society (London: 1983) pp.38-66
xi‘Professional Interpreters and the Matter of Britain’ (Cardiff, 1966) p.14 f. Quoted in Mac Cana (1992) pp.15-16.
xiiMarie de France was an Anglo-Norman poetess from the mid-twelfth century, whose works include a rendition of Aesop’s fables, as well as the musical ballads or laïs, which were often based on the themes of Arthurian Romance. Little is known about her life circumstances, although she is sometimes identified with Marie Abbess of Shaftesbury, cousin of William Counsul, Earl of Gloucestershire and the lordship of Glamorgan. Bromwich (1991, p. 287) notes that a number of motifs in Marie’s laïs can also be found in the Mabinogi.
xiiiSee Pronsias Macana The Mabinogi (Cardiff: 1992) p.13 ff. It should be noted this interpretation of the supposed anomalies of Middle Welsh prose is not universally accepted.
xivThe Fitzhammond dynasty, of whom Robert and William Counsel (see above) had a blood relationship with Angevin dynasty, and played an important role in the civil war between Stephen and Matilda, helping secure a powerbase for Henry II’s mother in the west of England. Robert Counsel was one of the key sponsors of Geoofrey of Monmouth’s Arthurian project, and in general terms (via figures such as Marie de France) they might be seen as the Anglo-Norman conduit through which Welsh material was mediated into the French cultural bloodstream.
xvSee Arthur of the Welsh (ed. Bromwich, Jarman and Roberts; Cardiff: 1991) pp. 57-58, 83. The abduction of Gwenhwyfar by the faerie king Melwas seems to have been a well-established theme in the early tradition. In the Life of Gildas, the hagiographer cleverly exploits the otherworldly aura of the names aestiva regio (‘Summer Land’ i.e. Somerset) and urbs vitriae (‘Glass City’ i.e. Glastonbury) to effectively euhemerise this archaic tradition (which may well have once formed part of a pagan conception myth), which in turn seems to have been ‘remythologised’ by the secular narrative tradition. Such a process is entirely typical of medieval Celtic narrative culture
xviThe Tristan tradition seems to owe its origin to a North British context (judging by the name of the hero, which appears to be derived from the Pictish Drystan). The influence of the Gaelic legend of Diarmuid and Gríanne is also evident.
xviiThis interpretation of Chrétien’s Erec, first proposed by Karl Uitti in 1981, has recently been explored at greater length by K. Sarah-Jane Murray (From Plato to Lancelot – A Preface to Chrétien de Troyes, Syracuse University Press, New York:2008)
xviiiFrederic Jameson ‘Magical Narratives: Romance as Genre’ New Literary Review VII pt.1 (1975) p.139
xixThe circumstances precipitating the emergence of this distinctive narrative form in the final decades of the twelfth-century remain unclear. Aside from the Welsh influence previously discussed, this literary fashion might be seen in the context of the more general cultural and intellectual foment that was taking place during the so-called ‘twelfth-century renaissance’, during which time European elite was becoming exposed to a variety of new influences. Contact with Arab civilisation, which was at that time both culturally and technologically superior to the feudal society of Latin Christendom, seems to have acted as a powerful stimulus on the European sensibility. The rediscovery of the literary voice of the classical antiquity (and that of Ovid in particular) may have also contributed to a more urbane and ‘knowing’ tone in the narrative compositions of the day. However, none of this quite explains the sudden protrusion of the fantastical into the courtly literature of the Medieval West. The truth is that the magical elements associated with Romance were not so much invented in the late-twelfth century, as rediscovered (perhaps initially within the native narrative culture of South Wales, where it lay closest to the surface) and reworked into a feudal narrative culture that had previously been dominated more exclusively by the martial preoccupations of the warrior-aristocracy.
xxJameson, ibid. p. 139
xxiMedievalist Beate Schmolke-Hasselman suggests that the tale may even have been composed at the request of Henry II (‘Henry II Plantagenêt, roi d’Angleterre, et la genèse d’Erec et Enid’ Cahiers de Civilization Médiévale 24 (1981) pp. 241-246). Geoffrey’s wedding took place in the Breton port of Nantes, where the coronation feast at the end of Erec is also set Nantes (also at Christmas). In the latter ‘dukes and kings from many different countries’ are described as attending ‘Normans, Bretons, Scots, Irish…powerful nobles from England and Cornwall, and from Wales to Anjou, [from] Maine [and]Poitou’. This does indeed sound reminiscent of the Angevin wedding at Nantes in 1169.
xxiiThe theory of a Breton origin has been supported by, among others, Rachel Bromwich, who pointed to the coincidence of nomenclature between the hero and heroine, and the archaic names of the Eastern Breton subkingdom of Gwened – which was also known as Bro-wened (after the Gallo-Roman tribe, the Veneti) and Bro-weroc (after Gweroc, apparently one of its early sub-Roman chieftain). This, Bromwich has suggested, implies a sovereignty myth at the base of the Romance, further evidence for which she finds in the archaic ‘White Stag’ motif. This suggestion is ingenious, and possibly valid. But we also need to bear in mind the evidence for an insular origin – which would include the fact that the basic germ of a similar story (with Edern fab Nudd as the hero, rather than Erec/Geraint) seems to have been known to William of Malmesbury in the early twelfth century, who gave the legend a West County backdrop (this appears William's history of Glastonbury Abbey, but may be a late interpolation). Given the festive context of all versions of this legend (Whitsuntide in Geraint, Christmas in Chrétien and William of Malmesbury), one explanation could be that it derives from a traditional Brythonic seasonal tale, which had been variously reworked in a variety of contexts: in Wales, Brittany and the West country.
xxiiiMab. p.xxx
xxivMurray, op cit. p. 215-216
xxvThe rivalry between Cligés and Alis mirrors that of the eleventh century Conan II and his uncle Eudes, who had held the regency during the young duke’s minority. Brittany was later to play a significant role in the politics of the kingdom of England, when the future Henry II would usurp Stephen with the help of a Breton army, recalling events in the first section of the tale, when the Greek hero marches with a Breton army to help the king put down an attempted coup led by Arthur’s steward Count Angres. Such was the strategic significance of Brittany in twelfth-century northwest Europe that the Angevin king continued to interfere in Breton dynastic politics, fomenting rebellion against the independent-minded Conan IV. For twelfth century readers connected with Britanny or the Angevin court, it would have been difficult not to have seen resonances of this situation in the Romance of Lancelot.
xxviOne wonders to what extent this scenario was shaped by the career of Marie de Champagne’s mother, Eleanor of Acquitaine, whose marriage to the weak and unworldly Louis VII was annulled in favour of the dashing young Henry II of England.
xxviiQuoting from the article on Owain elsewhere on these pages: “Owain son of Urien, we might remember, was a genuine historical figure who seems to have thrived in sixth-century Rheged, and his association with this other well-known Man of the North is one of a number of suggestions that the source-tradition was a northern dynastic legend which seems to have provided the Welsh Romance of Owain with additional material not found in Chrétien’s Yvain (and vice versa). In this respect, the accumulated evidence of this northern background might be seen as supportive of the ‘common source’ theory. This is a question to which we will return. However, for the moment, the key point to take on board is the possible significance of the fact that Cynon is a Lothian hero, associated with the Southeast of Scotland, just as Owain is a Rheged prince associated with the western area of the border region. Some interesting possibilities emerge when we consider the likelihood that these details were present in the original tradition which may well have been contingent on the wider social context of the sub-Roman north. In terms of the inter-tribal politics of the region, it would have been regarded as significant that Owain (the Rheged hero) succeeded where Cynon (the Lothian contender) had failed in what has often been read as magico-mythical bid for a kind of pan-Northern British sovereignty. As with much of the early Brythonic tradition, the focus of interest here seems to have been geo-political – a question of which groups or individuals had rights over which territorial claims, and for what reasons. There are reasons, as we will see, for believing that the source tradition underlying both Owain and Yvain may have been a tribal-political tractate of this kind, justifying the extension of a Rheged-based hegemony over the Eastern Scottish borders, and perhaps North Britain as a whole.
In the sixth and seventh centuries, when the Brythonic traditions of the Hen Gogledd were in their formative phase, the hegemony of the North was a live political issue. The strength of the Northumbrian advance had already destroyed a number of the smaller Brythonic kingdoms, and (as the saga englynnion tradition tends to imply) the ability of the North British combrogi to mount a concerted resistance was seriously undermined by factional in-fighting and rivalry between local dynasties. Under these conditions it would not be surprising if the House of Rheged, as the strongest of these kingdoms, might have presented itself as the potential rallying point of a pan-Brythonic northern coalition (much as Gwynedd was to do, with some success, in Wales in the thirteenth century in response to the Anglo-Norman threat). Rumours of Owain’s supernatural origins, as well as his seduction of the King of Leudonia’s daughter – apparently staples of Bythonic mythical lore – may well have begun life as articles of inter-dynastic propaganda arising from this particular sub-Roman North British context.
xxviiiArthurian Literature and Society p.68 ff (London: 1983) xxixThe modern reader might be tempted to equate the serpent with the reptilian ‘back-brain’ or automated impulses, while the lion would stand in for something closer to the human ego – a kind of pre-rational emotionalism that is also characteristic of higher social mammals.
xxxFor a concise summary of the debate, see Doris Edel ‘The Mabinogionfrage: Arthurian Literature between Orality and Literacy’ in (Re)Oralisierung ed. Hildegard L. C. Tristram (Tubingen: 1996)
xxxiExamples of these ‘later’ linguistic features include the third person preterite ending –awdd, which are relatively uncommon in CO and PKM (where the more archaic preterite form –wys is typically used), but increasingly prevalent in the rhamantau texts. However, the evidence here is by no means conclusive. The –wys preterite is still frequently used in the Three Romances, and in Owain in particular. There are also, as noted by R. L Thompson in his edition of that tale, Owain also exhibits some traces of what would appear to be twelfth century orthography, as well as various other archaic linguistic features
xxxiiTYP p.lxxix
xxxiiiFor a recent discussion of this concept, see Erich Poppe ‘Owein, Ystorya Bown and the problem of ‘relative distance’ Arthurian Literature 21 (2004) pp.73-94
xxxivIn The Legend of Arthur in the Middle Ages ed. P. B. Grout et al. (Cambridge, 1983) p.181
xxxvKristen Lee Over ‘Transcultural change: romance to rhamant’ in Medieval Celtic Literature and Society ed. Helen Fulton (Dublin: 2005) p.204
xxxvi'Medieval Welsh Tales or Romances? Problems of Genre and Terminology' Cambrian Medieval Celtic Studies 47 (2004) 42 ff.
xxxviie.g. R. L Thomson ‘Owain: Chedl Iarless Y Fyfynnon’ in The Arthur of Welsh ed. Bromwich, Jarman, Roberts; University of Wales Press (Cardiff: 1991) pp.167
xxxviiiOwein (DIAS: 1986) pp. xvi-xxii. When in the twelfth century rather depends on the respective dating of the Four Branches and the Breudwyt Macsen – texts which match the linguistic profile of Owain most closely. I have dated the Four Branches to the late twelfth century, on the basis of its typological references and the political geography reflected therein. Maxen I would place at some point in the thirteenth century between the translation of Brut Y Brenhinedd and the death of Llywelyn Fawr, i.e. 1200-1240. On this basis I would suggest a tentative dating of Owain between 1190 and 1210.
xxxixThis is explained more fully in the article on Owain below. An alternative explanation, ventured by R. S. Loomis, is that a Breton conteur heard the story while in Scotland, and it entered the French-speaking world through this route rather than via Wales ‘Scotland and the Arthurian Tradition’ Proceedings of the Society of Antiquaries of Scotland 89 19955/1956 p.9. While such a scenario is by no means impossible, like all theories involving ‘Breton Conteurs’ as a primary source of Arthurian legend, it requires ignoring the existing evidence of a substantial Welsh tradition in favour of a hypothetical Breton channel, the evidence for which has yet to be established. As such, it does not stand up well to Occam's razor.
xlThis observation was made by Roger Middleton, a leading authority on the text in Arthur of the Welsh (University Press, Cardiff:1991) p.147
xliThis psychological situation, well recognised in the field of clinical psychiatry and social anthropology is powerful described by Ernesto de Martino “The presence is surrounded by a perilous beyond, one that causes anguish and anxiety and disturbs the presence’s horizons; the world also sees its own horizon enter a critical state and is perpetually passing into this anxiety-ridden beyond. At its outer limits, every rapport of the presence with the world becomes a risk, a collapse of the horizon, an inability to maintain itself, a complete abdictation of control…Magic rises up to fight against this and sets itself against the process of disintegration. It establishes a series of institutions that warn of, and eventually overcome, the threat.” Primative Magic (1972/1988) p.152
xliiPlots and Powers (Florida: 2001) p. 24
xliiiibid p.10
xlivfrom Jacques Le Goff The Medieval Imagination (1988) pp.128-129
xlvLloyd-Morgan, op. cit. p.50 ff.
xlviGerald of Wales in BK1 Ch 14 of the Description of Wales notes: 'In Wales courtiers and even family men in their own homes are very funny in what they say. This amuses their guests and gives them the reputation of being great wits. They make the most droll comments, some of which are very clever. The love references, ambiguities and equivocal statements. Some of these are just for fun, but they can be very bitter.'
xlviiI have made use of the ‘magical plot’ model extensively throughout my analysis of the Mabinogion texts, in particular the rhamantau and the Four Branches of the Mabinogi. However, my approach differs from Wilson’s in a number of ways, and it might be worth acknowledging some points of departure. Unlike Wilson, I have not always found the magical plot (as I understand it) to be in conflict with the ‘rational’ level of the text, though I have noted that its purposes are often artfully concealed beneath a system of substitutions and counter-devices. This raises the question of the consciousness or otherwise on the part of the author and the audience of this covert agenda. Wilson appears to suggest that it was a largely unconscious process, and should be approached much as one would approach dream-material in classical psychoanalysis, i.e. from the assumption that the significance and purposes of the dream material were hidden from – and indeed resisted by – the dreamer’s conscious ego. I remain agnostic on the question of whether the creation of magical plots in medieval texts was a purely unconscious process, or part of the conscious artistry of a narrator like Chretien or the writers of the Welsh Romances. My initial impression, with particular reference to the rhamantau, is that it is often not incongruous with the other textual motivations, such as the relaying of antiquarian lore or the insertion of a contemporary typological reference. Where such imperatives could be reconciled, is the authors took as a sign of the ‘fittingness’ or validity of their narrative process.