Introduction

Whilst it can certainly be argued that the 'original' Arthur is probably a non-historical (folkloric or mythical) figure who became associated with historical deeds by the ninth century via a process of historicisation (Padel, 1994; Green, 1998; Green, 2007), it has to be recognised that the opposing view has often been taken too. With regards to this, it is fair to say that a vast literature has been generated by the search for historical characters who 'fit the facts' – that is to say, by the quest to identify the 'original' historical Arthur. The present piece is intended as a guide to the four of the most popular theories which have been proposed by those who choose to make the a priori assumption that there really was a historical Arthur at the core of the Arthurian legend. It is felt that such a guide is necessary due to the continuing popularity of this assumption, particularly outside of the academic community, and the potential difficulties for the interested reader in discriminating between the various theories propounded. The value of these theories in general, and of the search itself, is fully discussed elsewhere (Green, 1998; Green, 2007) and needs no further elaboration here, other than to simply say that an enormous number of theories can and have been proposed. In order for the following guide to work, the question of whether the search for a historical Arthur is a useful one is ignored. Similarly, the notion of 'no smoke without fire' – which is criticised heavily elsewhere – is treated as reasonable, i.e. the analyses below follow the theories they discuss in assuming that there probably was a historical Arthur.

Arthur, the Post-Roman War-leader

The notion that Arthur could have been a post-Roman war-leader has it origins in a study of one of the earliest and most important Arthurian sources, chapter 56 of the Historia Brittonum (written c. 829-30 A.D.), in particular the section which says 'Arthur fought against them [the Anglo-Saxon invaders] in those days, together with the kings of the Britons, but he was the leader in battles [dux bellorum].' In the most basic and popular form of this theory, the above sentence is treated as a literal statement that the historical Arthur was a great warrior and war-leader (with an implication, it is often suggested, that Arthur was not a king himself), who led the fight against the Anglo-Saxon invaders. The rest of the text lists a number of his supposed battles, although only one of these – the Battle of Badon – can be proven to have definitely taken place in the post-Roman period (Gildas’ De Excidio Britanniae of c. 540 A.D. mentions the battle, but not Arthur). This theory is essentially the 'default' concept of a historical Arthur for the academic community, and is used by those researchers who believe that Arthur probably existed but think that we can know nothing more of him without entering into the realms of speculation. This view takes the Historia Brittonum chapter 56 as (to some degree) evidence of the existence of Arthur and his basic nature and role, but frequently doesn't trust the contents of this chapter to provide reliable evidence with regards to the battles he fought (aside from Badon) or the region he operated in (see especially Green, 1998 and 2007, and Higham, 2002, on the reasons for this general academic scepticism about the reliability of the Historia Brittonum, particularly with regards to the battles ascribed by it to Arthur). Proponents of this theory of Arthur as a war-leader include Jackson (1959; 1969), who explicitly rejects any localisation of Arthur on the basis of the Historia battles, and Charles-Edwards (1991), who concluded his recent survey of the evidence for a historical Arthur by saying that:

it cannot be ruled out a priori that some useful information about the sixth century may, some day, be surmised on the basis of the [Historia Brittonum] text; but, at the moment, the prospects are poor. At this stage of the enquiry, one can only say that there may well have been an historical Arthur [but] that the historian can as yet say nothing of value about him. (Charles-Edwards, 1991: 29)

The advantages of this theory are (1) that it is based firmly on a critical appreciation of the early Arthurian sources, usually focussing on the Historia Brittonum as the only text worthy of serious consideration as a plausible source of useful information on the nature of the historical Arthur (see, for example, Charles-Edwards, 1991 and Green, 1998 with regards to this); (2) that it recognises the problems inherent in the use of the Historia and other sources; and (3) that, although it does require an a priori assumption that Arthur existed, otherwise this theory refuses to go beyond what can be established from these sources by a correct historical methodology. It does, however, leave us with a somewhat indistinct portrait of the historical Arthur.

Whilst many might accept the above as the most we can legitimately say of any historical Arthur, if we must have one and given the quality of the sources available to us, some have sought to expand this through various means. Jones (1964), for example, argues that the notice regarding Arthur's death at the Battle of Camlann in the Annales Cambriae should be treated as authentic and early and thus added to the above concept of a historical Arthur. Alcock (1971) would argue the same, seeing the Annales Cambriae entries (it also records the Battle of Badon as being fought by Arthur) as the most reliable source of information on any historical Arthur, rather than the Historia Brittonum account, arguing that they derive from sixth-century Easter tables. Both notions have, however, been hotly contested by more recent research into the Annales Cambriae (see for example Dumville, 1977 and Grabowski and Dumville, 1984, which indicate a probable early to mid-tenth-century date for the Arthurian annals) and no academic researcher now accepts the Annales notices as witnesses to the historical Arthur that can be relied upon.

Other attempts to fill-out the above concept of Arthur have focussed on trying to localise this Arthur. The most successful (and perhaps the most methodologically sound) of these is the 'Northern Arthur' theory of Bromwich and others, discussed below, which places the war-leader of the Historia Brittonum in the 'Old North' of Britain on the basis of a consideration of regional bias in the earliest stratum of Arthurian evidence (both historical and literary: see especially Bromwich, 1975-6). Many other theories try to identify and locate the battles of the Historia Brittonum in a particular region – for example, the south-east; the midlands; southern Scotland – in order to localise Arthur there (something the 'Northern Arthur' theory scrupulously avoids indulging in), but these are seriously undermined by Jackson's (1945-6; 1959) warnings about the impossibility of doing this – such attempts rely mainly on linguistic 'ingenuity' rather than sound scholarship – and the general and serious academic scepticism over the trustworthiness of the Historia's list of battles (see Bromwich, 1975-6; Green, 1998 and 2007; and Higham, 2002). Lastly, some attempt to argue from identifications of the battles in the Historia that Arthur was not associated with any particular locality but rather fought battles all over Britain from southern Scotland to south-western England (for example, Alcock, 1971 and 1972). This theory runs into major problems, however, with regards to both plausibility and (once again) the nature and reliability of the Historia Brittonum list of twelve battles, of which, it is worth noting once more, Badon is the only battle mentioned that we know actually took place in the post-Roman period (see Jackson, 1959, especially pp.7-8; Bromwich, 1975-6, especially p. 168ff.; Bromwich et al, 1991: 2-3; Padel, 1994; and Green, 2007 for (1) far better and less 'romantic' explanations of why Historia Brittonum chapter 56 ascribes to its historical Arthur battles in, for example, southern Britain, Chester and Coed Celyddon in southern Scotland, and (2) full discussions of why the Historia's list of battles cannot at all be treated as historically reliable). The simple fact of the matter is that it is now generally agreed that the Historia Brittonum's account is not trustworthy or reliable enough to allow any conclusions about the extent and area of activity of its supposedly historical Arthur to be drawn from it.

Finally, there are those who would return to the Historia Brittonum's statement that 'Arthur fought against them in those days, together with the kings of the Britons, but he was the leader in battles [dux bellorum]' for further inspiration. For them, the above statement indicates something more than simple war-leadership and, perhaps, an implied non-royal status for Arthur (see Jackson, 1959: 9 for an argument that this statement does not rule out Arthur having been a king, and further below): it is rather suggested that what is being described by the Historia Brittonum is Arthur being appointed to the control of some kind of combined British army – a general, if you will, appointed by the British kings to lead the fight against the invaders wherever he is required. Alcock (1971; 1972: 15-18) certainly seems to take this view, having Arthur as a 'general commanding a combined British force', with Arthur and his army riding around Britain and fighting in places as far apart as Bath and southern Scotland (see above on the latter part of this). Some would go even further, making the phrase dux bellorum, 'leader in battles', not a literal statement but an official title or position and analogous to the known Late Roman Comes Britanniae (for example, Rhys, 1884 and 1891: 6-8). Collingwood (1937: 321ff.) sees this post-Roman Dux being placed in charge of a roving Roman-style cavalry unit, whilst Bachrach (1990) favours – on an analysis of (again) the Historia Brittonum battles – having him in control of a fifth-century version of the Roman naval forces in the north and east of Britain.

Certainly these last suggestions offer a much more 'colourful' vision of any historical Arthur but they can be (and have been) accused of going far beyond – sometimes very far beyond – what can reasonably be inferred from the sources as to Arthur's status and his role in the defence of Britain.  In conclusion, most researchers who believe that a historical Arthur is at least possible have preferred to stick with the concept of Arthur described at the beginning of this section as the most that can be legitimately said (especially given the poor quality of our sources): that is, a concept of Arthur as a late fifth-/early sixth-century war-leader, famed for leading the fight against the invading Anglo-Saxons and winning a great victory at Badon, without any of the above speculations about appointed generalships, areas of operation and what-not. To quote Myres (1986: 16), 'if we add anything to the bare statement that Arthur may have lived and fought the Saxons, we pass at once from history to Romance.'

See further, for example, Jackson (1959: 8-9) for a critique of attempts to make Arthur an appointed general and dux bellorum a title, and Charles-Edwards (1991: 24-5, 28) for a discussion of the meaning of dux bellorum, where he argues that the phrase was coined by the author of the Historia Brittonum to reflect his view that Arthur’s role was much like that of Penda, king of Mercia, at the mid-seventh-century Battle of Winwæd, when Penda led a force of thirty other kings and leaders against the Northumbrians, a suggestion that re-opens the question of Arthur's royal status (or lack thereof) and is clearly incompatible with any speculation about 'appointed generals'. Charles-Edwards' comments do, of course, very powerfully raise the question of the extent to which we can really rely upon any of the statements of Historia Brittonum chapter 56 as a useful guide to the supposed fifth/sixth-century reality of Arthur's status and role; in this light, the sentence discussed above would instead represent an anachronism on the part of the author of the Historia Brittonum, projecting his ninth-century ideas about war-leadership onto an earlier age (Charles-Edwards, 1991: 28; see also Higham, 2002: 16-57, 164-5 and chapter one of Concepts of Arthur on this topic).

The Northern Arthur

The Northern Arthur theory is one of the most respectable theories of a historical Arthur, being supported by Thomas Jones, Rachel Bromwich and A.O.H. Jarman, amongst others. This model takes its concept of a historical Arthur from chapter 56 of the Historia Brittonum – that is, it sees him as a late fifth-/early sixth-century warrior famed for leading the fight against the invading Anglo-Saxons (see above). It then uses the nature and perceived regional bias of the very earliest stratum of Arthurian sources to argue that these sources imply that this Arthur was originally a hero of Y Gogledd, the 'Old North' (that is northern England and southern Scotland), and that his later fame throughout Britain was a later secondary development of his legend. The Arthurian reference in Y Gododdin (a poem from the 'Old North') is seen as particularly significant in this theory, as is the concentration of three or four early (c. 550-650 A.D.) 'Arthur' names in the 'Old North', including a prince of the royal house of Dalriada. Other important elements of the evidence for a 'Northern Arthur' include a possible northern British origin for chapter 56 of the Historia Brittonum and the Arthurian references in the Annales Cambriae (these elements are controversial, however: see on the Historia, for example, Bromwich, 1975-6; Dumville, 1977; Dumville, 1986; Koch, 1996: 247-8). Jarman has commented with regards to the Y Gododdin reference (often seen as the earliest reference to Arthur, dating from perhaps as early as c. 600 A.D.) that the poem is a very self-contained and insular work, concerned only with the 'Old North', and thus the mention of Arthur in it can be seen as implying that he was of that region (Jarman, 1989-90: 17-20). The most detailed examination of the evidence for a 'Northern Arthur' is that of Bromwich (1975-6), in which she strongly argues for such an identification and provides a context for Arthur's later, wider, fame by associating the proposed shift of the Arthurian legend to Wales with the well-established movement of early traditions concerning Northern heroes such as Urien Rheged and Llywarch Hen south to Wales by 'at least as early as the ninth century' (Bromwich, 1975-6: 180).

One very major advantage over similar theories of Arthur's geographical origins that this 'Northern Arthur' theory has is, of course, its scholarly rigour and the fact that it is grounded firmly in a detailed and learned analysis of the very earliest Arthurian sources. Another, as Bromwich notes, is its deliberate avoidance of getting tied up in the futile games many authors play in trying to identify the exact location of the battles mentioned in chapter 56 of the Historia Brittonum (Skene, 1868: I, 52-8 is one example of this tendency which might, if correct, support the Northern Arthur hypothesis; see though Jackson, 1945-6, for both criticisms of Skene and the general futility of all attempts to identify and locate more than a handful of the battles recorded in the Historia). Also counting in its favour is the proponents willingness to admit to problems with their theory (something that less reliable theorists hardly ever do), principally the fact that a member of the royal house of Dyfed in south Wales – who was probably born in the late sixth century – was named Arthur, and the reference to Arthur in the mid-seventh-century East Powys poem Marwnad Cynddylan (Bromwich, 1975-6: 177, 179; Jarman, 1981: 5; Jarman, 1989-90: 19. Bromwich offers possible solutions to both of these issues but these are not really satisfactory and do not resolve the issue: see further for discussion and alternative explanations chapter two of Concepts of Arthur, Green, 1998; and Padel, 1994; and below).

Finally, two things must be noted. First, the 'Northern Arthur' theory does naturally depend to some large degree upon the dating, nature and interpretation of the evidence mentioned above, and in this context it is worth noting the controversies surrounding this (Green, 1998 and 2007, and the references therein). Second, the 'Northern Arthur' theory does have questions to answer with regards to the Battle of Badon if – as is generally accepted – this battle was fought somewhere in southern England against the invading Anglo-Saxons. If this 'Northern Arthur' is associated with areas of the 'Old North' such as Rheged or Gododdin, then we have to assume that either (1) Badon was not in the south (which causes problems with both the archaeology and Gildas, though these may not be insurmountable); (2) Arthur ranged widely all over Britain (in which case the dubious and methodologically flawed theories of non-localised Arthur – such as Alcock, 1972 – are in fact correct and he was not an originally solely Northern figure as Bromwich et al argue the sources indicate); or (3) Badon was not originally fought by Arthur. If the latter is true then the entire case for Arthur as a historical personage and defeater of the Saxons starts to collapse, as this case (whatever you may think of its merits) is fundamentally based around the Arthur/Badon connection, with Badon supposedly being the reason for Arthur's fame amongst the Britons and, furthermore, the only thing that ties the Historia Brittonum account of Arthur to known history, with the Historia's account being the mainstay of the case for a historical Arthur (see Green, 1998, and above).

One possible solution to this issue (other than being forced to make the difficult argument for a northern Badon) may be to follow Bromwich in associating Arthur with 'the south-eastern corner of the "Old North", that is with the East Riding of Yorkshire and possibly with York itself' (the later Anglian kingdom of Deira), rather than the more northerly regions. This would put Arthur far enough south to fight fifth-century Anglian invaders (see, for example, the large early-Anglian cremation cemetery at Sancton in the East Riding of Yorkshire) and it is reasonably close to the most northerly of the candidates for Badon, Baumber in northern Lincolnshire. Then, when this area was lost to the invaders, the traditions of a great defender might have been passed northwards to the surviving 'Old North' kingdoms (see Bromwich, 1963; 1975-6: 180-1; 1978: 275; and Thompson, 1979: 215-9 for an argument that the East Riding – or the Vale of York – was in fact the area Gildas was talking about when he gave details of the settlement and rebellion of the Anglo-Saxon federates: this is highly debatable though, for example Wright, 1984; Higham, 1991; Dark, 1993: 260-66; Higham, 1994). Overall, this seems to be the most plausible variant of the 'Northern Arthur' theory.

Arthur the Emperor

The notion that Arthur was some sort of emperor has its origins firmly in the Middle Ages. Geoffrey of Monmouth, in his Historia Regum Britanniae, has Arthur as the ruler of an empire that eventually encompassed Britain, Brittany, Ireland, Iceland, Gaul and Norway, and even challenged Imperial Rome itself, and there may be traces of this conception in earlier texts such as the Old Welsh poem Gereint fil[ius] Erbin, where Arthur is called 'emperor, leader in toil [i.e. battle]', though the term translated as 'emperor', ameraudur, may be better read in this context as 'general' or 'commander' (Jarman, 1983: 106). In the twentieth and twenty-first centuries, however, the notion that the historical Arthur was genuinely an emperor, ruling over the whole of Britain, has not – to say the least – achieved widespread assent amongst academics. In fact the theory had its only serious modern champion in John Morris, who saw Arthur as the dominant figure of his age. Morris made an extensive study of 'The Age of Arthur' (as he termed it) and his main conclusions on Arthur's identity from this research were as follows:

He was the emperor, the all-powerful ruler of the whole of Britain, and the seat of his power was in the lowlands [Colchester, according to Morris, was Camelot]… [He restored] the government of [the] Roman emperor, equipped with a hierarchy of civil and military officers, on the model of that which had existed in the earlier fifth century… These institutions endured for at least thirty years after Badon … With Arthur died the unity of Britain, and all hope of reviving it under British rule… The rule of Arthur had been an age of order, truth and justice, to be praised in retrospect… Arthur dominates and unites the history of two centuries; his victory was the climax and consummation of the fifth-century struggles; and his undoing shaped the history of the sixth century, the mould wherein the future of the British Isles was formed. He was at once the last Roman emperor in the west, and the first medieval king of the country now called England … He left a golden legend, and he rescued a corner of the Roman world from barbarian rule for a short space. (Morris, 1973: 132-141)

Unlike some of the other theories of a historical Arthur discussed here, few would now be tempted to describe Morris' 'Arthur the Emperor' theory as a respectable work of scholarship in its totality, and especially with regards to Arthur. As has been argued at length by two distinguished reviewers, it is 'an outwardly impressive piece of scholarship' which 'crumbles upon inspection into a tangled tissue of fact and fantasy which is both misleading and misguided' (Kirby and Williams, 1975-6). This view is supported to some very large degree by David Dumville in his justly famous attack on both Morris and Alcock (1971), 'Sub-Roman Britain: History and Legend', where he demonstrates the utter invalidity of Morris' approach to the sources which renders his 'reconstruction' of events almost completely worthless (Dumville, 1977). Another reviewer, James Campbell, is slightly more generous, recognising the good hidden in amongst the bad, but he too admits that The Age of Arthur is a book so misleading, so idiosyncratic, so full of problems, difficulties, and traps for the unwary, that it should be used only by professional scholars – already familiar enough with the ongoing debates and the primary sources to ignore the many unreliable theories and passages in the book – and that it is manifestly not a work appropriate for amateurs or newcomers to the subject (Campbell, 1975). Unfortunately, this seems to be just the category of readers who make most use of the book nowadays, with very few professional researchers ever now returning to the tome due to these immense problems.

The South-Western Arthur

The belief that the historical Arthur belongs to the south-west of Britain has been supported recently by authors such as Ashe (1968: 50-1) and Dunning (1988), but it is an ancient and popular association. Thus, for example, he is called penn kadoed Kernyw, 'chief of the battalions of Cornwall', in the c. 1150 non- Galfridian poem Ymddiddan Arthur a'r Eryr, and in Herman's De Miraculis Sanctae Mariae Laudensis ('The Miracles of St. Mary of Laon') a visit to Cornwall in 1113 by some canons from Laon almost ended in violence and rioting when one of the visitors dared to argue that the people of Bodmin were wrong in their belief that Arthur 'still lived', indicating the great strength of interest and feeling amongst the local Cornish population of the early twelfth century towards Arthur (see Coe and Young, 1995: 44-7, and Padel, 1994: 8-9, on this). Another south-western association for Arthur is the fact that all the early native sources – the twelfth-century Welsh poets, the non-Galfridian Trioedd Ynys Prydein (the 'Welsh Triads'), and Culhwch ac Olwen (which has been variously dated from the mid-tenth century to the late eleventh century) – agree that Arthur's court was called Celliwig ('the forest grove') and was to be found in Cornwall. Indeed, Celliwig also seems to appear in the Arthurian poem Pa gur yv y porthaur?, which may date from as early as the ninth century, and as such there is a strong suggestion that the tradition of Celliwig as Arthur's court was one of considerable antiquity (in the poem it is not, however, stated where this place was to be found but, given that Celliwig is never located anywhere other than Cornwall in native tradition, a Cornish location can reasonably be assumed; see further on all the above Bromwich, 1978: 3-4; Padel, 1991: 234-40; Koch, 1994: 1127). Other evidence for an association of Arthur with south-western Britain includes the mid-late ninth-century poem Gereint fil[ius] Erbin; the Vita Prima Sancti Carantoci (c.1100?), which mentions a dragon-slaying episode in Somerset; the story of Gwenhwyfar's abduction and imprisonment at Glastonbury (and Arthur's summoning of the men of Cornwall and Devon to help free her) in the Vita Gildae of Caradoc of Llancarfan (1120s or 1130s); the belief that Glastonbury was Arthur's last resting place and Avalon (see below); and Geoffrey of Monmouth's story of Arthur's conception at Tintagel, Cornwall (many of these sources are discussed further in Green, 1998b and 2007).

These are the kinds of materials upon which the theory of a south-western Arthur has often been built (for example, Wilson, n.d.: 96-7). The problem with all of this is, of course, that it stems mainly from sources reflecting the Arthurian legend, rather than those, such as the Historia Brittonum, which are generally felt to reflect, to some degree, the Arthurian reality. Naturally, this does raise some very important methodological issues. Fundamentally this theory proceeds from the same basis as the 'Northern Arthur' theory, that is an attempt to locate the war-leader of Historia Brittonum chapter 56 by looking at the regional bias of the Arthurian sources. However, two things need to be noticed here. First, we have to recognise that the above sources for a 'South-Western Arthur' are generally far more 'legendary' in nature than those used by the 'Northern Arthur' theory. Second, whilst there are more of them, they are also largely later – sometimes much later – in date than those used by the 'Northern Arthur' theory (especially if the Y Gododdin reference can be dated to before c. 638 A.D., as Koch, 1997 has recently argued; see, however, Green, 1998a). If we are to see the development of the Arthurian legend as a general movement from sober history to fantastical (and increasingly popular) legend, then both of these features would tend to add weight to Bromwich's notion that the presence of the legend in the south-west reflects a secondary development of an originally northern legend and hero, though there are issues with this theory (Bromwich, 1975-6 and above).

Given the above considerations, the case for a 'South-Western Arthur' would seem to require further support if it is to be considered plausible. If the supposed discovery of Arthur's grave in the 1190s by the monks at Glastonbury Abbey (and their claim that Glastonbury was Avalon) could be proven to be genuine then this would obviously significantly alter the situation. C.A. Ralegh Radford (1968) and L. Alcock (1971) have attempted to, at least partially, argue this case, but they fail to convince (see Rahtz, 1993; Carey, 1999; Carley, 1999). Another possible link with reality comes from the Alcock's excavations at Cadbury Castle, Somerset, which showed that this important Iron-Age hill-fort was reoccupied and heavily refortified in the late fifth or sixth century by a very powerful war-lord (Alcock, 1972 and 1995), the Arthurian link being Leland in the sixteenth century who records that the local people thought that this site was Arthur's Camelot. Certainly the possibility is interesting, but the Arthurian link is based on very late traditions, first recorded more than 1000 years after the historical Arthur is supposed to have lived, which severely limits their value in constructing any theory of a historical Arthur; Cadbury-Camelot therefore cannot be taken as proof of a 'South-Western Arthur'.

The question must therefore be asked, is there any good reason to believe that the 'South-Western Arthur' theory is worth supporting? I think that we can cautiously answer 'yes, perhaps' here, on the following basis. Working with the critical study of the early materials as our foundation, we can say that most serious researchers – if they believe in Arthur at all – would argue that the Historia Brittonum is the only plausible source of information on any historical Arthur that we possess, and that the most that can be inferred from this source with any degree of confidence is that Arthur was a late fifth-/early sixth-century war-leader, famed for leading the fight against the Anglo-Saxon invaders and winning a great victory at the Battle of Badon. Now, given that the victory over the Anglo-Saxons at Badon is supposedly the main reason for Arthur's fame – and the fact that it is the only battle associated with Arthur in the Historia Brittonum (the source of the main academic concept of any historical Arthur) which we know from other historical sources actually took place in the post-Roman period – it does not seem unreasonable to take its location as some sort of a guide to the region of operation of any historical Arthur; it is the only remotely reliable clue that our 'historical' sources can provide for us. Of course this does require a degree of assumption, mainly that Badon would have been fought in roughly the region Arthur operated within, but this does not seem unreasonable either. The notion that any historical Arthur was a general or even an emperor(!) who led a combined British army in fighting battles all over Britain – as Alcock (1971 and 1972) and Morris (1973) make him –, not just his own general region (wherever that might have been), is no longer supported by serious researchers and can be dismissed on the grounds of both plausibility and the nature and reliability of the Historia Brittonum list of twelve battles, upon which it largely rests (as noted above; see the other references there for further details, especially Bromwich, 1975-6: 168ff., and above). One would not wish to claim that an identification of Badon might closely localise Arthur – it cannot be denied that post-Roman war-leaders might have ranged over a reasonably wide area, even if the notion of them travelling all over Britain can be dismissed – but it may give us some idea about the general region that he operated within (I am, of course, leaving to one side here the question of whether Badon was actually fought by Arthur, for the very good reasons set out above).

So, where was Badon? Most researchers agree that this battle was fought somewhere in southern Britain, with Jackson writing that 'no amount of ingenuity can make Badon, the most probably genuine of [Arthur's battles], anything but a battle against the Saxons or the Jutes in southern England' (Jackson, 1959: 10 – see also, for example, Bromwich, 1975-6: 172). Some dissenting voices have been raised against this consensus, preferring to argue the difficult case for a location of Badon in the north of Britain (Thompson, 1979: 215-19 and Dumville, 1984: 70-2), but this suggestion is highly contentious and has not received widespread support (see for example Higham, 1991; Dark, 1993: 260-6; Higham, 1994; Wright, 1984). Further, the general consensus that Badon probably belongs to southern Britain is supported and supplemented by the archaeology of fifth-century Anglo-Saxon settlement, which indicates that the earliest and most extensive settlements and conquests by the invading Anglo-Saxons occurred in the south and east of Britain, making this the most plausible region for the operation of a British war-leader fighting the invaders (good modern summaries and introductions to the archaeology include Carver, 1989; Hawkes, 1989; Hines, 1990; Higham, 1992; Welch, 1993; Scull, 1995; and Arnold, 1997, chapter 2).

If we can thus say that Badon probably belongs to southern Britain, where in southern Britain was it? There are two main theories with regards to this. The first theory argues that the name Badon would, when it was taken into Old English and if the site was a fortified hill, regularly become the modern English place-name 'Badbury' and variants (see Jackson, 1953-8; Gelling, 1988: 60-1). There are a number of 'Badbury' names in southern and eastern England that might thus have their origins in Badon, including Badbury Rings (Dorset), Liddington Castle (Wiltshire, this site being once known as Badbury Camp), Badbury Hill (Berkshire), and Baumber (Lincolnshire), with Badbury Rings in Dorset often being the favoured above the others (see especially Jackson, 1953-8). Another favoured 'Badbury' identification is Liddington Castle, but recent archaeological excavations there seem to cast doubt on the plausibility of any identification of this site with Badon, despite support for it from Chambers (1927), Myres (1986: 159) and others (see Hirst and Rahtz, 1996). The second theory follows Geoffrey of Monmouth's Historia Regum Britanniae in identifying Badon with Bath, an identification that has been taken up in modern times by Alcock (1971: 70-71) and the Burkitts (1990) on both philological and archaeological grounds. On the whole it cannot yet be said to be clear which of these competing theories should command our support. However, as the most recent survey of the evidence has commented, we can say on the most general level that the balance of probabilities and current scholarly opinion indicates that 'the battle probably took place in the south-west' (Hirst and Rahtz, 1996: 17).

In conclusion, if we assume (as it does not seem unreasonable to do) that the location of Badon can be seen as a rough guide to the general region that the Arthur of the Historia Brittonum chapter 56 operated in, then it follows that Arthur was quite possibly a figure from the southern portion of Britain and most probably the south-west. This notion has the major advantage of being based around what can be reasonably assumed from the 'historical' sources about any possible genuinely historical Arthur, rather than a perceived regional bias in the 'non-historical' material concerning Arthur, and if it is accepted then it does go a long way towards making the 'South-Western Arthur' theory the plausible theory that the later legendary material on its own cannot.

There are, naturally, a number of issues with this theory: First, by its very nature in arguing for a 'South-Western Arthur', this theory ignores or sets-aside the evidence for a northern bias in the early Arthurian material such as Y Gododdin, as observed by Bromwich (1975-6) and others (see above). To this we can add that there is also early evidence for a knowledge of Arthur and his legends in Wales, including Historia Brittonum chapter 73 (the traditions related here are considered to pre-date the ninth century: Bromwich and Evans, 1992: lxvi); the poem Preideu Annwfyn (which Koch argues should be dated to before the late eighth century: Koch, 1996: 263-5); the mid-seventh-century East Powys poem Marwnad Cynddylan; and the existence of Arthur map Petr of the Dyfed royal house, born c. 570 A.D. (Bromwich, 1975-6: 178-9). This is, of course, part of a wider issue which seriously affects the 'Northern Arthur' theory too: even the very earliest and most reliable evidence for the distribution of a knowledge of the Arthurian legend – the use of the name Arthur in the mid-late sixth century by the royal houses of both Dyfed and Dalriada – indicates that this knowledge was extremely widely spread, from south Wales to southern Scotland. The explanation of all this is very difficult, especially given that theories of the historical Arthur as an age-defining figure who fought all around Britain are no longer considered plausible or methodologically defensible, as noted above (for possible solutions to this problem, see Bromwich, 1975-6: 177ff.; Padel, 1994, especially p.24; Green, 1998a and chapter two of Concepts of Arthur: Bromwich's is the only solution that has been offered which would maintain Arthur as a historical figure, and it forms part of her 'Northern Arthur' theory). Of course, as formulated above, the 'South-Western Arthur' theory is primarily based around the 'historical' sources, not materials reflecting the growth of the Arthurian legend, but this does not mean that these considerations can be discounted or ignored; the question still has to be asked, if we argue that the historical Arthur belonged to the south-west, what then are we to make of this very early evidence for a knowledge of his legend in Wales and southern Scotland?

Second, the above argument for a 'South-Western Arthur' is based around an assumption that Badon would have been fought in roughly the same region that any historical Arthur operated within. Whilst this is not at all unreasonable, it is an assumption and it should be remembered that early medieval war-bands could be very mobile, raiding deep into enemy territory. Given, however, that we are only looking for a rough general region for Arthur's operation, rather than an exact location, this is much less of a concern and an issue than it might have been. And third and finally, we should not forget that Badon has not actually been securely identified. Whilst most agree that it was fought in southern Britain and most likely in the south-west, there are dissenting voices. It ought not to be forgotten that one of the ‘Badbury’-style place-names which may derive from Badon is located in Lincolnshire, still south of the Humber but in the East Midlands rather than the south-west. Furthermore, Badon may not, of course, have been any of the places so far suggested but rather some as-yet-undiscovered site elsewhere in Britain, where perhaps a new English or Scandinavian place-name has silently replaced and erased the earlier name ‘Badon’.

Bibliography

L. Alcock, Arthur's Britain: History and Archaeology A.D. 367-634 (London, 1971)

L. Alcock, 'By South Cadbury, is that Camelot...' Excavations at Cadbury Castle 1966-70 (London, 1972)

C.J. Arnold, An Archaeology of the Early Anglo-Saxon Kingdoms, second edition (London, 1997)

G. Ashe, 'The Arthurian Fact', in G. Ashe (ed.) The Quest for Arthur's Britain (London, 1968), pp. 27-57

B.S. Bachrach, 'The Questions of King Arthur's Existence and of Romano-British Naval Operations', The Journal of the Haskins Society, 2 (1990), pp. 13-28

R. Bromwich, 'Scotland and the Earliest Arthurian Tradition', Bulletin Bibliographique de la Société Internationale Arthurienne, 15 (1963), pp. 85-95

R. Bromwich, 'Concepts of Arthur', Studia Celtica, 10/11 (1975-6), pp. 163-81

R. Bromwich (ed. and trans.), Trioedd Ynys Prydein, second edition (Cardiff, 1978)

R. Bromwich et al, 'Introduction', in R. Bromwich et al (edd.) The Arthur of the Welsh. The Arthurian Legend in Medieval Welsh Literature (Cardiff, 1991), pp. 1-14

R. Bromwich and D. Simon Evans (edd.), Culhwch and Olwen. An Edition and Study of the Oldest Arthurian Tale (Cardiff, 1992)

T. Burkitt and A. Burkitt, 'The frontier zone and the siege of Mount Badon : a review of the evidence for their location', Proceedings of the Somerset Archaeological and Natural History Society, 134 (1990), pp.81-93

J. Campbell, 'The Age of Arthur', Studia Hibernica, 15 (1975), pp. 177-85

J. Carey, 'The Finding of Arthur's Grave: A Story from Clonmacnoise?', in J. Carey et al (edd.) Ildánach Ildírech. A Festschrift for Proinsias Mac Cana (Andover & Aberystwyth, 1999), pp. 1-14

J.P. Carley, 'Arthur in English History', in W.R.J. Barron (ed.) The Arthur of the English (Cardiff, 1999), pp. 47-57

M.O.H. Carver, 'Kingship and material culture in early Anglo-Saxon East Anglia', in S. Bassett (ed.) The Origins of Anglo-Saxon Kingdoms (London, 1989), pp. 141-58

E.K. Chambers, Arthur of Britain (London, 1927)

T. Charles-Edwards, 'The Arthur of History', in R. Bromwich et al (edd.) The Arthur of the Welsh. The Arthurian Legend in Medieval Welsh Literature (Cardiff, 1991), pp. 15-32

J.B. Coe and S. Young, The Celtic Sources for the Arthurian Legend (Felinfach, 1995)

R.G. Collingwood and J.N.L. Myres, Roman Britain and the English Settlements (Oxford, 1937)

K.R. Dark, Civitas to Kingdom: British Political Continuity, 300-800 (London, 1993)

D.N. Dumville, 'On the North British Section of the Historia Brittonum', Welsh History Review, 8.3 (1977), pp. 345-54

D.N. Dumville, 'Sub-Roman Britain: History and Legend', History, 62 (1977), pp. 173-92

D.N. Dumville, 'The chronology of De Excidio Britanniae, Book 1', in M. Lapidge and D.N. Dumville (edd.) Gildas : New Approaches (Woodbridge, 1984), pp. 61-84

D.N. Dumville, 'The Historical Value of the Historia Brittonum', Arthurian Literature, 6 (1986), pp. 1-26

R. Dunning, Arthur: King in the West (Gloucester, 1988)

M. Gelling, 'Towards a chronology for English place-names', in D. Hooke (ed.) Anglo-Saxon Settlements (Oxford, 1988), pp. 59-76

L.M. Gowans, Cei and the Arthurian Legend, Arthurian Studies XVIII (Cambridge, 1988)

K. Grabowski and D.N. Dumville, Chronicles and Annals of Medieval Ireland and Wales (Woodbridge, 1984)

T. Green, 'The Historicity and Historicisation of Arthur' (1998), archived at http://www.arthuriana.co.uk/historicity/arthur.htm

T. Green, 'A Bibliographic Guide to Welsh Arthurian Literature' (1998b), archived at http://www.arthuriana.co.uk/notes&queries/N&Q1_ArthLit.pdf

T. Green, Concepts of Arthur (Stroud, 2007)

S.C. Hawkes, 'The South-East After the Romans: The Saxon Settlement', in V.A. Maxfield (edd.) The Saxon Shore. A Handbook (Exeter, 1989), pp. 78-95

N.J. Higham, 'Old light on the Dark Age landscape: the description of Britain in the De Excidio Britanniae of Gildas', Journal of Historical Geography, 17 (1991), pp. 363-72

N.J. Higham, Rome, Britain and the Anglo-Saxons (London, 1992)

N.J. Higham, The English Conquest: Gildas and Britain in the Fifth Century (Manchester, 1994)

N.J. Higham, King Arthur, Myth-making and History (London, 2002)

J. Hines, 'Philology, Archaeology and the Adventus Saxonum vel Anglorum', in A. Bammesberger and A. Wollman (edd.) Britain 400-600: Language and History (Heidelberg, 1990), pp. 17-36

S. Hirst and P. Rahtz, 'Liddington Castle and the Battle of Badon : Excavation and Research 1976', The Archaeological Journal, 153 (1996), pp. 1-59

K.H. Jackson, 'Once Again Arthur's Battles', Modern Philology, 43 (1945-6), pp. 44-57

K.H. Jackson, 'The site of Mount Badon', Journal of Celtic Studies, 2.2 (1953-8), pp. 152-55

K.H. Jackson, 'The Arthur of History', in R. Loomis (ed.) Arthurian Literature in the Middle Ages (Oxford, 1959), pp. 1-11

K.H. Jackson, The Gododdin (Edinburgh, 1969)

A.O.H. Jarman, 'The Delineation of Arthur in Early Welsh Verse', in K. Varty (ed.) An Arthurian Tapestry: Essays in Memory of Lewis Thorpe (Glasgow, 1981), pp. 1-21

A.O.H. Jarman, 'The Arthurian Allusions in the Black Book of Carmarthen', in P.B. Grout et al (edd.) The Legend of Arthur in the Middle Ages (Cambridge, 1983), pp. 99-112

A.O.H. Jarman, 'The Arthurian Allusions in the Book of Aneirin', Studia Celtica, 24/25 (1989-90), pp. 15-25

T. Jones, 'The Early Evolution of the Legend of Arthur', Nottingham Medieval Studies, 8 (1964), pp. 3-21

D.P. Kirby and J.E.C. Williams, 'Review of The Age of Arthur', Studia Celtica, 10-11 (1975-6), pp. 454-86

J.T. Koch, 'Review of R. Bromwich et al (edd.) The Arthur of the Welsh', Speculum, 69.4 (October, 1994), pp. 1127-9

J.T. Koch, 'The Celtic Lands', in N.J. Lacy (ed.) Medieval Arthurian Literature: A Guide to Recent Research (New York), pp. 239-322

J.T. Koch, The Gododdin of Aneirin. Text and Context from Dark-Age North Britain (Cardiff, 1997)

J. Morris, The Age of Arthur (London, 1973)

J.N.L. Myres, 'Review of The Age of Arthur', English Historical Review, 90 (1975), pp. 113-6

J.N.L. Myres, The English Settlements (Oxford, 1986)

O.J. Padel, 'Some south-western sites with Arthurian associations', in R. Bromwich et al (edd.) The Arthur of the Welsh. The Arthurian Legend in Medieval Welsh Literature (Cardiff, 1991), pp. 229- 248

O.J. Padel, 'The Nature of Arthur', Cambrian Medieval Celtic Studies, 27 (Summer 1994), pp. 1-31

P. Rahtz, Glastonbury (London, 1993)

C.A. Ralegh Radford, 'Glastonbury Abbey', in G. Ashe (ed.) The Quest for Arthur's Britain (London, 1968), pp. 97-110

J. Rhys, Celtic Britain (London, 1884)

J. Rhys, Studies in the Arthurian Legend (Oxford, 1891)

C. Scull, 'Approaches to material culture and social dynamics of the migration period in eastern England', in J. Bintliff and H. Hamerow (edd.) Europe Between Late Antiquity and the Middle Ages (Oxford, 1995), pp. 71-83

W.F. Skene, The Four Ancient Books of Wales (Edinburgh, 1868)

E.A. Thompson, 'Gildas and the History of Britain', Britannia, 10 (1979), pp. 203-26

M. Welch, 'The archaeological evidence for federate settlement in Britain in the fifth century', in F. Vallet and M. Kazanski (edd.) L'Armée Romaine et les barbares du IVe au VIIe siècles (Paris, 1993), pp. 269-77

C. Wilson, 'Search for the Real Arthur', in B. Duxbury et al, King Arthur Country in Cornwall (St Teath, n.d.)

N. Wright, 'Gildas's Geographical Perspective: Some Problems', in M. Lapidge and D.N. Dumville (edd.) Gildas : New Approaches (Woodbridge, 1984), pp. 85-106

* * * * *