Many students of Arthuriana, amateur and professional, have delved into the "mystery" of the Arthur name, only to find out what has long been suspected or known: it comes from the Roman/Latin Artorius.
Another point being made about this name is, however, not anywhere near as convincing. For we have been assuming all along that the Artorius was preserved among the British because it was made famous by the camp prefect of that name at York. We know about him only because of the chance find of some fragmentary inscriptions.
A related assumption has been that the name would have been passed down among a royal family that traced its descent from the sub-Roman rulers of York. In other words, the famous Dark Age Arthur must belong somewhere in the North and not in the South.
The second assumption ignores two possibilities. First, that there may well have been another famous Roman Artorius elsewhere in Britain for whom no record is extant. Or, second, Artorius of York may have been so famous in Britain that the reuse of his name need not be restricted to the North. If a royal family in the South knew of him as well, then they might have borrowed the name for one of their own princely sons. Especially if a) bear-names were already present in their pedigree b) they worshiped a bear deity or c) they had a semi-divine ancestor with a bear-name.
For a couple of decades now, I've been looking for "respectable" historical Arthurian candidates. But in all this time, I've only managed to find two valid contenders.
One must be sought in the Irthing ("Little Bear") Valley between Cumbria and Northumbria. This is the most probable location of Arthwys, the 'man of the Arth', an eponym like the Glewys of Wales (from Glevensis, 'man of Glevum/Gloucester'). The river valley contains the Camboglanna Roman fort and the sub-Roman hall at Birdoswald. Not too many miles to the west is the Aballava/Avalana (Avalon?) Roman fort. The son of this Arthwys is called Ceidio, a hypocoristic form of a name which almost certainly would have conveyed a meaning identical to the "dux erat bellorum" descriptor applied to Arthur in Nennius. Most of the battles from the HISTORIA BRITTONUM list can be found running north and south of Hadrian's Wall along the Roman road called Dere Street. This arrangement of Arthur's martial activity strongly suggests that Dere Street was the frontier zone between the British and the English. This Arthur may well have gotten his name from the claimed connection of his family with that of York. Any Irish blood (necessary to account for the subsequent Arthurs, all of whom belong to Irish-descended dynasties in Britain) would have to derive from the Dalriadans, whose founder Fergus Mor is present in the relevant genealogies. These genealogies, of course, are not entirely reliable. They were doubtless manipulated and there are some authorities who believe they are outright fictions. In my book THE ARTHUR OF HISTORY, I successfully showed that a line of descent was often used to demonstrate a relationship between neighboring kingdoms or traditional tribal territories.
The second candidate was discussed in great detail in my second book, THE BEAR KING. While the idea has been around forever that Cerdic (the ealdorman = dux erat bellorum?) of the Gewissei might be Arthur, it was not until I was able to make a very strong argument in favor of Cerdic = Ceredig son of Cunedda that it occurred to me the notion may be correct. Ceredig was founder of a kingdom that embraced the Arth or "Bear" River, and his immediate descendants had a predilection for Arto- or 'bear' names. It didn't take much to recognize this "coincidence" and to propose that Arthur/Artorius was a decknamen used to replace an earlier Irish or British 'bear-king' title or name. The Arthurian battles of the HB seemed to be susceptible to comparative analysis with the battles of Cerdic and other Gewissei leaders as these were found arranged in annalistic form in the ANGLO-SAXON CHRONICLE. A few other stray hints seemed to lend weight to the thesis (e.g. that the name Arthur provided the impetus for the story of the lame son of Elafius in the VITA of St. Germanus of Auxerre). The weakness (or strength, depending on how one looks at this!) of the Cerdic/Ceredig = Arthur theory is that one must accept all the Arthurian battles of the HB as Welsh attempts to translate the place-names supplied by the ASC or to substitute for them approximate Welsh place-names. If this were done, and the name of Ceredig/Cerdic dropped in favor of Arthur, we must ask, simply, "WHY?" As the HB was written in NW Wales (where Cunedda and his sons - including Ceredig - had settled, and from whom all the prevailing princely houses ultimately traced their origins), a propagandist motive is likely present. Cerdic/Ceredig was an Irishman or Hiberno-British mercenary who fought with the English against British enemies of the High King of Wales. Arthur, on the other hand, was represented as being a thoroughly British champion who fought against the English. A name-change, then, might have been in order, and the battles of the ASC were cleverly disguised so that they did not resemble those supplied by an early English source document.
Given these two candidates, which is the real Arthur?
My hope was that the question could be decided by properly identifying Uther Pendragon. Yet even this accomplishment proved unfruitful, for Uther was none other than St. Illtud and Illtud was not Arthur's father! He was only made Arthur's father by Geoffrey of Monmouth or his source. Uther himself was confused (accidentally or intentionally) with the mil uathmar of the North, a name conjured in the Irish Compert Mongan to account for a corrupt form/spelling of the Degsastan/Dawston place-name in Liddesdale. Ironically, Liddesdale is said to have been ruled over by Gwenddolau (perhaps a place-name, 'White Dales') son of Ceidio son of Arthwys.
My very cautious interpretation of all this is as follows: the Welsh needed a counterpart to Cerdic of the Gewissei, reputed founder of the English kingdom of Wessex. The Northern Arthur was at hand, perhaps with a list of battle victories. Or these victories may have been at least in part fabricated (or borrowed from other Northern heroes, like Urien of Rheged, who is also said to have fought at Brewyn/Breguoin/Bremenium, and perhaps the later Arthur of Dalriada) to balance out those of the Gewissei.
An alternative reading is not nearly as attractive, but cannot be easily dismissed: Cerdic, seen as an English hero, was converted into a purely British hero by changing his name (or giving him his other name/title) and being supplied with a list a battle sites derived from a Welsh rendering of the Gewissei sites.
Until I uncover more evidence - or develop a better argument - this is as far as I can go in my surmises.