Sunday, October 21, 2018


Uxellodunum/Stanwix Roman Fort

Years ago, I looked at the Arthur Penuchel of a corrupt version of one of the Welsh Triads.  While it is easy to dispense with such a personage - precisely because his name and title do seem to be corruptions of the original - what I found to be of profound interest was that in this particular corruption the Arthur in question was made a son of Eliffer/Eleutherius of York.

Why of profound interest?  Because the only Roman period Artorius we know of was the camp prefect of York.  This is the well-known Lucius Artorius Castus, whom Dr. Linda Malcor has made much of.  While Dr. Malcor and I disagree quite strongly on LAC, as he is called, in so far as I believe he merely supplied his name to the North Britons and she subscribes to the notion (strictly from the standpoint of her expertise in the field of folklore) that much of the Arthurian legend should be traced to LAC and not to a 6th century personage), I cannot shake the feeling that Arthur must somehow belong to the North.

At one point, I had dared suggest that the reference to Arthur Penuchel was not, in fact, actually a corruption, but instead a correction of the original text. Not even a substitution, but an acknowledgement of the fact that there was a Northern Arthur of the 6th century who was somehow connected to the royal sub-Roman house of York.

My reasoning was pretty simple, even if it was based on a paucity of data.  Any legitimate Arthur candidate should have some stated association with York.  Otherwise, we were forced to try and explain why the rare name Artorius would have been given to a man who had no connection, no matter how distant, with LAC.  I could not at the time - and still cannot - adequately apply the name Arthur to someone in the south or extreme west.  Someone had to have family descent or presumed or desired family descent from LAC in order to bear the name.  This much was plain to me.  When I'm honest with myself, I still cannot account for an Artorius in the South.  I can imagine a sub-Roman Artorius in the North - but only with some caveats.

Before I go further, I'm pasting here all of Rachel Bromwich's relevant passages on Arthur Penuchel. This should be read carefully, and considered as authoritative.  Following her analysis of the corruption problem, I will delve more deeply in the genealogies of the Men of the North and try to come to some conclusion that might further our search for a legitimate Northern Arthur candidate.  Also provided here are some entries from P.C. Bartram's A CLASSICAL WELSH DICTIONARY.


 She is mentioned in Bonedd y Saint as the wife of Brochwel Ysgithrog and mother of St.Tysilio (§33 in EWGT p.59). The cognomen, Penasgell, ‘wing-headed’, occurs only in a minority of manuscripts. Compare Ceindrech ferch Eliffer Gosgorddfawr. It is said that Dolarddun, a former township in the parish of Castell Caereinion (WATU) was named after her. (MA2 417, LBS I.168).


She and her two brothers, Gwrgi and Peredur, constituted one of the 'Three Fair Womb Burdens’
of Ynys Prydain according to a triad (TYP no.70, Pen.50 version). The Pen.45 version
substitutes Arddun. Compare Arddun Benasgell.

PABO POST PRYDYN. (450) [of Papcastle/Deventio Roman fort in Cumbria; note that an attempt was made to connect him to Arthwys as well]

‘P. Pillar of Pictland’. He seems to have been a famous hero of North Britain though little is
now recorded of him, and he is mentioned chiefly as the father of Dunod Fwr, Sawyl Benisel, Cerwydd and Arddun Benasgell, the wife of Brochwel Ysgithrog. See the names. The earliest genealogical sources make him son of Ceneu ap Coel Hen (HG 11, 19, JC 38 in EWGT pp.11, 12, 48). However ‘Bonedd Gwŷr y Gogledd’ makes him son of Arthwys ap Mar ap Ceneu ap Coel (§4 in EWGT p.73). This longer version was copied in late versions of the ‘Hanesyn Hen’ tract (ByA §12 in EWGT p.88). There is little doubt that the earlier version is more correct, being chronologically more satisfactory. Compare Pabo, St.

I wish to now present what might be termed the "genealogical evidence" for Arthur Penuchel - although, I should hasten to add that the scholarly consensus regarding the early Welsh genealogies is that they are probably manufactured and thus utterly unreliable.  Yet in some cases they may preserve accurate historical traditions.  Whether these traditions are themselves of true historical value is debatable.

The following snippets from family trees (see P.C. Bartram for confirmation of details) are arranged vertically.  I have elsewhere proven that 'Mar' is a known variant of 'Mor', and so when we encounter Mar in the genealogies we know this is, in fact, Fergus Mor, the founder of Irish Dalriada in Scotland.

Gwrwst Ledlum (Fergus Mor)                Mar (=Mor, Fergus)                Mar (=Mor, Fergus)

                 Eliffer                                               Arthwys                                   Arthwys

          Arthur Penuchel                                       Ceidio                                       Eliffer

                                                                                                                    Arthur Penuchel

It will be noticed immediately that one ancestry trace makes Arthur Penuchel of the same generation as Ceidio.  In my book THE ARTHUR OF HISTORY (only now posted here to this blog site in its entirety), I made my case for Ceidio being Arthur, as his name is a shortened form of what would almost certainly translate into the 'dux erat bellorum' of Nennius's HISTORIA BRITTONUM [1].  An Arthur Penuchel son of Eliffer son of Arthwys places Arthur a generation too late to make the chronology work. Ceidio is fascinating in another respect, as he is father to Gwenddolau ('white dales', probably a place-name recorded at Carwinley).  Gwenddolau is the lord of Myrddin/Merlin.

Arthwys is 'man of the Arth', an eponym similar to Glywys ('man of Glevensis', i.e. of Glevum/Gloucester).  I had made a case for this place being the valley of the Irthing, where we find Banna and Camboglanna.  However, Dr. Andrew Breeze's etymology for Irthing derives it from a Cumbric word meaning 'little bear.'  This involves i-affection, as a diminutive suffix -yn (Old Welsh -inn) allows arth- to become erth-.  It is hard to hold to the idea that the man of the Arth/Bear had left his name in a river called the Little Bear.

Instead, I now propose a somewhat different etymology for the Irthing.  Could this not, I asked Brythonic place-name expert Alan James, be simply an Arth river to which English -ing (often added to river-names; see Ekwall) has been added?  His response:

"As far as English is concerned, note that a rather similar sound-change, generally called i-mutation or (as in German) umlaut, would have made *arthing > *ærthing, which would normally have reverted to *arthing in ME, but I think could possibly emerge as *erthing."

If this happened, then we could have had an original arth at the Irthing, rather than erthinn/erthyn.  Arthwys would belong to the valley of the Irthing.[2]

Arthwys, of course, can simply be a territorial designation.  He need not be a real man.  A son of his - or merely a chieftain originating from the valley of the Arth River - might well take on a name that was believed to contain an arth- element.  One like Arthur/Artorius, for example.  I've several times pointed out that noted Celticist Stefan Zimmer has provided an excellent treatment of Artorius was itself deriving from an ancient Celtic 'Bear-king' name.  Given the connection York had with Hadrian's Wall (the former was the headquarters of the governor of Northern Britain, and there was a unique relationship between York and Stanwix/Uxellodunum [3] at the west end of the Wall), it would not be surprising for the name of a famous camp prefect at York, passed along through generations of Britons who like to claim descent from the Romans, to be given to a man born on the Wall at the Bear River.  

In THE ARTHUR OF HISTORY, I suggested that Arthur's base of operations might have been Stanwix, the site of the largest Roman cavalry unit in all of Britain.  This was based on a number of factors.  First, a tradition that the place was called 'Arthur's burg.'  Second, I had wondered if there could be a connection between the Uxellodunum name and Arthur's Penuchel title (as Welsh uchel derives from British Uxello-).  Finally, the Petriana unit of the fort reminded me of the Dyfed Arthur, whose father was one Petrus.  Admittedly, these could all simply be coincidental resemblances.  Yet if Arthur really was as famous as he is made out to be in later tradition, there would be no better place for him to have ruled from that a location which in the Roman period, at least, was the headquarters of the Wall.  Uxellodunum was also pretty much exactly between Camboglanna and Aballava (Avalan/"Avalon"), with the Carwinley of Gwenddolau son of Ceidio just a little to the north.  Finally, there is the known strong link between York of the Roman governor and Stanwix - and it is from York that the name Arthur seems to have come.

I don't really have anything to add to this outline of a possible Northern Arthur.  It is at least as good as declaring Cerdic of Wessex/Ceredig son of Cunedda to be a Southern Arthur.

[1] “Ceidiaw is a 'pet' form of a name in *katu- 'battle' with the common hypocoristic ending -iaw (> Mod. Welsh -(i)o) found in Teilo (Old Welsh Teliau) etc., and still productive today (Jaco, Ianto etc.). And yes, it's not possible to say what the second element would have been. But the forms you suggest [Cadwaladr, (“Catu-walatros) ‘Battle-leader’, Caderyn (Catu-tigernos), ‘Battle-lord’, Cadfael (Catu-maglos), ‘Battle-prince’, Caturix (a Gaulish god), ‘Battle-king’] are among the candidates, especially as this man was a chieftain of Y Gogledd [the North] at the head of some of the royal genealogies. ” - Dr. Simon Rodway, The University of Wales

[2] The River Irt in Cumbria could have a "bear" etymology as well.  According to Alan James: "...a minority of early records do have Ert(h)(e), but I don't think a plural eirth, 'bears', is likely in a river-name. A fossilised genitive - *avon eirth - might be more plausible, though this would require rather special pleading."  Note there are other proposed etymologies for the Irt.  None are very satisfactory.  Once again from James:

"Ekwall (ERN 211) points out that irt happens to be a Middle Irish word for 'death'. I hadn't noticed that before, it would presumably be related to Welsh aer < eCelt *agrā < IE *Haeĝ-reHa-, probably present in the r-n Aeron Ceredigion < *Agronā (but not, in spite of never-ending speculation about Aeron in CA, Ayr) Watson CPNS 97 derives Hiort/ Hirta (St Kilda) from irt, though I - and others - would reserve judgement. It's an interesting idea, there's some evidence of Gaelic influence (presumably 10th - 11th ct, maybe - following David Parsons' recent argument - Gall Ghaidheil) in this part of Cumbria. But I'm dubious as to the idea of Gaelic speakers at that date giving such a name.

Watson, discussing Hirta, refers to the opinions of some (unspecified) antiquaries who associated it with the ‘isles of the blessed’ etc., and it would be possible to weave some Celtic-mythological notions around such a name for a river. But I tend to steer clear of such notions (and Watson thought St Kilda was more of a hell on earth!).

PNCmb 18 (followed by Watts CamDEPN) misrepresents Ekwall as giving Welsh ir 'fresh, green' as the origin, though actually he doubts that as the second vowel never occurs in records for Irt or Irthing. All the same, it can't be ruled out,  the suffix -et is reasonably common.

I think *ir-et ‘fresh, green’ or even Gaelic irt ‘death’ (though a bit far-fetched) are possible, but wouldn’t rule out *(avon) eirth."

Irthing and Irt Rivers in Cumbria with Roman Roads

[3] I have this information on the significance of Uxellodunum via a personal communication from Professor Anthony Birley:

Dear Mr Hunt,

That the praef. alae Petrianae at Stanwix was the "senior officer" of the Wall garrison is simply a statement of fact: he was the only prefect of an ala milliaria in the entire province and thus was in the quarta militia, the elite highest grade for equestrian officers, probably only created in the early 2nd century. For the regiment see e.g. M.G. Jarrett in the journal Britannia for 1994. Whether this officer ex officio "controlled" the Wall is another matter; but he no doubt at least had the authority to give orders in an emergency without having to wait for authorization from the legionary legate at York (from Caracalla = at the same time the governor of Britannia Inferior) or the consular governor of undivided Britain further south.

The place-name: this is a conjecture by Mark W.C. Hassall, in Aspects of the Notitia (1976), 112f., edd. R. Goodburn and P. Bartholomew, who convincingly restores [Banna] after tribunus cohortis primae Aeliae Dacorum in line 44 in the Duke's list and inserts [tribunus cohortis secundae Tungrorum] before [C]amboglanna, making Banna the name of Birdoswald and Camboglanna that of Castlesteads; and replacing Petrianis after alae Petrianae in line 45 with Uxel(l)oduno, and Axeloduno in line 49 with Mais. This is now generally accepted, see e.g. A.L.F. Rivet & C. Smith, The Place-Names of Roman Britain (179) 220f. Cf. also in Britannia for 2004 on the Staffordshire pan, with another list of place-names from the western sector of the Wall.

Best wishes,

Anthony Birley

I discuss the archaeological evidence for sub-Roman continuation at Stanwix in my book THE ARTHUR OF HISTORY.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.