This is a PDF transfer of the review article �The Historicity and Historicisation of Arthur� (archived at http://www.arthuriana.co.uk/historicity/arthur.htm), the first version of which appeared online in 1998. This transfer has been undertaken in order to ease both the reading and referencing of this article. The text itself remains unchanged from the HTML version, aside from being necessarily re-formatted and having the separate notes incorporated as end-notes. An up-to-date expansion, development and revision of the views presented below can be found in my Concepts of Arthur.
Thomas Green, Exeter College
University of Oxford
14th May 2006
* * *
The Historicity and Historicisation of Arthur
Many different theories are available as to the �identity� of Arthur and some brief methodological notes will be found here regarding the making of such identifications. While these theories are interesting, they fail to address fully one important question � was there a historical post-Roman Arthur? Many books, articles and web-pages simply make the a priori assumption that there has to be a historical figure behind the Arthurian legends. Such an assumption is totally unjustified. As anyone at all familiar with medieval literature in general will know, the historicisation of nonhistorical/mythical personages � often through association with some important event of the past � is not in any way an unusual occurrence. Some examples of this that will probably particularly interest readers of this article are Hengest and Horsa, who were Kentish totemic horse-gods historicised by the 8th-century with an important role in the 5th-century Anglo-Saxon conquest of eastern Britain (see Turville-Petre, 1953-7; Ward, 1969; Brooks, 1989; Yorke, 1993); Merlin (Welsh Myrddin), who was an eponymous founder-figure derived from the place-name Caer-fyrddin and historicised with the deeds of one Lailoken (see Jarman, 1991); and the Norse demigod Sigurd/Siegfried who was historicised by being associated with a famous historical battle between the Huns and the Burgundians dated 437AD, in the Nibelungenlied (Thomas, 1995, p.390).1
Given this, no a priori judgements can be made as to whether a figure is, in origin, historical, mythical or fictional � each individual case must (and can only) be decided by a close examination of all the relevant material. When we have figures such as Arthur being portrayed as historical we are therefore, on a very basic level, looking at either a historical figure or a legendary figure who became historicised, with neither explanation enjoying priority on a priori grounds � it must be recognised that one can only say that there has to have been a historical Arthur once all the material has been evaluated and this has been shown to be the case; there is no possible justification for simply assuming this. The following article is intended to provide a summary account and bibliography of the latest academic research into Arthur with a particular focus on the question of historicity. Aside from the various articles and books cited, much of what is below has been discussed in detail on the discussion list of the International Arthurian Society, Arthurnet, in a moderated debate that I had the great pleasure of chairing. The results of this discussion, including all
posted comments, can be found in the Arthurnet archives.
The Historical Arthur: an Analytical and Bibliographic Survey
Any inquiry into the �historical� Arthur must proceed from the sources. One of the most important sources for the student of post-Roman Britain is archaeology and, indeed, the case is sometimes made that it is our only reliable source (see, for example, Arnold, 1984). When looking at Arthur�s possible historicity however, archaeology cannot really help as it deals with sites not people � it can show that a site was occupied in the right period but only very rarely (that is, when we have an inscription) can it tell us who the occupier was. The only piece of archaeological data
which might have been significant to the debate is the Glastonbury cross naming King Arthur as the occupant of the grave it was supposedly found in by the monks of Glastonbury in 1191. Some have suggested a mid 10th- or 11th-century date for this (for example, Radford, 1968; Alcock, 1971) but it is now clear that it was the product
of a late 12th-century fraud and derivative of Geoffrey of Monmouth�s Historia Regum Britanniae, and thus of no use in the search for a historical Arthur (see Rahtz, 1993; Carey, 1999; Carley, 1999; Gransden, 1976; Somerset and Dorset Notes & Queries for 1984 &c.; there was a copy of Geoffrey�s Historia at Glastonbury from c.1170.
The early 6th-century inscribed stone that has recently been found at Tintagel does not refer to Arthur, contrary to reports by English Heritage and the media). Given the above, any conclusions regarding Arthur�s historicity, or lack thereof, must be drawn from the textual references to him. The King Arthur we encounter in the later medieval texts (and with which people are often most familiar) is not the Arthur of earlier works � shortly before A.D. 1139
Geoffrey of Monmouth (Galfridus Monemutensis) completed his Historia Regum Britanniae (�History of the Kings of Britain�) which glorified Arthur and made him an international warlord. This work quickly became influential throughout western
Europe and affected the Arthurian legend in all areas with the result that, in general, scholars look to sources written before Geoffrey�s Historia for the �original� Arthur (that is, in the �pre-Galfridian� sources). One well known dissenter from this is Geoffrey Ashe (1981; 1985; 1995) who argues that Riotamus, a 5th-century �king� of the Britons who campaigned on the continent, is the actual historical prototype of Arthur and Geoffrey of Monmouth drew on this tradition when writing his magnum opus. While this theory is quite popular it is rightly dismissed by academics as
nothing more than �straws in the wind� (Bromwich et al, 1991, p.6. See also Padel, 1994, p.31, n.113; Hanning, 1995; Padel, 1995) on the grounds that, while Riotamus (or Breton traditions about this figure) could be the (partial) inspiration for Geoffrey�s portrayal of Arthur, he has nothing at all in common with the insular traditions of Arthur and thus cannot be the prototype for Arthur as a whole (indeed, he doesn�t even have the correct name. Ashe explains this by saying that Riotamus was a title and Arthur was his real name but a recent reviewer (Padel, 1995) has shown this to be incorrect). The above means that the historical Arthur, if he existed, will be found in the pre-Galfridian texts and it is to these we must now turn.
The pre-Galfridian sources for Arthur can be most conveniently read in Coe and Young (1995), which provides facing text and translation. Some earlier historians, such as John Morris (1973), tried to make use of, as historical texts, all the sources which mentioned Arthur including, for example, the Saints� Lives and late poetry. This tendency has been correctly and heavily criticised by David Dumville (1977a), amongst others, mainly because these sources cannot be seen as in any way historically reliable � we are therefore, when looking at a possibly historical Arthur and in the light of Dr Dumville�s learned comments, essentially confined to four pieces of evidence which might contain information of real historical value: the Annales Cambriae (Phillimore, 1888; Morris, 1980); the Historia Brittonum (Morris, 1980; Dumville, 1985; Koch and Carey, 1995); the collection of heroic death-songs known as Y Gododdin (Jackson, 1969; Jarman, 1988; Koch, 1997); and the four or five occurrences of the name Arthur in 6th- and 7th-century contexts (Barber, 1972; Bromwich 1975-6; Coe and Young, 1995, pp.156-165).
Dealing with the last of these first, the occurrence of four (or possibly five) people named �Arthur� in 6th- and 7th-century western Scotland and Wales has often been seen as one of the best pieces of evidence for a historical Arthur � the argument is, essentially, that the appearance of these names reflects the commemoration of an
earlier historical figure (see, for example, Chadwick and Chadwick, 1932).2 However such a commemoration by name of an earlier historical hero would be totally unparalleled in the Celtic world and as such cannot be at all supported as an explanation of these names (see Bromwich, 1975-6, pp.178-9). Thus these names cannot be used as evidence for a historical Arthur and as long as we continue to see Arthur as genuinely historical they are likely to remain a lasting crux (at present there is only one viable explanation of these names, that proposed by Dr Oliver Padel (1994, p.24) � see below on this. It is worth noting that none of these �Arthurs� can be seen as the �original� Arthur, pace Barber, 1972 � see Bromwich, 1975-6, p.179; Jackson, 1973; Roberts, 1973-4).3
The second source for consideration is the collection of heroic death-songs known as Y Gododdin, relating to a battle fought in the late 6th-century. In recent years there has been considerable debate over the statement in Y Gododdin that Gordur �fed black ravens on the rampart of a fort, although he was no Arthur� (B.38. Koch (1997)
numbers this B�.38). Thomas Charles-Edwards (1991, p.14), building on his theory of textual transmission (set forth in Charles-Edwards, 1978), concluded that, as the reference only occurs in the B version and not the A version of Y Gododdin, it need be no older than the 9th- or 10th-century. Recently, however, John Koch (1997) has
attempted a �reconstruction� of the �original� text of Y Gododdin and includes the �Arthurian� reference in this text, dated by him to pre-638AD. Whilst his is certainly an interesting exercise in discovering how Y Gododdin might have looked if it was of 6th- or 7th-century date, the limitations of this �reconstruction� must be recognised.
As one reviewer has noted, Koch�s text is, in reality, a translation of Y Gododdin into the language of c.600AD and in this it must be seen in the same light as Jarman�s earlier translation of this text into modern Welsh (Jarman, 1988) � Koch has not shown that Y Gododdin was composed in this period, only what it might have looked like if it had been (Padel, 1998). Indeed, Isaac has demonstrated that Koch�s whole theory of the creation and transmission of Y Gododdin, including the idea that B� represents the Ur-text, cannot be at all supported (Isaac, 1999).
Similar caveats have been shown to apply to Koch�s �reconstruction� of the poem Gweith Gwen Ystrat, with Isaac demonstrating that whilst one can undertake such an exercise and show how this poem would have looked if it had been composed c.600AD, such a reconstruction is entirely unwarranted and there is no reason to think that the text
was composed inthis period (Isaac, 1998). Given the above, it seems clear that, despite Koch�s assertions, �[t]he date of composition [of Y Gododdin] remains as unclear as ever.� (Padel, 1998, p.55). Indeed Isacc (1996; 1999) has recently followed D. Simon Evans (1978) in arguing that there is no linguistic evidence that would necessitate dating Y
Gododdin as a whole before the 9th- or 10th-century and, in light of all of this, Charles-Edwards� comments on the antiquity of the Arthurian references in this text must stand.
Turning to the �Arthurian� awdl (�stanza�) of Y Gododdin, how does this reference affect the question of Arthur�s historicity, given that Arthur only appears as a comparison to a warrior of (supposedly) the late 6th-century? One common argument is that in works such as Y Gododdin the figures named are always believed to be historical and therefore the Arthurian awdl would seem to indicate that by the 9th- or 10th-century Arthur was believed to have been a historical personage, at least by the author of Y Gododdin (see Jarman, 1989-90; Bromwich et al, 1991). Whilst
superficially convincing, there are considerable problems with such a judgement. First, the simple fact of the matter is that we can only identify a few of the characters that appear in early Welsh heroic poetry; many of the people in the poems appear only there, so that we have no knowledge of whether they were (or were thought to be) historical or not � it is an assumption, nothing more, that everyone in these poems was a real historical figure and as such we cannot take Arthur�s presence in Y Gododdin as evidence either for his historicity or a belief in his historicity . Second, in Y Gododdin Arthur is in the remarkable position of appearing �only not to appear� (Padel, 1994, p.14).
Unlike Gordur or the other warriors he is not actually present at the battle: �In the allusion, Arthur is presented as the unrivalled paragon of martial valour and is thus used to form a highly unusual comparison by rendering explicitly inferior the honorand of the awdl (�stanza�). Therefore, if the relevant awdl and lines can be sustained as Aneirin�s original, this would tell us that by the later sixth century there existed in North Britain a tradition of a Brittonic superhero Arthur...� (Koch, 1996, p.242). Whilst we might not be able to accept Koch�s assertions on dating, we can say that Arthur is essentially a �highly unusual comparison�, not a warrior who is being honoured; he is not envisaged as being present at the battle and he is a military �superhero�, someone to whose heights of valour not even a man who killed 300 in one rush could compare.
He is therefore in a different league to the rest of the figures who appear in Y Gododdin and, as such, there is no reason to think that assumptions drawn from the identifications of a few characters in the text as a whole, even if they were viable, would apply to him. All the Y Gododdin reference tells us is that Arthur was seen, by the 9th- or 10th-century, as �the impossible comparison� (Padel, 1994, p.14), a �superhero� to whom not even the greatest living warrior could compare; it does not tell us whether this reflects a mythical �superhero� named Arthur or a historical Arthur mythicized and Arthur is, in the text, in no way associated with the defence of post-Roman Britain or any specific period of history.
In light of the fact that neither of the above can help in the investigation of Arthur�s possible �historicity�, the case for a historical Arthur rests entirely on two sources, the Historia Brittonum and the Annales Cambriae, both of which would appear to have a concept of Arthur that is (at least partly) unequivocally historical. The Historia Brittonum was written anonymously in A.D. 829/30, the ascription to one �Nennius� now being regarded as false (Dumville, 1974; 1975-6, though see Field, 1996). There is considerable debate over the nature of the text (see, for example, Dumville, 1986;
Charles-Edwards, 1991; Dumville, 1994; Koch, 1997; Howlett, 1998) but it now seems clear that the writer of the Historia was not an ignorant and incompetent compiler who simply �made a heap� of earlier sources but rather �author� who wrote the Historia Brittonum with a unity of structure and outlook and engaged in the active processing of his sources, and this conclusion is endorsed by the researches of Dr David Howlett who sees the Historia as a work of architectonic genius making use of the sophisticated �Biblical style� in its construction (Howlett, pers. comm.; 1998,
chp.5. For the Celtic-Latin tradition of Biblical style see Howlett, 1995).
Given the above, we must question to what extent the author altered his sources for his own purposes, what were the nature of his sources, and thus how far can we trust what we read in the Historia? Dumville (1986) took a very pessimistic line on this, arguing that it was a source only for the 9th-century and its concerns. While this view
has been challenged by Thomas Charles-Edwards (1991), who identifies the Historia as a fusion of the two historical genres, historia gentis and historia ecclesiastica, it is still clearly the case that �even where credit might be given to the supposed source [of a section of the Historia], the author�s methods...do not encourage us to be confident
about the possibility of recovering usable information about the period whose history he was narrating. His procedures were synthetic and interpretive, his sources overwhelmingly non-contemporaneous with the events which they purport to describe� (Dumville, 1994, p.419).
As such the Historia is of very dubious historical value, for example, in addition to many of its sources being of a similar date to itself and suspect in nature, the Historia can be shown to portray characters who are decidedly mythical in origin, such as Hengest and Horsa (see Turville-Petre, 1953-7; Ward, 1969; Brooks, 1989; Yorke, 1993; Green, n.d.), as genuinely historical. Indeed, as a number of recent commentators have recognised, the Historia Brittonum is in fact a synchronising and synthetic history of the type well known from medieval Ireland, fusing sources for its own political ends and involved in the creation of a full national pseudo-history, a process which was closely allied with the historicising of legend (Padel, 1994, p.23; Carey, 1994; Dumville, 1994; Coe and Young, 1995, p.6).
Directly relevant to this question of the �historical value� of the Historia Brittonum is the fact that the author of the Historia was not writing �history� as we know it today but was rather engaging in something more akin to that which we would call sermonising, and this must be remembered in any analysis of the Historia. To try and read such works as the Historia as linear history is completely false to the methods and assumptions with which they were composed (see Hanning, 1966; Howlett, 1998; N. Hinton, pers. comm).
This leads us to Chapter 56 of the Historia Brittonum, which contains the references to a �historical� Arthur. This is �a pseudo-historical account of a suspiciously formulaic list of twelve battles against Germanic invaders� (Coe and Young, 1995, p.6), supposedly fought by Arthur. Some have suggested (for example, Chadwick and Chadwick, 1932; Jones, 1964) that Chapter 56 could have been based on a poem written in Welsh that was translated into Latin by the author of the Historia. Whilst this is an interesting suggestion it has to be recognised that such a notion is speculation and it does not allow us to give this section of the Historia an early date.
Indeed, various considerations indicate that any such hypothetical poem would date to much the same period as the Historia anyway (see Jackson, 1945-6, p.57; Jackson, 1959a, pp.7-8; Dumville, 1977a, p.188; Jarman, 1981, pp.2-3; Dumville, 1986, pp.13- 4; Charles-Edwards, 1991, pp.21-29; Padel, 1994).6 Furthermore it must not be
forgotten that, with the writer of the Historia Brittonum now seen as an author actively manipulating his text to create a synthetic pseudo-history rather than a simple compiler, Chapter 56 was, to some large extent, his creation. This is underlined by Dr. Howlett�s (1998, chp. 5) discovery that this section is written in the highly complex
�Biblical Style�, showing that Chapter 56 was an integral part of the Historia that was created, engineered and planned by the author in accordance with his aims and methodology. As such the notion that Chapter 56 might represent anything like a postulated earlier source incorporated bodily into the text of the Historia can be rejected.
Instead it seems clear that this chapter, along with its concept of Arthur, cannot be separated from the Historia as a whole, the aims, methodology, unity of structure and outlook with which this was created, or, indeed, the general comments of Dr Dumville and others on the nature of the Historia and its sources noted above (see further Hanning, 1966; Barber, 1972, p.101ff.; Charles-Edwards, 1991, p.21ff. on Chapter 56 as an integral and inseparable part of the Historia). The best we can therefore honestly say is that in the Historia Brittonum, a source of very dubious
historical value (which can be shown to portray mythical figures as genuinely historical), we have evidence for the idea that Arthur was a historical figure being current by A.D. 829/30 at the latest.
Our last source, the Annales Cambriae, was compiled in 950s and is sometimes seen as providing good evidence for Arthur being a historical figure (see Grabowski and Dumville, 1984 for the dating. Studies and commentaries on the text include Jones, 1964; Alcock, 1971; Hughes, 1980; Grabowski and Dumville, 1984; Dumville, 1990; Charles-Edwards, 1991 and Koch, 1996. Dumville apparently has a new study of the Annales forthcoming). It mentions Arthur in two entries: that for A.D. 516 which tells of the �battle of Badon, in which Arthur carried the cross of our Lord Jesus Christ on his shoulders for three days and three nights, and the Britons were the victors� and that for A.D. 537 concerning �the battle of Camlann, in which Arthur and Medraut fell�. In assessing the value of these entries, considerable attention should be paid to the date of these annals. Jones (1964) and Alcock (1971) were both inclined to see at least one of these annals as a contemporary record of Arthur and, if it could be accepted, such a conclusion would �prove� Arthur�s historicity.
However, the late Dr. Hughes (1980) in her important and extensive studies of the Annales reached a rather different (and convincing) conclusion, and this has been built upon by Dumville (in Grabowski and Dumville, 1984) and Charles-Edwards (1991) � the Annales Cambriae to 613 is basically a version of the �Chronicle of Ireland�, with the sections from 613 to 777 being based on North British materials; there is absolutely no justification for thinking that any of the pre-613 British entries are drawn from contemporary or even near-contemporary sources and, rather, they should be seen as retrospective interpolations dating from between the very late 8th-century (the period in which the
�Chronicle of Ireland� was first brought together with the post-613 North British materials at St David�s in order to extend backwards a chronicle kept by that community from the closing years of the 8th-century onwards) and the mid 10thcentury (when the Annales reached something like its final form). Indeed, in light of Dr Dumville�s further researches into the date of this bringing together, the above 7 terminus post quem for the interpolations might well be shifted forward to the earlymid 10th-century.
Looking at the annals themselves, one very important point must be made: the Badon entry in the Annales is not an independent witness to Arthur�s historicity. Instead it is clearly related to the Historia Brittonum�s account (Chapter 56) of Arthur�s eighth battle at Guinnion Castle, in which Arthur carries an icon on his shoulders into battle
with him, and as such the Annales account either derives from the Historia Brittonum or its source. Thomas Charles-Edwards has suggested (1991, pp.25-8) that they be seen as dual elaborations of single original, the entry in neither case being very much older than the text it is contained in (829/30 for the Historia and the 950s for the
Annales). However, a more convincing explanation has been provided by Professor John Koch. Koch observes that both the Historia Brittonum and the Annales Cambriae have the probable confusion of Old Welsh scuit �shield� and scuid
�shoulder� in them and notes that �that error of transmission is hardly likely to have come about twice�. He goes on to say that �In all details, the Annales Cambriae entry is more easily understood as derived from Historia Brittonum�s account�, which would appear to be the most probable scenario on the present evidence and is sound even without the support of the scuit/d confusion (see Koch, 1996, pp.252-3 for discussion; also Barber, 1972, p.105). Similarly the second entry regarding Camlann is best viewed as non-traditional and as having mid 10th-century origins (see Charles-
Edwards, 1991, pp.25-7, 28; Ashe, 1986, pp.76-8; Wood, 1981, pp.59-60; Bromwich, 1978a, p.487; Jarman, 1983, p.109), with the consequence that the Annales Cambriae cannot really be seen to be of any independent value in making the case for a �historical Arthur�. As a result we are forced to return to the text of the Historia Brittonum.
Whilst general comments on Chapter 56 of the Historia Brittonum have been made, a more detailed examination of the information contained within it may prove enlightening. It is easy to assume that all the battles mentioned in Chapter 56 were remembered as being those fought by Arthur but such assumptions may well be incorrect. Perhaps the most famous �Arthurian� battle is that of Badon (in montis badonis) but the reference to this has serious problems. It has long been accepted that this is the same battle as the obsessio Badonici montis of Gildas�s De Excidio
Britanniae � 26 (see Winterbottom, 1978 for an edition and translation. The date of publication of this work, and thus the date of Badon, has been much discussed � see for example Miller, 1975; O�Sullivan, 1978; Sims-Williams, 1983; Lapidge and Dumville (edd.), 1984; Higham, 1994; Howlett, 1998)7 and one of the arguments against Arthur�s historicity has always been that Gildas fails to mention Arthur in his reference to the battle.
It is usually countered (as Jackson 1959a) that he was deliberately omitted, either because Gildas didn�t approve of him or because his contribution to the victory was too well known, but recent work suggests that the reason Arthur was not mentioned was indeed because he was not associated with the battle when Gildas wrote. Rather than not naming anyone as the British leader at Badon, Gildas does indeed assign Badon a victor � Ambrosius Aurelianus. The idea that this figure was the true victor has been previously dismissed on the grounds that the manuscript (British Library, Cotton Vitellius A.vi) implies a major interval between Ambrosius and Badon. Dr Oliver Padel has returned to the original manuscript however and has been able to show that the break evident in Winterbottom�s edition (1978) has no manuscript authority and rather that Mount Badon now �reads naturally as the victory that crowned the career of Ambrosius Aurelianus� (Padel, 1994, pp.16-18 at p.17. For further very good reasons to doubt the attribution of Badon to Arthur see Jones, 1964; Bromwich, 1978a, p.276; Bromwich et al, 1991, pp.3-4. There seems to be good evidence for the existence of traditions about Badon which did not associate it with Arthur � see Bromwich, 1978a).
This is all, of course, of the utmost significance as it further undermines our faith in the �traditions� recorded in the Historia Brittonum � it seems very probable that in the case of Badon we are seeing a battle that had originally been fought by another leader being attributed to Arthur by the 9th-century (It is interesting to note that this conclusion has also recently been reached � apparently without knowldge of Padel�s work � by Woods (1999, pp.34-38) who, like Dr Padel, returns to the original manuscript and finds the un-edited text clearly indicating that Gildas saw Badon as
being won by Ambrosius). This tendency would appear not to be restricted to the battle of Badon � similar cases can be made for the eleventh, ninth and seventh battles (see Jackson, 1945-6; Jackson, 1949; Bromwich 1975-6 and Padel, 1994, pp.18-19). The other battles are largely unidentifiable,9 though the tenth, the �battle on the bank
of a river which is called Tribruit�, is recorded elsewhere in very early sources as a traditional battle against werewolves, thus casting further doubt on the Historia�s value; similarly a good case can be made for seeing Cat Coit Celidon in Chapter 56 as the entirely mythical battle of trees recorded in the archaic poem from the Book of
Taliesin, Kat Godeu.
Other elements within the body of Chapter 56 appear similarly suspect. For example, Prof. Hanning (1966, pp.119-20) and Prof. Charles-Edwards (1991, pp.24-5 and 28) have respectively shown that both the number of battles and the reference to Arthur as dux bellorum would seem to reflect the needs of the author of the Historia rather than any postulated earlier source. Whether or not all of the above conclusions regarding the identification of the battles are accepted it can be said, bringing all this together, that in the Historia Brittonum, our only really usable source for a �historical� Arthur, we have a text which cannot be at all relied upon to predate the 9th-century and the
contents of which can be described as being, at the very least, suspect � as such it can tell us virtually nothing certain about any possible �historical� Arthur. Indeed, the whole portrayal of Arthur in the Historia Brittonum might be seen to reflect the needs and aims of the 9th-century author rather than genuinely ancient tradition, as we
might expect given the nature of the text as a whole (see Hanning, 1966; Dumville, 1986; Charles-Edwards, 1991, pp.21-9; Dumville, 1994; Coe and Young, 1995, pp.6- 7; Howlett, 1998). The failure of the Historia as a source of information regarding any historical Arthur and the consequent intangibility of this �historical� Arthur is a fact
which has often been remarked upon: as Dr Dumville has written, �This is not the stuff of which history can be made� (1977a, p.188. See further Jackson, 1945-6; Jackson, 1959a; Jones, 1964; Bromwich, 1974-5; Dumville, 1977a; Charles-Edwards, 1991; Padel, 1994, and also Dumville�s (1994) comments on the Historia as a whole).
What then of the case for Arthur�s historicity? It should be obvious that, even when we restrict ourselves to the best sources for a �historical� Arthur, as discussed above, we can come to no solid conclusions regarding historicity. The four occurrences of the name Arthur in southern Scotland and southern Wales in the 6th- and 7th-century
cannot be seen as evidence for a historical Arthur; indeed they defy interpretation if we have a historical Arthur. The Y Gododdin reference clearly reflects a 9th- or 10thcentury (and possibly earlier) concept of Arthur as a military �superhero� but this concept of Arthur could result either from a mythical figure being used as �the
impossible comparison� or a historical figure being mythicized as a paragon of valour � thus this reference cannot help us to reach any solid conclusions. The case for a historical Arthur must therefore be based on only two sources, the Historia Brittonum and the Annales Cambriae, and neither of these can be seen as a reliable witness to historicity, both being late in date and suspect in content, with the latter very probably being derivative of the former and the former being a synthetic pseudo-history known to portray mythical figures as historical � as such, these sources cannot in any way prove that there was a historical 5th-/6th-century Arthur and no contemporary or nearcontemporary source makes any mention of him.10 The best we can say is that there existed by the 9th-century at the latest a concept of Arthur as a historical figure; our sources are simply not of the quality that would allow us to come to any firmer conclusion than this.
Against this we have to set the evidence for the existence of a concept of Arthur as a legendary figure. Whatever else we might say about it, Y Gododdin (and, it might be added, Marwnad Cynddylan) very clearly possesses a concept of Arthur as a mythical �superhero�, not a historical figure. Similarly in the Historia Brittonum, the earliest
source to portray Arthur as �historical�, Arthur appears not only in the �historical� light of Chapter 56 but also in a manifestly legendary folkloric light in Chapter 73 (an important point that is too often overlooked, particularly as the legends recorded here are considered to pre-date the 9th-century, see Bromwich and Evans, 1992, p.lxvi),
and this same concept of Arthur as a mythical hero is found in a number of other early sources, such as the 8th-century Preideu Annwfyn (Padel, 1994; Koch, 1996, pp.263- 5, etc.. See further below).
Given this, a concept of Arthur as a figure of myth and legend can be demonstrated to be present as early as (and, indeed, earlier than) a concept of Arthur as a historical figure. Here we must return to the methodological comments made at the beginning of this study. As was there noted, there are numerous examples of mythical or fictional figures being historicised, often in association with some important event of the past, and consequently �no a priori judgements can be made as to whether a figure is, in origin, historical, mythical or fictional � each individual case must (and can only) be decided by a close examination of all the relevant material.� Each of these possibilities is equally as
likely to be true, on a priori grounds, as the others; the burden of proof lies with all sides. In the absence of such proof we simply cannot assume � in the �no smoke without fire� mould � that one explanation of figures such as Arthur enjoys priority over the others: it does not. Thus whilst the above �legendary Arthur� might be the result of a historical figure being mythicised, it is at least equally as likely that, in the absence of good evidence either way, the above �historical Arthur� was a result of a legendary figure being historicised (it is perhaps worth noting with regards to this that the �process of historicising legends was a widespread feature of Celtic literary activity in the Middle Ages� (Padel, 1994, p.23)).
Hence in answer to the question �Was there a historical Arthur?�, the sources being questioned (i.e. the Historia Brittonum and the Annales Cambriae) can only answer �perhaps, maybe� � they cannot say �no there wasn�t� for obvious reasons but equally they cannot say �yes there was�: the nature and quality of the sources for a �historical� Arthur is quite simply such that they neither show nor demand a historical figure to lie behind them and we most definitely cannot assume one in the absence of this. Whilst it is possible that Chapter 56 of the Historia reflects, to some extent, the distorted but genuine traditions of a �historical Arthur�, it is at least equally as likely, given the nature of our sources, their claims to reliability and the fact that a concept of Arthur as a mythical hero existed from at least the 8th-century, that the opposite is true and that these references simply reflect a legendary figure (such as that of Chapter 73 of the Historia) historicised by the 9th-century. Arthur could well be a mythical figure
portrayed as historical by the author of the Historia Brittonum in just the same way as Hengest and Horsa were mythical figures portrayed as historical by both Bede and the author of the Historia.
In the absence of a priori assumptions regarding historicity, a detailed investigation of the �relevant material� (as required by the above methodology) has left us with a situation in which the information contained within these late references could still reflect either a historical figure or a legendary figure historicised with no convincing reason, from the internal evidence of these few sources, for accepting one alternative over the other. To put it another way, there is no obvious reason from the material discussed above to prefer the portrayal of Arthur in Chapter 56 of the 9th-century Historia Brittonum over that in Chapter 73, or vice versa.
Part of the problem, of course, lies with methodology. When the case for a historical 5th-/6th-century Arthur is made, it involves trawling the pre-Galfridian source material for anything that might be used to back it up. The interest is not with the pre- Galfridian material itself and with what it tells us but rather with what it can tell us about a possibly historical figure called Arthur. The texts selected to answer this question, as in the above analysis, are thus divorced from the context of the whole body of pre-Galfridian material in which they must surely be viewed and of which they form an integral part. By asking �Was there a historical Arthur?� one forces the texts to answer �perhaps, maybe�; they have no other choice because, on the basis of the few sources selected and the viewing of these few sources in isolation, they are incapable of denying that there was such a figure just as they are incapable of
As such this �perhaps, maybe� is in reality valueless. What this means is that conclusions regarding Arthur�s historicity can and should only be drawn via a sound methodology, namely by looking at all the available evidence and allowing it to �lead�, not forcing it to conform to preconceived notions. The Historia Brittonum and Annales Cambriae references must be seen in the context of all the early Arthurian material, not as discrete pieces of information that can be mined for �facts�. No judgements of any value can be made by attacking the pre-Galfridian corpus in a
piecemeal fashion � one has to look at the weight of the body of evidence as a whole. To quote Dr Padel, �the nature of the inquiry, which hitherto has always started with the natural question �was there a historical Arthur?�, has determined its outcome (�Yes, perhaps�)� (Padel, 1994, p.2. Ashe (1995) also makes this point).
By commencing an examination of the pre-Galfridian material with a view to discovering (or, at least, investigating) a truly historical figure of the post-Roman period the conclusions reached are unavoidably biased and the investigation ignores the majority of the available early evidence. We must therefore ask, what is the nature of Arthur in the pre-Galfridian sources with which we are here primarily concerned with? Where does the �weight� of the evidence
�lead� us? What is the context of the �historical� sources? The most recent attempt to define this �nature� (which then proceeds, after doing this, to adopt the above methodology and look at the Annales and Historia references in the context of this nature) is by Oliver Padel of Cambridge University (in Cambrian Medieval Celtic Studies 27 (Summer 1994) pp.1-31.
We cannot possibly do justice to Padel�s detailed analysis in a short summary account and only the main conclusions are here reproduced � those interested in the details and wishing to see the evidence on which these conclusions are based should most definitely read his article, along with the other studies listed in the bibliography at the end of this article). The conclusion reached is that, when the pre-Galfridian sources are approached without such preconceived agendas and a priori assumptions as described above, the results prove to be most interesting: �if the collective evidence is first allowed to speak for itself, its weight is quite different.� (Padel, 1994, p.2). In non-Galfridian tradition, Arthur was very clearly �the leader of a band of heroes who live outside of society, whose main world is one of magical animals, giants and other wonderful happenings, located in the wild parts of the landscape.� (ibid., p.14); Arthur is portrayed as a figure of pan-Brittonic folklore and mythology, associated with the Otherworld, supernatural
enemies and superhuman deeds, not history. This concept of Arthur occurs in both the very earliest of these sources (earlier than and contemporary with the earliest references to a possibly �historical Arthur�) and, indeed, in the vast majority of the non-Galfridian sources, with these sources consistent in their portrayal of Arthur.
For example, it appears in the 8th-century or earlier mythological poem Preideu Annwfyn (see Koch, 1996, pp.263-5), the very early mythological poem Kat Godeu (see Ford, 1977 for a translation), Chapter 73 of the Historia Brittonum (the folklore contained in which is considered to be �already ancient by the ninth century� (Bromwich and Evans 1992, p.lxvi)), Pa gur yv y porthaur? (which might be as early as the 9thcentury, or even the 8th, and is, itself, simply a summary of many earlier entirely mythical Arthurian tales (Bromwich, 1978b, p.21; Koch, 1996; Koch, 1994, p.1127; Edel, 1983)), and Culhwch ac Olwen (which was written in the 11th-century but is a literary composition based on a number of earlier legendary Arthurian tales brought together with the �giant�s daughter� folklore tale-type � the Arthurian material is generally considered to represent the same body of very early non-historical tales as
Pa gur yv y porthaur?, Historia Brittonum Chapter 73 and Preideu Annwfyn do: see Edel, 1983; Bromwich and Evans, 1992). Padel is not at all alone in seeing this as the context of the Historia Brittonum and Annales Cambriae references, though he has given the subject its fullest treatment.
Two of the foremost authorities on early Arthurian literature, Dr Rachel Bromwich and Professor D. Simon Evans, have recently written that �Arthur was above all else... a defender of his country against every kind of danger, both internal and external: a slayer of giants and witches, a hunter of monstrous animals � giant boars, a savage cat monster, a winged serpent (or dragon) � and also, as it appears from Culhwch and Preiddeu Annwn, a releaser of prisoners. This concept of Arthur is substantiated from all the early sources: the poems Pa Gur and Preiddeu Annwn, the Triads, the Saints Lives, and the Mirabilia attached to the Historia Brittonum... in early literature he belongs, like Fionn, to the realm of mythology rather than to that of history.� (Bromwich and Evans (edd.) 1992, pp. xxviii-xxix. See Ford (1983) for some very interesting supplementary evidence for the view that the pre-Galfridian Arthur belongs to the realms of mythology. The above comments on the �nature of Arthur� in early literature represent the general view among Celticists of this question, see for example Ford 1986; Jarman, 1983; Ross, 2001, chapter 4; and note 16 below).
In essence, the vast majority of the non-Galfridian material, including the earliest sources, paints a notably consistent picture of Arthur as a pan-Brittonic folkloric hero, a peerless warrior of giant-like stature who leads a band of superhuman heroes that roam the wild places of the landscape, who raids the Otherworld whilst being intimately associated with it, who fights and protects Britain from supernatural enemies, who hunts wondrous animals and who takes part in mythical battles, and hence the �weight� of this evidence indicates not a historical origin for Arthur but
rather a legendary one (it is particularly worthy of note that Arthur is never associated with either the Saxons or Badon in the vast majority of the material, despite the fact that such an association is usually said to be the reason for his fame, and when this association does appear it is only present in those sources which are directly derivative
of Historia Brittonum Chapter 56). In fact, the Fionn parallel in the above quote is also noted by Dr Padel in his article � it is his convincing conclusion that the nature of Arthur evidenced in the pre-Galfridian sources is very similar indeed to the nature of Fionn in Gaelic literature, this Fionn being an entirely mythical character (originally a god) who became associated (i.e. historicised) with the repelling of the Viking invasions of Ireland and who had a list of battles against his �foes� attached to his name (for Fionn see � h�g�in, 1988; Padel (1994) summarises some of the parallels
on pp.19-23). Professor Van Hamel made some very similar observations regarding the nature of Arthur in the early sources and the very close parallels between him and Fionn, noting that it was but a natural, logical step �to represent a hero of this type [i.e. a protector of Britain against supernatural threats] as a victor over the Saxons� (1934,
quote at p.231. See also Murphy, 1953, pp.213-17; MacKillop, 1986, pp.63-4; Koch, 1996, p.261; Ross, 2001, chapter 4).
How does this affect the question of Arthur�s historicity? What then of those references to a �historical� Arthur which, when viewed in isolation, can only answer the question �Was there a historical Arthur?� with �perhaps; maybe� and could at least just as easily represent a legendary figure historicised as the distorted remembrances of a �genuinely� historical figure? To recapitulate, the conclusions resulting from the above discussion are:
(A) that one cannot assume that a character is historical simply because a medieval source claims that this is the case: such a priori assumptions are demonstrably false (Hengest & Horsa and Fionn being good examples of mythical figures historicised by later writers) and are thus unacceptable. One can only say that there was / has to have
been a historical Arthur once all the material has been evaluated and this is shown to be the case. There is no possible justification for simply assuming this to be the case � �historical� explanations of figures such as Arthur do not, on a priori grounds, enjoy priority over other explanations. Indeed, it should be remembered that the �process of
historicising legends was a widespread feature of Celtic literary activity in the Middle Ages.� (Padel, 1994, p.23).
(B) that the few usable sources that we have which portray Arthur as �historical� could very easily represent either a legendary figure historicised or the distorted traditions of a genuinely historical Arthur. Each possibility is equally as likely as the other judging from the internal evidence of the sources and, as such, no conclusions can be reached on the matter of historicity � there may have been a historical Arthur but at least equally as well there may not have been.
(C) that whilst it is true to say, as in (B) above, that Historia Brittonum Chapter 56 etc. could just as easily reflect a legendary figure historicised as a genuinely historical personage, this method of analysis fails to answer the question of Arthur�s historicity satisfactorily. By treating the �historical Arthur� sources in isolation rather than in the context of the whole body of non-Galfridian Arthurian literature of which they form an integral part, valuable information is ignored that is essential to the interpretation of these sources and, as such, no conclusions of any value can be drawn. To give an example, we might have a charter purporting to be a grant of land to a monastery from a king.
When this charter is viewed on its own the evidence internal to the charter may be such that no decision can be made over whether it is genuine or a forgery � in the absence of convincing evidence for either option each possibility might be said to be equally as likely. If, however, this charter is looked at in the context of all the other charters from that monastery then the situation is rather different: thus if, for example, all the other charters from that monastery appear to be forgeries then it seems very likely indeed that this charter too is a forgery. In the context of the body of material of which it forms an integral and inseparable part, it becomes clear that the two possibilities allowed by the internal evidence are not in fact equally as likely � when viewed in light of all the other material it remains remotely possible that the charter may be genuine but it is infinitely more probable that it is a forgery.
In other words,
the serious possibility that the charter is genuine only really existed because the charter was being analysed outside of the body of material of which it is an integral part, something which caused information essential to the interpretation of the charter to be ignored � when it is viewed within the context of all the material, there is simply
no reason to think that it might be genuine; the charter�s context is such that this is not, in the absence of evidence in its favour, a serious possibility. In the same way, conclusions regarding historicity can only be drawn from looking at the �historical Arthur� texts in the context of the whole body of early material. The Historia Brittonum and Annales Cambriae references must be seen in the context of all the early Arthurian material, not as discrete pieces of information that can be mined for �facts�; no judgements of any value can be made by attacking the pre-Galfridian
corpus in a piecemeal fashion � one has to look at the weight of the body of evidence as a whole and allow it to �lead�. To do otherwise simply biases the conclusions and ignores the vast majority of the available early evidence.
(D) that the weight of the non-Galfridian material (early and late) provides, as numerous scholars have noted, a very clear and consistent picture of Arthur as a thoroughly legendary figure of folklore and myth not associated in any way with either the Saxons or Badon, and with this figure resembling in many of its characteristics (and, indeed the development of its legend) the Gaelic Fionn who was a mythical figure � originally a god � later historicised with battles against foreign invaders.
These four relatively uncontroversial conclusions have, as should be obvious, some very interesting consequences for the question of Arthur�s historicity. Following them through, it seems clear that if those few references which portray Arthur as historical are seen in the context of the material as a whole � as they have to be � then the
weight of the material is such that there is absolutely no justification for believing there to have been a historical figure of the 5th- or 6th-century named Arthur who is the basis for all later legends. When the �historical� references are pulled out of their context and viewed in isolation then, as we have seen, they may possibly represent the
distorted traditions of a historical figure but at least equally as well they may not. However, when they are viewed, as they must be, in the context of the body of material of which they are an integral part this �maybe� evaporates. All the other evidence, the vast majority of the early material, portrays Arthur as an entirely legendary figure from the same mould as the Gaelic Fionn, and he is never connected in this material in any way with either the Saxons or Badon.
As such there is simply no reason to think that there was a historical Arthur. The �maybe� only appears when it is forced to, when the few references to a �historical� Arthur are divorced from their context and made to answer questions regarding the possibility of a historical Arthur. If we ask what the material actually says rather than try and force any preconceived notions upon it then it appears, as Dr Padel has observed, to very clearly tell of a
legendary figure of folklore named Art(h)ur who was historicised in much the same way as Hengest or Fionn were � the serious possibility of there ever having been a �historical Arthur� who was the �original� from whom all the later tales spring is simply a construct based on a misuse of the sources.
Therefore, rather than the folkloric Arthur evidenced in the Historia Brittonum Chapter 73 being an elaboration
of the �historical� Arthur of Chapter 56, this �legendary� Arthur would appear to be �the true one, and the �historical� Arthur... the secondary development.� (Padel, 1994, p.30), a logical extension of his folkloric role, with not only the existence of Arthur but also his association with the fifth and sixth centuries being seen as most probably
spurious (with regards to this, it should be noted that the post-Roman period was not the only period into which Arthur was historicised � see below). To put it another way, the context of the few �historical� references is such that the onus of proof would seem to come to lie firmly on the shoulders of those who would have a historical 5th-
/6th-century Arthur as the basis for all the later legends � in the absence of proof of historicity (and in the absence of a priori assumptions and the forcing of preconceived agendas onto the sources) there is simply no reason to think that a �historical Arthur� is a serious possibility.
We must consequently ask, can the �evidence� for a �historical� Arthur of the 5th-/6thcentury live up to this burden of proof? Does it provide any reason to believe that there was a 5th- or 6th-century figure named Arthur? Taken on its own, it can be legitimately said that the answer to this is �no�. Even when viewed outside of the context of the whole body of early material, thus in the most advantageous circumstances, it could (as has been seen above) only produce the answer �perhaps; maybe�; the Arthur portrayed in the Historia Brittonum and the Annales Cambriae
could be easily understood as either a historical figure or a legendary one historicised. In the context of the pre-Galfridian material this answer becomes meaningless due to the shifting of the burden of proof � as such the �maybe� has to be taken as a �no�. The Historia and the Annales do not provide the necessary proof that would allow us to
disregard the context of the pre-Galfridian material (particularly as the latter is very probably derivative of the former, and the former is known to portray mythical figures as historical) and thus on the basis of these pieces of evidence we are forced to conclude that there is, at present, no cogent reason to think that there was a historical
Instead he is best seen, like Fionn for the Gaelic regions, as a folkloric hero, living in the wilds of the landscape and protecting Britain from all kinds of supernatural threats, just as the vast majority of the evidence suggests.16
Indeed it is worth pointing out once more that the Historia Brittonum�s account of Arthur in Chapter 56 not only appears to include deeds of a number of earlier warriors such as Urien of Rheged and Ambrosius Aurelianus, but also identifiable mythical elements which have been historicised in this text � the possibly very early poem Kat
Godeu would appear to be concerned with a mythical battle in which Arthur plays some (perhaps major) part and in which the trees of Coed Celyddon are magically animated to fight, thus showing the battle of Coit Celidon (�the Caledonian Forest�) recorded in Chapter 56 of the Historia Brittonum in a very interesting light.
Similarly, the �battle on the bank of a river which is called Tribruit� in Chapter 56 of the Historia appears elsewhere, in the early Pa gur yv y porthaur? (which summarises a number of pre-existing Arthurian tales) as an entirely mythical battle against werewolves (With regards to the battles named in the Historia Brittonum, it should perhaps be emphasised that there is no reason to think that all of the battles used to historicise Arthur were real historical battles � at least some of the battles used to historicise Fionn seem to have been invented spontaneously for the purposes of historicisation and this could well be the case here (a fact that may well explain some of the problems in identifying the battles in Historia Brittonum Chapter 56, see Padel, 1994, p.21; Jackson, 1945-6)).
The above conclusions may well help explain certain puzzling features of the Arthurian legend, in particular the strange absence which has often been noted (e.g. Bromwich, 1978a, p.274; Thomas, 1995, p.389) of Arthur from the early Welsh genealogies. Such texts are perhaps best understood as dynastic �propaganda� (see Dumville, 1977a; 1977b) and if Arthur was generally held to have been a great historical leader at the time of their compilation, his absence would be very puzzling; if, on the other hand, he was not viewed in this light but instead as a pan-Brittonic
folkloric hero then his absence is entirely comprehensible (see Gowans, 1988 for a similar situation involving Cei). This notion, of a reluctance to use the name of a national folkloric hero, can also provide the only viable explanation of one of the first pieces of evidence examined here, that is the four (or five) occurrences of the name Arthur in 6th- and 7th-century contexts, as Dr Padel has recently noted (1994, p.24).
Padel observes, as others have done before him, that all the occurrences of the name �Arthur� are recorded in Gaelic sources and occur in the context of the Irish settlers in western Wales and Scotland (see Bromwich 1975-6; Barber, 1972) and he suggests that the absence of this name in British contexts is due to Arthur being regarded �with exceptional awe� as a legendary hero and Protector of Britain, whilst the Irish �when they came into contact with the folklore as a result of their settlements in western Britain, need not have felt such reverence or reluctance�(Padel, 1994, p.24) and consequently they made use of this name (the date of adoption of this name would, of course, be dependent on complex cultural interactions and developments and thus the fact that it was not immediately adopted should not be seen as significant).
As well as explaining satisfactorily all the available evidence this suggestion gains a considerable amount of credence from the fact that detailed study of the Welsh genealogical tracts reveals that not one single person of British descent in Wales bore the name �Arthur� in the genealogies until the late 16th-century at the earliest, a situation Bartrum suggests may well be because the name carried some sort of superstition with it (Bartrum, 1965). If Arthur was to be viewed as historical rather than legendary, then explanation of these three pieces of information (the absence of Arthur from the early royal genealogies; the sudden occurrence of four people named Arthur in the context
of the Irish settlers in Wales and Scotland; the fact that not one single person in Wales of British descent can be shown to bear the name Arthur until at least the late 16th century) would be a very difficult problem.
Another �puzzling� feature particularly worthy of note is the fact that, outside of the Historia Brittonum Chapter 56, the Annales Cambriae (which is derivative of the Historia Brittonum), the possibly 11th-century Breton Life of Saint Goueznou (which paraphrases the Historia Brittonum) and William of Malmesbury�s 12th-century Gesta Regum (which again paraphrases the Historia Brittonum and the Annales Cambriae), Arthur is never associated in the whole body of pre-Galfridian literature with the post-Roman defeat of the Saxons � a very strange situation surely for one who is supposed to be famed because of such an association. However, it fits with the fact that there seems to be good reason to believe that there was a separate non-Arthurian tradition regarding the battle of Badon (which, again, is only ever associated with Arthur in the few sources (above) that are directly derivative of the Historia Brittonum � in sources that are not connected with the Historia, Badon is not linked with Arthur nor is Arthur linked with Badon, see Bromwich, 1978a), the single event which puts Arthur�s supposed victories into the realms of history and which, in
essence, defines his role as defeater of the Saxons.
Both of these features, especially when taken together, appear highly suggestive. One has to ask, why, if the reason that Arthur was so honoured in Welsh tradition was that he led the British resistance and won the famous battle of Badon, this is ignored and even perhaps doubted by the �guardians of Welsh tradition�? Why, in the vast majority of cases, both early and late, did they instead paint a consistent picture of Arthur as a figure of folklore who was very similar indeed to the Gaelic Fionn, an entirely mythical figure who came to be historicised with great battles against the Viking invaders of Ireland? Indeed, one might further ask why, if Arthur was universally famous not for being a folkloric Protector of Britain but rather the defeater of the Saxons, the Cornish felt perfectly able to totally ignore his Saxon associations and instead historicise him into distant antiquity and into the period of the Viking incursions (see Hunt, 1881; Courtney, 1890)?
Such considerations as those above, quite apart from the fact that the adoption of a sound methodology forces us to conclude that Arthur was in all probability a folkloric �Protector of Britain�, suggest that such an interpretation is the correct one. A historical 5th- or 6th-century Arthur is not in anyway necessary to the understanding of the pre-Galfridian Arthur and the evidence we have makes the postulation of such a figure not only unnecessary but also completely unjustifiable.
The Historicisation of Arthur
Whatever else Arthur is, he is a composite figure. Through the centuries the concept of Arthur did not stay the same � there is no �standard� Arthurian legend as this legend is the result of Arthur attracting to himself both the deeds and characteristics of other tales and characters. This bears directly on the above question � we cannot
conclude that there was no historical Arthur as there was, to the extent that certain texts, notably the Historia Brittonum, the Annales Cambriae and Geoffrey�s Historia Regum Britanniae, have a concept of Arthur that is clearly historical. While the Arthur they portray cannot be seen, in light of the above, as the �original�, it is surely still a valuable exercise to inquire as to whose deeds were being later attributed to Arthur, as these deeds are an integral part of many later portrayals of Arthur and as such do constitute part of the origins of Arthur. What then of the Arthur of Historia Brittonum Chapter 56? While we might legitimately look for an �original� for each of the battles, we also have to ask whether the whole concept presented in Chapter 56 of the Historia is based on a single figure.
The prime candidate for this �honour� has to be, naturally, Ambrosius Aurelianus. In Gildas�s De Excidio Britanniae Ambrosius is given prominence as the initiator of the British counterattack which, after the fighting of several battles, culminates in the battle of Badon, just as Arthur in the Historia Brittonum initiates the British counterattack which, after the fighting of several battles, culminates in the battle of Badon. On the basis of this we may well be able to say that, to some extent, we do have a historical Arthur � Ambrosius � in the sense that the concept of Arthur as a
historical figure and the framework for historicisation was based on his deeds.
With regards to the individual battles, this is perhaps more difficult. As noted in the preceding discussion, the �battle on the bank of a river which is called Tribruit� and Cat Coit Celidon may well be actual Arthurian mythic battles. Others may be �real� or they could be invented: Badon, as has been argued above, can be easily associated with Ambrosius, just like the whole framework of historicisation, and Breguion appears elsewhere in very early sources as a battle fought by Urien of Rheged. Others however could simply be made up, as is thought to be the case for the battles used to historicise Fionn in his battle-list and as has been suggested earlier in this study. The problem with undertaking any exercise of this kind is the fact that the names given to the battles could represent many areas � only a few can actually be called certain and on the basis of this list theories of a Southern Arthur, a Midland Arthur and a
Northern Arthur have all been constructed.
A partial solution is to split the list up into separate characters as above but it should be remembered that it can only be taken so far. The desire to identify these battles is often great but this should not prevent us from recognising that with sufficient �ingenuity� they can be made to fit just about any area and many may not, in fact, be identifiable or even have identifications. With regards to the whole question of historicity and historicisation, it has been
suggested that, rather than ask whether there is any justification for postulating a historical Arthur, we should ask whether any candidate fits the �facts� � certainly the undert