Pre-Medieval Landscapes in Medieval Court Romance

by A. B. Graves.

Abstract
Introduction
History and Arthurian Legend: Previous Approaches
Landscape and Memory in the Romances
'Perlesvaus' and the 'Quest'
The Landscape of the Grail
1: The Political Landscape
2: The Forest and the Religious Landscape
Conclusion
Bibliography
Acknowledgements

 Abstract

This essay examines the appearance in the early phases of Arthurian literature of landscape features recognisable to modern archaeologists as belonging to earlier periods, both historic and prehistoric. It argues that this is the case principally in those tales derived most closely from Welsh and Breton oral traditions, and postulates the use of a classical-style memory system as a means by which such features may have been preserved in oral tales. Finally it examines the underlying landscape of the tales themselves, with comments on revealed features (frequency and possible origin), comparisons to current interpretations of the past (particularly the Iron Age), and the potential of such 'recovered' literary material for the interpretation and reconstruction of past cultures.

 Introduction

The subjects of King Arthur and of the Grail really need little by way of introduction. They have been interpreted in ways which range from the historical (Ashe, 1968; Alcock, 1971) to the mystical (see any New Age bookshop), with all parties involved attracting fervent supporters and detractors. One way or another these stories have proved a source of speculation, entertainment, and annoyance for the better part of a thousand years. Such is the power of a good tale!

This piece presents a fresh approach to extracting information about the past from literary works derived from an extant oral or 'folklore' tradition. This is not a search for the origins of the tales (that bugbear of folklore studies), although in passing I will inevitably be mentioning some ideas on the subject. Rather I am here concerned with the scenery: the locations for events within the tales themselves; how, and with what internal consistency those locations are used; and finally, how to interpret this layer of information with regard to current achaeological speculations concerning pre-medieval societies and land usage. This paper is intended to demonstrate the potential of such studies for providing material to those archaeologists interested in the attempted reconstruction of past cultures.

I have worked mainly with two of the 'Grail' texts (though referring to others) in this study: the Perlesvaus ( anonymously authored), and The Quest of the Holy Grail (also anonymously authored but for many years wrongly attributed to Walter Map) (Matarasso, 1969: 25). Both are early thirteenth-century in date though the Perlesvaus is likely to be up to twenty years older than the Quest and is certainly a very different text with respect to my research aims outlined above.

 History and Arthurian Legend: Previous Approaches

The question of what, if anything, can be gained in any meaningful archaeological sense from the study of arthurian literature, is one which has been largely concentrated for the past thirty to forty years on the search for the ' historical Arthur'. Taking as a starting point the assumption that the tales refer in part at least to the activities of a real ruler of the post-Roman, pre-Saxon period in Britain, researchers such as Geoffrey Ashe and Leslie Alcock have attempted to match the texts with definite geographical locations: the most well-known being the tentative identification of South Cadbury near the Somerset-Dorset border as the site of the castle of Camelot (Ashe, 1968: 123- 148). Without wishing to diminish the efforts of these researchers (who have made great contributions to our understanding of fifth century Britain) I do find myself questioning their focus on an empirical historical authenticity for individuals, events and locations mentioned in this corpus of tales. Keith Thomas has commented on this type of activity as a historical phenomenon: 'This half-historic, half-mythical past was firmly anchored in popular consciousness by the widespread habit of attaching heroes and events to specific localities' (Thomas, 1983: 4). Very little written material from the fifth or sixth centuries has survived - if, indeed, much ever existed - and considering the social turmoil of post-Roman Britain, with repeated Saxon incursions and internal political fragmentation followed by complete collapse (Ashe, 1968: 33-39), this is not surprising. The monk Gildas, writing in the mid-sixth century, mentions the Battle of Mount Badon as a notable defeat for the Saxons but fails to provide the location or any names associated with the (briefly) victorious Britons (Alcock, 1973: 26). Badon (the location of which remains uncertain: see Ashe, 1968: 41) is one of the lynchpins of the arguments for a historical Arthur since this name was associated with the battle in 'Nennius'' History of the Britons (BM Harleian MS 3859): the version we possess being a twelfth-century copy of a tenth-century compilation of odd documents and annals (Alcock, 1973: 29). While Alcock questions the attribution of Christianity to Arthur in this text on linguistic and other grounds (Alcock, 1973: 50-52) he sees no reason for suspicion concerning the use of the name itself, asserting that 'no convincing case can be made for regarding the reference to Arthur at Badon as an interpolation' (Alcock, 1973: 53). I would suggest that given the proximity of dates between this manuscript (early 1100s) and Geoffrey of Monmouth's Historia Regum Brittaniae (c.1136), and the lack of documentary evidence for a historical Arthur prior to this period, there may be no convincing case but the circumstantial evidence leaves some grounds for suspicion. (But see note at the end of this piece).

One cannot study the Arthurian corpus without at some point running into Geoffrey of Monmouth, his Historia being the work which seems to have inspired so much later literary endeavour. In the context of the doubts I have already voiced concerning the origins of a historical Arthur, it is worth examining Geoffrey's motives in writing what he did. These were at least partly of a political nature: he wished to provide 'a precedent for the dominions and ambitions of the Norman kings' (Tatlock, 1950: 426 quoted in Thorpe, 1966: 10) and at the time, given the dearth of available historical material, he appears to have seen, and grasped, his opportunity to use the medium of history to this end. He was not the first, and certainly not the last, to make use of the past in such a manner: to this day references to past glories remain a standard method for legitimising or laying claim to political or spiritual authority. As Ann Williams puts it, in referring to a pre and post-conquest legal work (Quadripartus et Leges Henrici Primi): 'the " formidable authority of the royal majesty" rested on the accumulated prestige of the Old English past' (Williams, 1995: 158). Keith Thomas further points out that as recently as the seventeenth century 'the only respectable justification for the study of the past was that it could be of service to the present. The idea of history for history's sake was no more acceptable (or even intelligible) than that of art for art's sake' (Thomas, 1983: 1). The active redefinition of the past is such a common phenomenon that we might expect to see it here. Any surprise lies more in the fact that it took so long to occur in written form: the Historia was completed some seventy years after the Norman conquest was begun. Considerable sophistication is evinced by the Normans in that they did not simply replace many Anglo-Saxon institutions but suborned them to their own ends, as with the legal system mentioned above: the conquest was not completed by erasing what had gone before, but by laying claim to it. Geoffrey, himself thought to have been of Welsh or Breton descent (Thorpe, 1966: 13), appears to have played his part in this process by 'filling out' the gaps in the historical record with material from the culture into which he was born: a ready made and, perhaps, a personally satisfying source. Not the greatest of wordsmiths himself, but part of a culture which perpetuated itself largely through the medium of poetry, Geoffrey was in a position to appreciate that 'the moral or aesthetic value of such legends was at least as important as their historical veracity' (Thomas, 1983: 2). The material upon which he drew requires some attention in this investigation since the same stratum of (mostly at this time oral) tradition was to be drawn upon by the authors of the French romances in the wake of the Historia.

Although, as I have stated above, there are no definite references to a historical Arthur prior to the twelfth century, this is not to say that he and his court went unrecorded in the period of Anglo-Saxon domination in Britain. In fact the deeds of Arthur were put to writing by a near contemporary to Gildas not long after the Saxon incursions had devastated the remnants of Romano-British culture. The bard Taliesin, himself a part of that retreating culture, composing his works in the later sixth century for patrons (mostly in the North of England) who were still at war with the invaders (Ashe, 1968: 173), is thought to have composed the Preiddeu Annwfn or Harryings of Hades. In order to make some points, and as I shall later be referring back to it, I believe it is worth quoting a translation in full. Note that the name 'Arthur' is identical in the Welsh original.

(1) I adore the noble prince and high king
Who extended his sway over the world's strand.
Perfect was the captivity of Gwair in Caer Sidi,
Through the warning of Pwyll and Pryderi.
Before him no one entered into it,
Into the heavy dark chain a trusty youth guarded;
And at the harryings of Hades grievously he did sing,
And till doom will remain a bard afterwards.
Three freights of Prydwen went we into it -
Seven alone did we return from Caer Sidi.

(2) I am a seeker of praise, if my song be heard:
In Caer Pedryvan...from the cauldron it would be spoken
By the breath of nine maidens it would be kindled.
The head of Hades' cauldron - what is it like?
A rim it has, with pearls, round its border:
It boils not a cowards food: it would not be perjured.
The sword of Llwch Lleawc would be lifted to it,
And in the hand of Lleminawc was it left.
And before the door of Hell's gate lamps were burning
And when we accompanied Arthur, a brilliant effort,
Seven alone did we return from Caer Veddwit.

(3) I am a seeker of praise, my song being heard:
At Caer Pedryvan in Quick-door Island,
At dusk and in the blackness of night they mix
The sparkling wine, their drink before their retinue,
Three freights of Prydwen went we on sea:
Seven alone did we return from Caer Rigor.

(4) I merit not the laurel of the ruler of letters -
Beyond the Glass Fort they had not seen Arthur's valour.
Three score hundreds stood on the wall:
Hard it was found to converse with their sentinel.
Three freights of Prydwen were they that went with Arthur,
Seven alone did they return from Caer Goludd.

(5) I merit not the laurel of them of the long shields:
They know not which is the ruler's day or who he is,
At what hour of early day he was born or where,
Who made... went not...
They know not the Speckled Ox with the stout halter,
With seven score joints in his collar.
When we went with Arthur, anxious visit,
Seven alone did we return from Caer Vanddwy.

(6) I merit not the laurel of those of long...
They know not which is the day of the ruler and chief,
At what hour of early day was born the owner,
Or what myriad guards the silver of the head.
When we went with Arthur, anxious contest,
Seven alone did we return from Caer Ochren.

(Trans. J.Rhys, introduction to Malory, 1985: xxviii-xxx).

It is clear from the above that Arthur, as presented to a sixth-century audience still struggling with the Anglo-Saxons, is not being held up as an example of a destroyer of these particular enemies - as we might expect if a historical Arthur had indeed been responsible for the victory at Mount Badon and an ensuing 'generation of peace and relative order' (Ashe, 1968: 38). In fact Arthur is shown on an 'otherworld journey' in pursuit of a magical cauldron (quite possibly the prototype for the later grail: see Matarasso, 1969: 13), and his victories, if such they are, must be described as pyrrhic at best! Indeed his first port of call on this journey is Caer Sidi: a place quite distinctly set apart from the ordinary, everyday world and which crops up time and again in folklore as a gateway to, for want of a better term, 'elsewhere'. Another bardic poem of this period, a dialogue between the king and a gatekeeper reluctant to allow him access (Malory, 1985: xxiv- xxvii), has him name and describe his companions and their deeds. Alongside figures such as Kei and Bedwyr familiar from the later romances, we find such distinct mythological personages as Mabon son of Modron and Manawydan son of Llyr: also familiar but this time as protagonists in the tales of the Mabinogion. (Both can also be found in the French romances as minor figures: see Darrah, 1981: 72). Descriptions of their feats, while eulogising their skills in battle, make no mention of Saxons and mostly tell of conflict with wild beasts and supernatural entities: a hag, nine witches, and the mysterious 'Palug's Cat' (unfortunately undescribed).

Both Ashe and Alcock, in what I have seen of their work, make no mention of these lengthy 'Arthurian' poems in their brief discussions of bardic poetry; I suspect because the poems enmesh Arthur so firmly into a well-known mythological context so near in date to their hypothesised historical figure. Ashe is quite vehement in his opposition to a purely mythological role for the king; allowing for only one possibility he states:

Another theory - that Arthur was a Celtic god, the centre of a pagan patriotic revival - will not bear looking into. Admittedly the Arthur of legend has mythical attributes. He rides through the sky, he slays giants, he takes the form of a raven. But a Celtic deity with a Roman name seems unlikely. The extant records of Celtic religion show no trace of him' (Ashe, 1968: 39).

Well, the poetry discussed above might be considered 'extant records of Celtic religion' (without getting sidetracked into the contentious issue of Celtic ethnicity), as might the Mabinogion itself. By this I mean that they are mythological tales concerned with the relationship between deities and heroes, and the performance of supernatural feats. It should also be remembered that the culture in which these tales were current was not only the inheritor of 'Celtic' mythological lore but had come through three and a half centuries of intense romanisation, which was not abruptly cast off when Roman rule was withdrawn in the early fifth century:

Such myths as may have existed must have been coloured (in Gaul and Britain) by Roman thought....Many of the statues and reliefs from Gaul and Britain exhibit decidedly Roman traits, and often the name of a Roman deity accompanies (where it does not actually replace) the Celtic one (Gantz, 1976: 14).

Under these cicumstances is a Roman name really so unlikely? Finally, as regards Arthur's 'mythical attributes', I would suggest that, on this reading of the literary evidence, if he looks like myth and behaves like myth, then on the balance of probability he is myth: a part of the indigenous late Romano-British (or Romano-Celtic) mythos.

What part might such a figure have played in the myths of the time? The fleeting references in other bardic poetry thought to date from the sixth century attribute to Arthur qualities associated with bloody victory in battle. For example from Y Gododdin:

He glutted black ravens on the wall of the fort
although he was not Arthur (Alcock, 1973: 15).

Also note in this respect the pyrrhic victories implied in the 'Harryings of Hades', quoted above. If we accept that myth is, as with other aspects of a culture, subject to change or mutation in the face of external and internal pressures, then the emergence of a new myth associated with victory and slaughter (seemingly against the odds) tells us something of the perceived needs of the people in the society in which it develops. If it were the case that a series of 'Arthur myths' developed in the later phases of Romano-British culture they would represent not so much a ' pagan patriotic revival' (Ashe, 1968: 39) but rather the cultural response of a desperate people who saw themselves as very much in need of the qualities displayed in the myths themselves. For a description of this process in a more modern context see Deren, 1975: 65.

Moving on in the search for written antecedents to the romance tales, it is worth examining the collection of Welsh tales known as the Mabinogion. Here we come up at once against the problem of dating, since the bulk of this work, in the form of The Red Book of Hergest, turns out to be more recent than many of the romances; having been written down around 1400. Fragmentary earlier versions do exist, but the earliest of these is thought to date from the 1220s (Gantz, 1976: 29), contemporaneous with various romance tales including the 'Grail' material I have been examining. They do, however, provide the flavour of medieval 'Celtic' myth, and given the ethnic background of Geoffrey of Monmouth, he is likely to have been familiar with at least some of this material. Here, in contrast to the earlier bardic poems (of which the tales of the Mabinogion appear to be a development), we do find the figure of Arthur being treated as though he was a part of a properly 'historical' tradition. He is not the only one: 'characters like Rhiannon and Math (in the Four Branches), while exhibiting divine aspects and powers, are treated as human, even historical figures' (Gantz, 1976: 14). It is likely that in the face of twelfth-century claims for the historical authenticity of one individual from this mythos, the oral tradition in which the tales were maintained was adapted to fit in with the expectations of an audience: expectations quite different from those of Taleisin's audiences in the sixth century. The sad lack of, say, a tenth or eleventh-century text dealing with these themes makes this possibility hard to assess, but since these tales were part of an ongoing oral tradition, memorized in outline and with many details extemporised at the time of telling (Gantz, 1976: 13), an adaptation of this sort is a reasonable assumption. The Mabinogion as it stands cannot be considered a forerunner of the romances, but should be viewed as coming from the same stock of oral tales from which the romances unquestionably derive.

 Landscape and Memory in the Romances

Geoffrey of Monmouth's sober Latin prose was translated and adapted into French verse by the Jerseyman Wace, whose version, known as the Roman de Brut (after Brutus, the legendary founder of Britain) came out in 1155. Like all medieval 'translators', Wace altered and expanded his source as he saw fit, giving greater colour to the story and bringing it closer to the genre of the romance that Chretien de Troyes was to develop a few years later (Cable, 1971: 11).

'Within the next thirty years the Tristan romances, the lais of Marie de France, and the romances of Chretien de Troyes had all seen the light of day' (Matarasso, 1969: 11). The impact on the storytellers to the French nobility of the 'Historia', as transmitted by Wace, was immense, but should be seen in context. Available already as 'respectable' sources for court tales at the time were the 'Matter of France' (concerning Charlemagne, whose mystique was of great importance to the French throne) and the 'Matter of Rome' (the Roman and Greek classics: see Ashe, 1968: 1-2). Both sources came with certain constraints, either political or cultural, as to interpretation; in particular where humour is concerned. One can imagine the fate of an incautious storyteller who made a mockery of Charlemagne in front of the French royal court! With the arrival, courtesy of Monmouth, Wace and de Troyes, of the 'Matter of Britain', came an opportunity to explore and exploit courtly tales in new ways. The humour and salaciousness of much of the earlier Arthurian Romance material is something which has failed to make any great impression upon public perception (except as presented by the Monty Python team), but from the tales of the Welsh and Breton people was derived, for presentation to the nobility, a robust and often bawdy body of literature which must have entertained by shock as much as by novelty. In support of this view Jean Misrahi has pointed out that at least one contemporary of de Troyes (Jean Bodel: Prologue to Les Saines) maintained that 'the matter of Rome was wise and instructive, that of France was historically truthful, while the tales of Arthurian Britain were frivolous and pleasant' (Bryant, 1978: 3). The Church disapproved of and fulminated against these tales, providing as they did both backdrop and precedent for the tales of courtly love already becoming prevalent among the european aristocracy (Ashe, 1968: 8-9).

A fine example of the genre is Beroul's Tristan: the oldest version of this story to have survived from the twelfth century (Fedrick, 1971: 10-11). The translator in his introduction feels compelled to point out some of the oddities in this text: 'We are told countless times that Yseut is noble, wise and fair and that Tristan is noble, brave and strong; but a certain amount of explanation is needed, for it must be admitted that they do not always act in a way which is noticeably noble or wise' (Fedrick, 1970: 17). (Nor, I might add, is Tristan dauntlessly brave). Further: 'An unmistakable aspect of Beroul's artistry is his ability to create and sustain humour of all kinds, varying from broad farce to subtle irony' (Fedrick, 1970: 31).

The tale of Tristan and Yseut appears to be one of those deriving from the first flush of interest showed by the French storytellers in the Breton and Welsh oral material. It is firmly set in Cornwall and Brittany, and in the version of Beroul a few mentions of landscape features are indeed strikingly reminiscent of those areas: 'a chapel on a hill, built on a ledge of rock. It overlooked the sea, facing north' (Fedrick, 1970: 68), would still serve as an apt tourist description today. But what are we to make of other descriptions, seemingly of features of no obvious relevance to a medieval courtly audience? A hiding place, close by a well-travelled path: 'Here is a thick bush, surrounded by a ditch. Let us get inside' (my italics). The venue for a secret meeting: 'there is a marsh he knows at the end of a plank bridge at the Malpas…Let him stand on the mound at the end of the plank bridge on this side of the Blanche Lande'… (Fedrick, 1970: 69, 121). From the perspective of the audience such references doubtless provided a sense of local colour, rounding out the tale being heard, and may have been recognisable as ancient and mysterious features in the countryside. From the current perspective of landscape archaeology we are able to discern in these locations elements of medieval - and earlier - land usage. In the first case the hiding place appears to be a circular ditched enclosure, small but unused for agriculture though in a well frequented and otherwise (by implication) cleared area. Perhaps therefore this is a place subject to taboos concerning usage: the memory of a sacred space long since fallen into disuse. In the second case the mound associated with the plank bridge and the marsh is more compellingly suggestive of such a space (Flag Fen near Peterborough springs particularly to mind), but here the description reads as though the place were still in use. This is not to suggest that plank bridges were not still being made and used in the twelfth century, but those three elements, together with an evident function as a marker in land division (the Blanche Lande), might preserve memories of a sacred or special area which would not have been out of place in the Bronze Age or even earlier. There is no question of being able to identify an actual location for these features from the descriptions in the text, and they are by no means to be found only in the south-west of Britain. A more appropriate question would be to ask what they are doing in Beroul's text at all?

I must here aknowledge a debt to John Darrah, who in one of his appendices first drew my attention to the archaeological potential of the romances. He cites a passage from the Norse Tristramssaga (1226), a translation of a now lost French version of the tale:

They found a secret place beside a certain water, and in the hillside, that heathen men let hew and adorn in olden time with mickle skill and fair craft, and this was all vaulted and the entrance digged deep in the ground, and there was a secret way in running along below ground. Over the house lay much earth and thereon stood the fairest tree upon the hillside (Darrah, 1981: 145).

Darrah goes on to list four points of correspondance between this description and the fogou at St Euny, Cornwall (noted by S. C. Harris) which are as follows:

(Darrah,1981: 145).

It is possible (though I think no more than that) that this is an accurate interpretation of the passage quoted, however the same description could equally well be applied to any number of ancient monuments to be found in the south-west of Britain or for that matter elsewhere: long barrows and chambered cairns to name but two varieties. What is incontrovertable is that the description is to be found in an early thirteenth-century romance text, that it clearly refers to an ancient monument of some sort (built by 'heathen men...in olden time') and that this monument appears to be (though well enough known to the original audience to be described in recognisable terms) a proscribed place where fugitives can hide in relative safety from their pursuers. In other words a place subject to taboos on access and use, as I have suggested for the 'hiding place' in Beroul's Tristan.

What, then, are references to features such as these doing in the courtly romances of the twelfth to thirteenth centuries? The composers of the romances were not noticeably interested in the landscape per se, and the concept of landscape as we use it today is a relatively recent one. Where it is used at all in the romances, landscape description tends to be minimal and is used principally as a background to the far more important events of the story. Landscape features may have been used as a way of providing a little 'local colour' and a dash of mystery. At times the surrounding landscape reflects the type of event which is taking place or about to occur, which can be taken to be a dramatic device of the storyteller. A 'bad' landscape (wild and overgrown, or dark and shadowy) almost invariably introduces a 'bad' event (often dire combat), while 'good' landscapes (fair meadows, cultivated lands, open and sunny valleys) tend to introduce 'good' events (safe lodgings and/or an explanation of recent events to the protagonist from a monk, nun, or hermit). This remains a standard part of storytelling today (think of any action or horror movie) and has written antecedents going back to ancient Greece: for example the rocky and mountainous island of the cyclops, arrived at 'through the murky night' (Rieu, 1946: 143) by Odysseus and his companions. For the purposes of this study I have winnowed out such references where they are otherwise undescriptive of landscape features, leaving a variety of types of location within a landscape where types of event may occur within a tale. I have alluded already to the technique of memorisation thought to have been used in the storytelling tradition from which the Mabinogion was drawn: extemporising details while memorising the basic pattern of a tale. This seems an appropriate juncture at which to examine memory techniques which might have been used in fixing those 'basic patterns' in the minds of the storytellers.

It has been surmised from learning experiments with chimpanzees that memory is a mental function which operates through association (Deacon, 1998: 88-89), but putting aside the language of science and the use of valuable chimpanzee time, this is hardly a new conclusion. Associative methods of training the human memory, reliant upon visual imagery, are known to us from the classical world and have been intensively studied by Frances Yates, who notes that 'in the ages before printing a trained memory was vitally important; and the manipulation of images in memory must always to some extent involve the psyche as a whole' (Yates, 1992: 11). These systems, or variants thereof, are still in use by 'memory masters' today and are regularly advertised as 'new' methods in the press for those who feel a better memory might improve their lives or careers. Yates accepts the assertion of Cicero that the principles of the 'Art of Memory' were invented by the poet Simonides of Ceos (c.556-468 BC). His development of the technique using visual imagery derived from architecture formed the basis of the memory techniques used in classical oratory and is described in texts such as the Ad Herennium, a guide to rhetoric thought to date from 86-82 BC (and still in use in the Middle Ages), where it is considered as 'artificial', as opposed to 'natural' memory (Yates, 1992: 20-32). There is no need here to describe the methods in great detail, save to draw attention to the fact that they were used by orators, not to memorise speeches word by word, but to memorise themes or events to be extemporised, by juxtaposing striking or humorous visual imagery with a series of locations on an imaginary journey through a building or public space.

What I find significant here is that Simonides, the reputed inventor of the memory systems described in the Ad Herennium, was himself a poet: an inheritor of the oral tradition of the Greek epics. It takes no great stretch of the imagination to suggest that he may also have inherited the mnemotechnic principles by which (say) the Homeric Epics were passed down through many illiterate generations. Where he perhaps parted company with his predecessors was in applying these techniques outside of the realm of poetry and storytelling, and in making them available to the intelligensia of the time. If we accept this as a possibility, then it could follow that a similar set of techniques for extemporising along set formats in public performance was at some time prevalent in non-literate cultures across Europe: as we have seen, the Welsh and Breton bards of the twelfth-century are thought to have still been doing much the same in the continuance of their localised oral traditions. This is, I must concede, a lot of 'possibles', and any proof of something as transient as a memory system, particularly where so much oral material has been irretrievably lost, must inevitably be by inference. Nonetheless the landscape and architecture of a memory system is the best explanation I have been able to find for the prevalence of archaic features in the (orally derived) early phases of Arthurian court romance. To emphasise this possibility, consider again the Preiddeu Annwfn quoted above, which fits rather neatly with the type of 'artificial' memory examined by Yates: a journey, punctuated by loci (the forts and island), upon each of which is imprinted an image and event. Caer Sidi is associated with the liberation of a captive - a familiar motif for the place in British and Irish folktales; Caer Veddwit is linked with the enchanted cauldron; the Glass Fort with a negotiation made difficult by what appear to be differences in language or dialect. In the context of the survival of archaic features through a memory system, this latter example is interesting in that, if it refers to anything derived from the real world (as opposed to the purely fantastic in myth) it must be to one of the vitrified forts of Scotland. These structures were coming to the end of their useful life during Romano-British times: 'although the various dating techniques produce somewhat contrary results, construction and use extend at least through the Iron Age into the early historic period' (Edwards/Ralston, 1997: 178), but they were still apparently part of the bard Taliesin's mental repertoire in the late sixth century. Since the 'early historic period' can be taken to include the sixth century this is not unlikely, but they were by this time archaic, outmoded (if still striking) structures, and the inclusion of one of them in this poem provides some insight into the way in which the elements of past landscapes may be transformed into the elements of current tales. In the oral tradition of which Taliesin was a part they became one of the mental loci upon which were imprinted the events and descriptions of stories being told, retold, and reinvented through extemporisation. On the evidence of the early medieval Arthurian material I would argue that the 'glass forts' are not alone in receiving this treatment.

 'Perlesvaus' and the 'Quest'

The author of the Perlesvaus was one of several medieval writers who attempted to complete the 'Romance of Perceval' left unfinished at his death by Chretien de Troyes (Bryant, 1978: 2-3), and his prose is very much in the racy, burlesque style I have already noted as being the norm for the earlier orally-derived Arthurian romances such as Beroul's Tristan. His treatment of the 'Matter of Britain' is lighthearted and his intention is to entertain, with novel oddities, humour, and irony as part of his repertoire. One of Gawain's adventures on the quest makes this abundantly clear: Gawain being portrayed in the romances as the courtliest and lustiest of the knights, he spends much of this quest trying hard to behave and to live down his reputation:

In the middle of the forest he came across a wide clearing through which a stream flowed from a spring, and looking towards the head of the clearing, right at the edge of the woods, he could see a great tent... [Gawain enters, and is fed and made comfortable before meeting two maidens]… "Sire", they said, "may God give you the power and strength tomorrow to destroy the evil custom of this tent". "Is there then an evil custom, damsel?" said Gawain. "Yes, sire, and ugly it is indeed. It grieves us much, but you seem to be a knight fine enough to put it right"… [It transpires that Gawain is expected to spend the night with one of the maidens; an offer he - for once - turns down, much to their annoyance]... "In faith", said one, "if this were Gawain, the nephew of King Arthur, he would speak to us differently, and we should find in him more entertainment than in this man: this Gawain is an imposter"... [After a warily sleepless night Gawain finds he is expected to fight two knights for "the food you ate and the honour you were paid". The maidens watch, commenting on the fight as he spears his opponents]... "By my life", said the elder maiden, "the false Sir Gawain is doing better than he did last night!"... and the dwarf cried: "Damsels, your Gawain is fighting well!" "He will indeed be our Gawain henceforth", they said, "if he is willing"... (Bryant, 1978: 62-65) .

While not strictly speaking representative of the whole text this abridged excerpt shows the tone of some of the humour and, more importantly for my purposes, provides a fine example of a forest glade, complete with spring and an associated 'evil' custom. I will return to these themes later.

By contrast the author of 'The Quest of the Holy Grail' has no truck with episodes such as the one described above; in fact this incident, along with much else, has been completely expunged from the tale. From the outset it is clear that the Quest has been composed to fit a very specific agenda: that of moral instruction. Knights who have led sinful lives in the past either meet with no adventures at all (like the irredeemable Gawain) or are mercilessly harried and mocked by the forces of righteousness until they fall by the wayside or (like the just barely redeemable Lancelot) acknowledge the error of their ways and become contrite penitents. Descriptions of the locations where events occur are, in contrast to the Perlesvaus, sparse and in many cases non-existent; where they appear at all they tend to fit the dramatic convention of bad land equals evil event (and vice versa), or may represent the spiritual state of the current protagonist. Here we see Lancelot in penitent mode, as an example to be followed:

He spent that night on a wild and towering crag, with no other company save that of god, dividing the hours of darkness between prayer and sleep. When he saw dawn break on the morrow he made the sign of the cross on his forehead and, facing east, prostrated himself on knees and elbows and prayed as was now his wont (Matarasso, 1969: 160).

Matarasso, in his introduction to the Quest, draws attention to the author's opposition to the courtly ideals of the time, showing how the characters' 'much-vaunted attributes lead them to the outcome one would least have looked for; the last are perhaps not first, but he makes no bones about showing how the first are last' (Matarasso, 1969: 15). He also supports the conclusions of Etienne Gilson, 'author of an illuminating article on the Cistercian inspiration of the Quest which he sees as illustrating the Bernardine doctrine of grace and mystical union with God' (Matarasso, 1969: 15). Certainly where monks and hermits are encountered in this text they are almost invariably 'white monks', in contrast to the Perlesvaus where, if they are given any affiliation at all, it is to St. Augustine. Indeed in many respects the general background of the Perlesvaus would seem to be the missionary efforts of Augustine's followers, establishing themselves at sites of importance to pagans: wells and springs in particular. By usurping this position for the Cistercians (founded long after these events in 1098) the author of the Quest provides a false sense of antiquity, and hence authority, for the order. With his dual agenda of moral instruction and the promotion of Cistercian ideals, the author has no time for the novel entertainments offered in the Perlesvaus, and I suspect had no interest in using the body of oral tales in which the author of that text found so many of those novelties. The Quest is, by comparison, a dry and long-winded tale, relatively bare of landscape features.

 The Landscape of the Grail

In presenting this study it would have been simple enough to produce an exhaustive list of landscape elements from both of the main texts being studied. However many such references are fleeting, especially in the Quest, and a list of single words taken out of context would be of little use except as statistics. This being a preliminary investigation I have opted instead to look at types of feature, from the broad scale of the countryside as a whole to specific features within it, with examples of each and comments on frequency and interpretative possibilities.

 1: The Political Landscape

The grail quest is set firmly in the land of Britain, but this is by no means the political entity with which we are familiar today; nor is it the political entity with which a medieval audience might have been familiar. It is a land in a constant state of flux, with a seemingly endless supply of tiny kingdoms and petty rulers. This is explained as being 'because of the strange adventures which God had brought to pass... they came to a forest or an isle where they encountered an adventure, and returned some time later and found different fortresses and castles and adventures...' (Bryant, 1978: 181). This might be no more than a literary device to explain the miraculous events of the quest, but it might also be intended to explain the curious (to a medieval audience) mobility of entire peoples and 'countries' in the background to the tales. There is no sign in the grail quest of the Saxon invasions of the Mort d'Arthur, but the constant feuding with neighbours and shifting of alliances seems to keep people on the move and changing their centres of power. The culture being described certainly thrives (or rather labours) under the concept of the blood feud, as Perceval makes clear:

The knight recognised the horse that Perceval was riding, and said: "Sir knight, that horse belonged to My Lord the Red Knight of the Deep Forest. Now at last I know that it was you who killed him". "That may be so", said Perceval, "but it was only right that I killed him, for he had beheaded my uncle's son, whose head this maiden had been carrying for a long while". "By my life", cried the knight, "since you killed him, you are my mortal enemy!"' (Bryant, 1978: 237).

The maiden with the head is one of many who crop up throughout the romances, and Darrah has argued for this to be a survival of the late Iron Age head cult (Darrah, 1981: 14, 39).

What of these 'different fortresses and castles'? A knight can rarely ride for more than a day or two, even in the darkest, wildest parts of the forest, without encountering one, with or without occupants or associated cleared land. In part they appear to serve as places used to protect cattle from those feuding neighbours: 'Brien of the Isles has Sir Kay the seneschal in his company and is burning the king's land and stealing the cattle outside his castles' ( Bryant, 1978: 193). Some castles are described as having streets, making them sound more like settlements than a single fortified structure (Matarasso, 1969: 74, 239). Many of the castles are to be found in forest clearings: 'he debouched in a vast clearing in the forest and saw ahead of him a strong and well-positioned castle, beset with walls and ditches' (Matarasso, 1969: 155-6). (Note the multiple walls and ditches). The occupants of at least one of these forest castles have taken it upon themselves to guard 'the forest and its paths' (Bryant, 1978: 55), while other strongholds are used by robber knights (Bryant, 1978: 131). Many of these places, like certain individuals, have strange customs associated with them, which range from the innocuous never asking the names of visitors (Bryant, 1978: 55) to the more familiar combat outside the gates - complete, in one case, with the ritual shaving of the loser ( apparently to make hair shirts for the hermits of the forest) (Bryant, 1978: 86). Better physical descriptions are forthcoming with some of the larger non-forest structures:

...he passed out of the forest and saw before him a most rich and pleasant land bounded by a huge wall which stretched a great way. He rode on and found that there was but one gate; and when he passed through he saw the fairest and most abundant land that ever a man beheld, with the most beautiful orchards. The land was no more than three leagues wide, but in the middle stood a great tower upon a rock, and at the top of the tower a crane had built its nest, and cried out whenever a stranger entered the land. (Bryant, 1978: 67).

The astonishing concept of an entire (if small) walled country, with restricted access and a central watchtower, calls to mind the extensive earthwork boundary features found in association with a number of hillforts in southwest England and the wall and ditched features in the Yorkshire Wolds. Notably, a neighbouring country is described in almost identical terms, save that it is desolate and uninhabited (Bryant, 1978: 69). This indicates similar landscape features in the same area, simultaneously in and out of use: supporting the implication of a mobile population. Here and elsewhere in this text, the extent of rulership appears to be limited by the distance the sound of a horn or other signalling device will carry: one mighty horn is said to have a range of ten leagues! (Matarasso, 1969: 74). 'At that time there was no bell in Britain or in Brittany - people were summoned by a horn instead, or in some places with a steel drum, and in others with wooden clappers' (Bryant, 1978: 178). This may account for oddities such as the Copper Tower. Treated as a god by the inhabitants it '…was in the middle of the castle, standing on four columns of copper, and at all hours of the day it let out such terrible roars that it could be heard for a league around; and there was an evil spirit within...' (Bryant, 1978: 164). A high prestige, and highly audible symbol of rulership, but not one for which I am aware of any historical or archaeological precedent. If such a device ever existed, perhaps producing an eerie sound in the same manner as wind over a bottletop or used as a sound amplifier, it is sadly unlikely to be proven by excavation.

The most complete descriptions of fortresses are of those containing spiritual treasures, as we might hope in tales concerning the quest for one such. These places have certain features almost invariably in common: they are surrounded by one or more 'rivers', they are difficult of access due to multiple narrow bridges, and they are associated with a nearby ancient tomb. At Kamaalot, birthplace of the grail knight:

...he came across a magnificent tomb with a beautiful lid, which seemed to be quite near the castle; there appeared to be a little cemetery there, for it was fenced all around, and yet there seemed to be no other tombs...as he approached the castle entrance he saw that there were three long and terrible bridges to cross, with three great rivers flowing beneath (Bryant, 1978: 76).

At the Forbidden Castle: 'Then they looked before them and saw a castle standing in the middle of the open meadows, surrounded by rushing rivers and rings of wall' (Bryant, 1978: 159). And at the Castle of the Grail itself: 'two bow- shots from the bridge was a chapel just like the one at Kamaalot, where there was a tomb and no-one knew what lay within' (Bryant, 1978: 169). This last fortress has no fewer than nine bridges over rivers, a number of entrances suggestive of nothing so much as a causewayed enclosure. If we consider the multiple walls and ditches (rather than rivers), and the proximity of ancient tombs in these descriptions, we can only be reminded once more of hillforts - also at times associated with causewayed enclosures. Inferring from these sources we can get some impression of the very special supernatural prestige enjoyed by the association of such structures with more ancient features. It is not that this is a new idea; it is of course a part of current landscape interpretation, but to find it in a thirteenth-century text derived from oral sources is a little startling, and suggests that these elements have made a powerful imprint on memory.

Finally, it is worth noting the abandoned castles encountered in the quest. Or rather not quite abandoned; for although in the course of the quest the knights come across a number of castles and 'big houses' which they at first take for empty ruins, these always turn out to have occupants. The lands round about may be empty and uninhabited, or in one case seemingly in the process of abandonment (the Poor Castle: see Bryant, 1978: 81-82), but at least one person (or whole family) will be left behind: 'Then he and Gawain rode on until they came to an old and ruined castle standing in a forest; it was a most beautiful castle, or would have been if it had been lived in, but there was no-one there save an old priest and his clerk who lived by their own labour' (Bryant, 1978: 197). Other occupants of ruins in desolate areas include a woman forced to stay as part of a religious penance (with the unpleasant task of collecting the bodies of those killed in the forest) (Bryant, 1978: 176-177), and a giant (or supernatural entity) ( Bryant, 1978: 69). This again can be taken as evidence for the mobility of the population, and it is tempting to infer a religious element to the selection of those who remained behind (perhaps) as 'territorial caretakers'.

Cities are few and far between in the grail texts, but unlike towns, villages, or farmsteads (unless, as is possible, the term 'castle' actually covers these three) at least there are one or two. The burning city is in use and awaiting the appointment of an annual king (Bryant, 1978: 106-107), being set in open farmland and with no indication of fortifications. The Waste City (Bryant, 1978: 90, 182) is fortified though crumbling, and appears abandoned; though the remaining few inhabitants also await an annual king. The description of this city, in two visits by Lancelot, is strikingly similar to the Anglo-Saxon description of an abandoned Roman city in a poem called ' The Ruin' thought to refer to Bath: 'The city was so huge that it seemed to fill an entire country...he could see its walls crumbling round about, and the gates leaning with age...its great palaces derelict and waste, its markets and exchanges empty, its vast graveyards full of tombs...' (Bryant, 1978: 90).

Splendid this rampart is, though fate destroyed it,
The city buildings fell apart, the works
Of giants crumble. Tumbled are the towers,
Ruined the roofs, and broken the barred gate,
(Hamer, 1970: 27).

Perhaps these similarities are inevitable; a ruined city can only be described in so many ways, but possible Anglo- Saxon influence can be detected elsewhere (in the Perlesvaus at any rate) in the tales associated with the Augustinian hermits, which may relate in part to Dark Age hagiography. The scale of the city is here worthy of note; 'it seemed to fill an entire country' suggests that it cannot be considered one of the enclosed townships implied in some of the 'castle' descriptions cited above, and further emphasises the small size of countries in the romances. Whether of Anglo-Saxon or Welsh/Breton provenance, a memory is undoubtedly preserved here of the post-Roman decline and abandonment of many large centres of population.

 The Forest and the Religious Landscape

Since the bulk of the adventures of the Grail take place in the setting of the 'Perilous Forest', and since this forest seems to be inhabited almost entirely by religious recluses wherever it is not punctuated by castles, it makes sense to deal with the two together. The forest itself is divided into a number of parts, mostly left unnamed. Those named areas which appear include the Waste Forest (Matarasso, 1969: 80, 207), the Deep Forest (Bryant, 1978: 233), and the Wild Forest (Bryant, 1978: 259): progressively more dangerous as the protagonists proceed and suggestive of a wide nomenclature for types of woodland. They are used in the tales largely to emphasise the darkness and mystery of what lies within: places apparently set apart from everyday human use. This is a misleading impression however as almost every forest encountered is riddled with well maintained trackways; the Waste Forest even has a wide road passing through it (Matarasso, 1969: 207). Linked by the trackways are innumerable open glades, many of which seem to have been developed around natural springs. The main paths are on the whole safe to travel: few life-threatening incidents occur upon them and some are patrolled by local warriors (Bryant, 1978: 55). The places of danger (or, if a hermit or friendly castle is present, succour) are the glades, in some cases fenced in and with access restricted by a gate:

…he came at tierce to one of the most beautiful glades ever seen, with a swing gate at the entrance...[he is then told by a maiden that]... the glade and the forest around it are so perilous that any knight who enters returns either dead or maimed. [After the inevitable combat and a narrow escape the protagonist is told:] Sire, you may dismount now. You are safe this side of the gate (Bryant, 1978: 25-30).

Combat is such a regular event in the forest glades that Darrah may well be correct in attributing this to a pre- Roman religious practice and, again, the severed head cult (the incident at the glade above includes a beheading) whether or not one accepts his heavily Frazerian interpretation (Darrah, 1981: 15, 26). What strikes me is that in the broad perspective of the forest as a whole we find some glades occupied by communities (which may or may not welcome strangers and which may be housed in 'pavilions' rather than castles); some glades, usually with springs, occupied by religious devotees (not all of them Christian if the maidens are anything to go by); and some simply described as 'beautiful' or 'ugly'. These latter terms as applied to the countryside in the romances often simply indicate productive or unproductive areas: 'Gawain passed through the ugly forest and entered the beautiful, vast and tall forest, full of wild creatures' (Bryant, 1978: 41). Thus the 'beautiful' glades appear to represent areas of food production, presumably for nearby forest-dwelling communities. Bearing in mind the cattle stealing and pillaging of farmland already mentioned and the implied political situation, these are likely to have been guarded at least for some of the time. Whatever the reason for combat in glades, it appears to have been common enough at some stage to have become firmly fixed as a motif in the memories through which the oral tales were passed down, eventually to become one of the novelties of the romances.

I have already mentioned the hermits of the forest in the context of the mission of St Augustine, which began according to Bede in 597 AD (Sherley-Price, 1968: 68), and this saint is the only one I have found mentioned in connection with religious houses in either the Perlesvaus or the Quest. Those who came with and after him were (if hagiography is to be believed) known for their austere, often solitary lifestyles and for the christianising of pagan sites.

The growing compendium of churches known or believed to be sited on or next to prehistoric barrows is suggestive of the intentional assimilation of these monuments by the Christian church, e.g. Fimber, Taplow, Ludlow, Bampton (OXF) and Yatesbury (WLT). The Life of Guthlac describes the battle of the saint with the demons in the burial mound and his success. He drives them from the place and then constructs his monastery on the site (Semple, 1998: 120).

I have been unable to locate any clear references to burial mounds in the grail stories examined here, although there are plenty of examples of chapels on hills (e.g. 'a knoll crowned by a hermitage') (Matarasso, 1969: 86) which may in some cases amount to the same thing. The hermits of the Perilous Forest seem particularly concerned with establishing themselves at springs and wells: places already of religious import to pagans. Significantly some of these remote chapels and hermitages are themselves haunted by evil spirits, connected with both springs and tombs: ' Galahad waited no more but set out towards the tomb, and as he drew near he heard a rending shriek as of a being in torment, and a voice...he saw smoke and flame belch out, followed at once by a thing most foul and hideous, shaped like a man' (Matarasso, 1969: 62). One must assume that the abbey at which this incident is set, though also connected with a mysterious tomb, was not the one founded by St. Guthlac mentioned above.

Most descriptions of religious retreats are as slight as those of secular structures: 'he came to the foot of a mountain, where he found a hermitage beside a spring' (Bryant, 1978: 110) is a typical example. A few give an idea of dimensions: 'he came to a hermit's cell, a hut so low that a horse could not enter, and the chapel was no bigger. The good man had not come out of his house for at least forty years' (Bryant, 1978: 75). One fine description is worthy of mention however:

…on the mountain he saw a newly built chapel, which was handsome indeed and very rich; it was roofed with lead, and on top were two crosses which seemed to be made of gold. Beside the chapel were three houses, most pleasantly situated: each stood on its own but was joined to the chapel. And round about was a most beautiful cemetery, surrounded by the forest on all sides; and a clear spring flowed down from the forest heights and ran before the chapel and then tumbled down into the valley. Each of the houses had a big orchard, and each had its own enclosure ( Bryant, 1978: 204).

This is the religious community at the Isle of Avalon, and though small it is clearly presented as being wealthy: enclosed or walled orchards are usually associated with the finer castles in the grail romances. It would be interesting to see if any excavations of early Christian communities in Britain (whether Augustinian or not) have revealed the sort of layout described in this passage, and what sort of enclosures (if any) have been interpreted as orchards.

Finally I should like to look at cemeteries. These are generally found in association with chapels, and surrounded by a distinct boundary: 'Just where the forest thinned out, he saw a tall cross at the entrance to a cemetery, which was bounded by hedges and hawthorn bushes, and the road led right through the graveyard' (Bryant, 1978: 88). The prevalence of bushes is reminiscent of Tristan's hiding place (Fedrick, 1970: 69), and might allow further identification of this small enclosure specifically as an ancient feature. We might also infer from the above passage the deliberate slighting of an earlier site (by the road being driven through) before its reconsecration to Christian use. The boundaries vary; a striking example relating to (presumably) non-Christian burial practices appears in the Quest: 'His journeyings, which followed no set course, brought him one day to the tombs which stand within their palisade of swords' (Matarasso, 1969: 268). Unusually there is no mention of a chapel connected to these particular tombs, while elsewhere chapels have clearly been built on the sites of much older cemeteries: 'he ordered that his body and the bodies of the other knights be taken to an ancient burial-pit beside an old chapel in the forest' (Bryant, 1978: 152). In this latter case a wicked ruler, the Lord of the Fens, and his followers are to be buried. They have not died in battle but have been captured and executed in a particularly bloody manner on the orders of Perceval, and in death they are treated as criminals. Sarah Semple has outlined the changes in use of ancient burial sites during Anglo-Saxon times, showing that as churchyards developed so the ancient sites became used specifically for the burial of criminals: 'The pagan cemetery and the barrow at a later period become the focus for a criminal cemetery, when the place is then perceived as a heathen and perhaps haunted place' (Semple, 1998: 118). In the incident described above we see exactly this occurring, and this can also be used as an example of the method of use of the memory system by which (I suspect) these tales were originally passed on. The loci upon which events are impressed in memory remain the same (the burial site), while events (the burial) are tailored to fit the expectations of a contemporary audience: the art of extemporisation in action. Immersed in the Christian ethos the loci themselves - at least those connected in Christian thought with religious activity - are adapted for the audience: hence, perhaps, the inclusion of the 'old' chapel. It is quite possible that this was by the twelfth century a permanent change in the mental imagery used by the Welsh and Breton storytellers where such loci were still recognised as being of religious import.

 Conclusion

In researching this piece and developing and arguing the case for the survival of pre-medieval landscape features in the early Arthurian romances, I have had to cover a great deal of ground. The unattested use of a classical-style memory system in european preliterate cultures remains a worry, and if used by bards in the sixth century may derive instead from the romanisation of the areas in question. This should not invalidate my observations since memory training of some sort was certainly a feature of Iron Age 'Celtic' culture (see for example Caesar's comments on the Druids in Handford, 1982: 141) and a conflation of the two, even if the techniques differed, could still be expected to preserve earlier imagery. To a great extent studying the texts themselves has been an exercise in reading between the lines and developing expanded interpretations from (often) tiny scraps of evidence; a risky business but comparable to the activities of landscape archaeologists in, say, walkover surveys. There is always the danger of overstepping the mark. A major problem, aside from being unable to dig for proof, has been in disentangling dramatic literary devices and the symbolism of landscape elements (to the original target audience of the romances or oral tales) from the features themselves. Further, care must always be taken in dealing with translated texts, let alone ( as here) with translations of translations: I am not a scholar of ancient Welsh or medieval French and Latin. My own biases inevitably show through; thus while I suspect there are more elements of Anglo-Saxon and Romano-British landscapes hidden away in the texts (such as the ruined city), I have concentrated on identifying what appear to be prehistoric elements; that being my principal area of study in archaeology.

As for the prehistoric features themselves, they do form a coherent pattern which seems to relate principally to the Iron Age. Hillforts, and the implied chaotic political situation (perhaps driven in part by the blood feud and the warrior ethic) certainly fit into current interpretations of that period, as do the 'walled countries' and the prestigious associations with older monumental structures. A major surprise was in finding so many implications of a mobile population, since the Iron Age is seen as the time when permanent, fixed settlements and territories became the norm. This may be taken to refer back to an earlier period still, but the implication is drawn from the descriptions associated with the use of specifically Iron Age features; again, the hillforts and associated linear features. Perhaps this can be taken to support the idea that permanent, fixed settlement was incompatible with the methods of cereal agriculture in use when it was first tried - hence the numerous 'desolate' or 'waste' countries: still claimed by particular groups but no longer productive having been farmed to exhaustion. It is possible there is a connection with the early Iron Age climatic 'hiccup' which contributed, with rising population, to the abandonment of much previously fertile agricultural land. Indeed some combination of these factors might have produced the sort of cultural trauma in the early Iron Age which lingers long in the images and motifs through which memories are transmitted. The implied religious nature of the 'caretakers' left at the associated abandoned castles might indicate more than just a territorial imperative: left in these locations do these people represent an attempt to regenerate these impoverished areas by the maintenance of ritual, in the hope of making them once more fertile for use by the larger community (now) elsewhere? It is a striking image of the problems associated with the early phases of settled communities which seems to have survived as the backdrop to a body of tales. Or perhaps fixed settlement itself began a little closer to historic times than is currently thought, and what has been taken as evidence for early permanent settlement should be considered semi-permanent at best.

Descriptions of the forest itself produce the impression of a network of trackways connecting cleared areas associated with settlement (permanent or not), food production, and religious practice (mainly connected with springs). This image, while still being quite recognisable in medieval times, is I suspect the most ancient stratum of background material to be translated from the oral tradition into the romances. It could certainly, on our present understanding, fit in with cultures in Britain at least as far back as the Neolithic. Though the actual myths of that time are long lost the setting for some of them may have survived, reinterpreted and reused with new myths in the memories of storytellers over many centuries.

Archaeological interpretation is itself a form of storytelling. We are engaged in creating modern myths around the physical remains of past cultures through logic and intuition, much as past generations reinterpreted the past in their own terms: 'Even the most scrupulous historian is busy myth-making' (Thomas, 1983: 24). Where I think a study such as this one is useful is in fleshing out our own myths with those of the past by finding points of correspondance between the two; perhaps creating room for the development of fresh interpretations. After completing this study I have no doubt that a great deal more material relevant to this purpose could be extracted from the early phases of Arthurian literature, and probably from many other literary works where these are derived from a pre-existing body of oral lore: ancient bardic material in particular could bear fresh examination from this standpoint. The use of memory systems provides us with a rationale for the survival of visual landscape imagery in storytelling, and the way in which that imagery is used in a tale gives us clues as to its import; both at the time of telling and, by inference, in earlier periods. How such information is used (if at all) will inevitably remain a personal choice according to the particular biases or interests of the interpreter: this is in the nature of the beast and I am as culpable as the next person in this respect. I should emphasise, if I have not already made it clear, that this study and any others like it can only produce possibilities. If these possibilities support current ideas they are still not necessarily correct, and the converse must also apply.

Note. Close to the date of publication I came across the following in a verse attributed to Taleisin:

As erst in Badon's fight, -
With Arthur of liberal ones
The head, with long red blades;
(Matthews, 1998: 86).

So it appears there is a sixth century source connecting Arthur with Badon and this obviously enhances the historical viewpoint with respect to this character. It also emphasises the desirability of further research into bardic material.

Bibliography

Alcock, L: 1971: Arthur's Britain: Harmondsworth, Penguin.

Arnold, C. J / Davies, J. L: 2000: Roman and Early Medieval Wales: Stroud, Sutton.

Arthur, R. G: 1996: Three Arthurian Romances: London, Orion.

Ashe, G: 1968: The Quest for Arthur's Britain: Aylesbury, Watson and Viney.

Bryant, N: 1978: The High Book of the Grail: Cambridge, Drewer.

Darrah, J: 1981: The real Camelot: London, Thames and Hudson.

Deacon, T: 1997: The Symbolic Species: Harmondsworth, Penguin.

Deren, M: 1975: The Voodoo Gods: St. Albans, Paladin.

Edwards, K. J, Ralston, I. B. M: 1997: Scotland, Environment and Archaeology. 8000BC-AD 1000: Chichester, Wiley and Sons.

Fedrick, A. S: 1970: The Romance of Tristan: Harmondsworth, Penguin.

Gantz, J: 1976: The Mabinogion: Harmondsworth, Penguin.

Hamer, R: 1970: A Choice of Anglo-Saxon Verse: London, Faber and Faber.

Handford, S. A: 1982: Caesar. The Conquest of Gaul: Harmondsworth, Penguin.

Hogg, A, H, A: 1984: A Guide to the Hill-Forts of Britain: London, Paladin.

Malory, T: 1985: Le Morte D'Arthur: Ware, Omega.

Matarasso, P, M: 1969: The Quest of the Holy Grail: Harmondsworth, Penguin.

Matthews, J: 1998: The Bardic Source Book: London, Blandford.

Rieu, E. V: 1946: The Odyssey: Harmondsworth, Penguin.

Semple, S: 1998: 'Fear of the Past' in World Archaeology. Vol.30 (1).

Sherley-Price, L: 1968: A History of the English Church and People: Harmondsworth, Penguin.

Thomas, K: 1983: The Perception of the Past in Early Modern England: London, University of London Press.

Thorpe, L: 1966: The History of the Kings of Britain: Harmondsworth, Penguin.

Williams, A: 1995: The English and the Norman Conquest: Woodbridge, Boydell.

Yates, F. A: 1994: The Art of Memory: London, Pimlico.

 Acknowledgements

M. Edmonds and H. Ullerthorne, whose extensive libraries and archaeological knowledge I have plundered. N. Perkins for many long discussions. My family for putting up with my obsessions. T. Neal and other friends for their support and gentle prodding.

A.B. Graves (biography)

The author is a graduate of the Landscape Achaeology MA at the University of Sheffield, and appreciates the latitude shown by staff there which has allowed him to develop ideas such as those in the above article. He has a particular interest in archaeological interpretation and in using non (archaeologically) standard means to explore the implications of theoretical speculation - such as creative fiction. Despite this he is no relation to Robert Graves. He hopes to continue working along the lines set out in the above article to produce a greatly expanded and more inclusive thesis for a doctorate should that prove possible.
bengraves2004@yahoo.co.uk