A KERFUFFLE OVER KAWELL or JUST HOW CAN THE UTHER ELEGY HELP US?

Hadrian's Wall

Only a week or so ago, I announced that the famed editor of the Uther Pendragon elegiac poem, Prof. Marged Haycock, had told me the word kawell - 'basket'- could be allowed to stand. This meant that I was free to theoretically link kawell to Ceawlin, as AS ceawl = 'basket.'

Unfortunately, I've continued discussing kawell with the professor. One of the things I was curious about was another word, cafell, which like cawell had been derived from L. cauellus. The word meant "sanctuary" or "temple", or even the Biblical Holy of Holies. I had asked Dr. Simon Rodway about it years ago.

To my surprise, she did not have a problem with cawell for cafell. And, indeed, cafell seemed a reasonable, common sense meaning for the phrase in question: pen cafell would be a title for God in the same line and would mean "Chief of the Sanctuary."

So:

It's I who's a leader in the darkness

May our God, Chief of the Sanctuary,

transform me

It's I who's like (or who's a second) kawyl in the gloom

While this is a decent reading for these lines, we haven't made much progress. For we lose kawell for Ceawlin and we are stuck wondering if Sawyl for kawyl (emended for kawyl through the copying process of eye-skip) can be retained.  For on May 9, 2023, Dr. Simon Rodway told me:

"Every line in this poem has end-rhyme.  Kawell forms proest (a type of half-rhyme) with tywyll, so that might be okay, although there are no other examples of proest in the poem.

Considering that double n is often written single in Middle Welsh, and that e for y is extremely common, I don’t see a great difficulty in reading kan(n)wyll for kawell.

n could have been written for nn in an exemplar with a suspension mark, and then the suspension mark omitted."

Professor Peter Schrijver was most helpful in giving his take of the poem:

"l. 23 gyhyr shows “Irish” rhyme (dd and r belong to the same class of consonants: voiced continuants)

l. 28 geinc shows Irish rhyme with -eint (t and k belong to the same class of consonant: voiceless plosives)

l. 34 goruawr gyghallen: last word does not rhyme, but it looks like this is compensated by preceding goruawr (rhyme in -awr); note that the commentary wrestles with the shortness of the lines and wonders whether the text is corrupt.

l. 40 gwrthglodyat – byt: same situation as in l. 34: byt does not rhyme (but does rhyme with the first word in l. 41) but gwrthglodyat does (in at)

So yes, there are other lines with rhyme problems. But they fall into different categories than kawell – tywyll would if taken at face value (proest/consonantal rhyme, if that is what it is). So there is no certainty that kawell/tywyll cannot be taken at face value but just the likelihood that they cannot."

That kawell represents one of only two proest-style rhymes in the entire poem suggests pretty strongly the word is corrupt and that we are justified in seeking to emend it.  Doing so brings it in line with the end-rhyme scheme of the rest of the poem.  And the emendation is not a wild one, but a simple and allowable one.  Logic dictates that we accept such. 

We can go even further with this. It would make no sense for cannwyll, a frequent rhyme partner to the tywyll found at end of the line before kawell, to instead be placed mid-line after kawell. 

If kawell is kan(n)wyll and refers to God -

May our God, the Pen Cannwyll, transform me

- then the only other possibility for the following kawyl is, in fact, Sawyl.  And as the Biblical Samuel was responsible for the lamp of God within the Shiloh shrine, Pen Cannwyll as 'Chief of the Lamp' (lamp being one of cannwyll's attested transf. meanings) would be poetically apt. 

How do we decide between the various options?

It would, in this case, be logical to go back to two things: the name Arthur itself, and the Arthurian battles. 

Arthur is from Latin Artorius. The linguistics work. No other etymology works. The temptation, then, is to look towards Carvoran Roman fort on the Wall, where a Dalmatian unit was long in garrison and a woman from the Salona of the Artorii was buried. Carvoran was near Birdoswald, itself in the valley of the *Artenses or Bear-people. Birdoswald was manned by the draco revering Dacians and may even have been referred to as the fort of the Aelian dragon. We know there was an extraordinary royal hall there during Arthur's floruit.

L. Artorius Castus was prefect of the Sixth at York before he led some British legionary troops against ARM[...]S and then became procurator of Liburnia.  He may have been born in Dalmatia, but at the very least had Dalmatian connections and the Artorii in Dalmatia are probably descended from him. 

On the other hand, that a Sawyl ruled from the Ribchester of the Sarmatian veterans points to another possibility, viz. that Arm[...]s is for ARMORICOS, not for ARMENIOS.  The first would allow Castus to be in Britain when the Sarmatians were there.  The second puts him in Britain prior to the arrival there of the Sarmatians.

We might suppose, without too much of a stretch, that the Artorius name was known of and preserved for several generations at Carvoran.  While purely speculative, it does not strain credulity to have Arthur's mother hail from Carvoran and his father, the Terrible Chief-dragon (or magister draconum?), be the ruler at Birdoswald.

As I've mentioned many times, Camboglanna/Camlann is just west os Birdoswald in the same Bear-people's valley, and Aballava/Avalana/"Avalon" is not far west of Camboglanna.

But if Artorius used Sarmatian troops in Britain and possibly on the Continent (see below), his name might well have been remembered in the vicinity of Ribchester as well.  

We can even keep all the same northern sites while retaining Sawyl of Ribchester. Let's look at those before we return to our discussion of the Birdoswald-Ribchester dichotomy.

All of Arthur's battles are easily locatable in the North without going through linguistic contortions or creative translations. In addition, traditions recorded in the Pa Gur poem, annal entries and saints' lives from the Irish sources and additional medieval period folk-names argue rather forcefully for at least some of these battles being strictly northern. One Welsh story even firmly places Badon in the North (see below).

Of course, before we can "go" with all that, we must be willing to ignore a great deal of what would be spurious tradition in the South. Much of the early Welsh material would have to be accepted as the usual legend relocation that occurs when borders recede, and the Celtic fringe became all that was left after conquest. This kind of thing happened in Wales and Cornwall, and even in Brittany.

The biggest problem with someone like Cerdic of Wessex/Ceredig son of Cunedda as Arthur? Well, first, we can't demonstrate how Artorius as a name would have been transmitted to Ceredig. Or why it would have been substituted as a decknamen for a Celtic bear name. I mean, he has three immediate descendents in his pedigree who have Brittonic bear names. Why did none of them find it necessary to use a Latin derived bear name (like Ursius) or a Latin name they perceived to be a bear name (like Artorius)? This badly damages  - if not totally destroys - the concept that Artorius was chosen as a decknamen for Ceredig.

So what if Arthur is (for lack of a better way of putting it!) simply Arthur?  And Uther Uther?  What if most, if not all, of the traditional lore I've been treating of is misleading and useless when it comes to trying to trace a historical figure?

Well, as I've hinted at already above, it's not all useless. Instead of being able to provide a tentative connection to one of the ancient Welsh genealogies (which were, of course, often preserved in corrupt form, manipulated for various reasons and sometimes literally manufactured), we must confine ourselves to the following "facts":

1) Arthur is from the Roman name Artorius. While it is certainly possible there were other Artorii in Britain besides L. Artorius Castus who could have lent the name to a subsequent generation, the only man we know of was Castus.  Furthermore, he not only acquired very high position as an equestrian, he would have been renowned for his service in Armenia with British troops.  That he had strong Dalmatian connections and ended up in Dalmatia (where several Artorii have been attested), and that we have a Dalmatian garrisoned fort on Hadrian's Wall just a few miles east of Birdoswald may also be significant. Artorius may also have been preserved at Ribchester, which was the fort of the Sarmatian veterans and was always subject to heavy influence from Castus's York.  Commanders from York actually led groups of Sarmatian cavalry. 

2) If the draco is to be properly associated with Uther Pendragon, and given the presence at Dacian-garrisoned Birdoswald of the sub-Roman/early Medieval royal hall, we could make a case for Uther's origin lying at the Banna Roman fort. My previous idea - that Arthur may have originated from the Ribchester Roman fort of the Sarmatian veterans, loses some steam when we realize the Sarmatians did not, in fact, have a draco standard - something that I've aptly demonstrated. However, he could still have been there if we allow for the whole draco and dragon-star episode being concocted from Geoffrey of Monmouth via his misinterpretation of the epithet Pendragon as the Dragon's Head. 

4) In a corrupt Welsh TRIAD, Arthur Benuchel is made a son of Eliffer (who almost certainly belongs at York, his 'great retinue' being a poetic reference to the Sixth Legion based there, and his son Peredur being a Welsh attempt at Praetor - not *Pritorix; see Rachel Bromwich’s Triads of the Island of Britain, p. 561). This looks attractive, given Castus' being stationed at York, but when one examines the original TRIAD and understands how these kinds of corruptions occur, we can easily dispense with this possibility.  It is true, however (and I have this through extensive correspondence with Professor Roger Tomlin) that the PRAESDIUM of the "Notitia Dignitatum", manned by Dalmatian cavalry in the late period, may well have been just across the River Ouse from York. 

5) There is a fair amount of traditional and historical evidence for the placement of the Arthurian battles in the North. The Bassas battle conforms very well to Dunipace, both in terms of probable etymology and a double historical/folkloristic "fix" at the site. The same is true of the Tribruit battle, which the 'Pa Gur' quite specifically pinpoints as the trajectus at North Queensferry.  The Welsh story

(late though it is) "The Dream of Rhonabwy" describes Badon as being Buxton. The City of the Legion can be nothing other than York. In fact, there is no other legionary city in Britain that makes sense as the site of a battle against the English during Arthur's floruit. Breguoin is perfectly derived from Brewyn, the Roman Bremenium at High Rochester, and Agned or Agued is a reference to Catterick, a Roman fort in the "Gododdin" poem (a poem that compares one of the warriors at Catterick with Arthur).  All of that taken together with the acceptable identification of Guinnion (for Guinuion), the Celidon Wood with the Welsh Lowland forest of that name (centered on the Caddon Water), the mouth of the Glen with the mouth of the Northumberland Glen, and Dubglas in Linnuis with the Devil's Water at Linnels near Corbridge, makes it nigh impossible for us to dislodge Arthur's arena of military activity from the North.  All of these battles run up and down or to either side of the Roman Dere Street, extending north and south of the Wall.  The perfect control node for such a series of battles would be the central portion of Hadrian's Wall.  In other words, someplace exactly like Birdoswald.  

Now we can circle back to Sawyl vs. a chieftain at Birdoswald.

Sawyl has some advantages as a potential paternal candidate.  Firstly, we know his wife was an Irish princess.  This is vitally important, as all subsequent Arthurs belonged to Irish-descended dynasties in Britain.  The only way we can really explain this fact is if we allow the first, more famous Arthur to have been part Irish. The Irish would have then wanted to claim the name, while the British may have been chary to do so.  Second, it is difficult to dispense with Sawyl's son Madog Ailithir, when we are told Uther had a sone Madog and Madog a son Eliwlad.  Ailithir and Eliwlad certainly appear to be semantically identical or at the very least Eliwlad looks to have been fashioned to resemble Ailithir.  

Third, we have a tradition which I have shown wrongly identifies Uther Pendragon with a northern Sawyl. This came about because Illtud's Latin military ranks/titles could easily be rendered into Welsh as Uther Pendragon.  [An attempt to suggest Illtud was actually Uther was abandoned, as doing that once again imposed upon us an unworkable southern sphere of military activity.] Then we find Illtud and Sawyl exchanged for each other in a St. Cadog episode in the saints' Lives. Geoffrey of Monmouth compares Illtud (= Eldadus) with the Biblical Samuel. 

It is perhaps most likely that Uther Pendragon actually was originally a designation for Illtud. But when Illtud/Uther was poetically compared to the Biblical Samuel - in Welsh Sawyl - he was wrongly identified as Arthur's father because there was a Sawyl at Ribchester who really was Arthur's father. This may sound overly convoluted, but in the realm of legend formation such things happen.

We know Geoffrey took the gorlasar epithet of Uther and created from it an entirely separate character - Gorlois, Duke of Cornwall. So it does not take much for Uther Pendragon to take on individuality himself. Once he is said to be a second Sawyl, transformed by God into that form, he becomes hopelessly entangled with Sawyl of Ribchester, Arthur's father. And hence we end up with Arthur son of Uther Pendragon.

True, as mentioned above, Geoffrey knew of the Illtud-Samuel comparison. And that would seem to complicate my chain of reasoning. But as he either mistakenly or intentionally converted gorlasar into Gorlois, and then freely displays the relationsip between Uther and the gorlasar epithet by having Uther transform into the likeness of Gorlois, it wouldn't take much to have him use Uther as Arthur's father rather than Sawyl. In truth, the poem says that God transforms Uther into a second Samuel, while Geoffrey has Merlin/Myrddin transform Uther into Gorlois.  

One almost wonders if there were a decided effort on Geoffrey's part to force Illtud's cryptic name/title into the Arthurian canon precisely because by doing so he was able to have Arthur's father's origin in SE Wales next to his own Monmouth, rather than in Lancashire. I have shown that Illtud's "Llydaw" and father Bicanus are representative of Lydbrook and Bicknor ( = Llangystennin) near Ganarew/Little Doward close to Monmouth.

I should add that there is no Galfridian influence apparent in the MARWNAT VTHYR PEN.

While no one seems to much like that idea that L.  Artorius Castus fought in Armorica during the Deserter's War and had his procuratorship bestowed upon him by Cleander, there is nothing really wrong with the idea.  Yes, there is some evidence that Liburnia was founded c. 169/170 A.D., but it is not a requirement that Castus be the new province's first procurator.   It also remains true that the only literary account of a British mission to the Continent with the exact equivalent of three legionary detachments is the force said to go to Rome demanding the execution of Perennis.

On the whole, then, if we forsake Sawyl for the Birdoswald Arthur, we lose everything that is so attractive about the former. With Birdoswald, we do get a place that may have had a bear name, and we do get Dacians with a draco and Carvoran with its Dalmatians. Otherwise, we are, essentially, just depositing Arthur there because the site looks good and we can envisage someone like Arthur having been there. However, there is no genealogical trace, there is no Irish connection, and most critically we lose the Uther-Illtud-Sawyl comparison - a comparison that is pretty much impossible to ignore.

So where does all this leave us?

Well, the only clue we have to Uther's true identity lies in the PA GUR poem.  If my treatment of that poem is correct, Uther is either Illtud or Sawyl.  Illtud seems highly unlikely and reads like a misidentification.  Leaving us with Sawyl.  Had Sawyl been located somewhere other than Ribchester, the debate would be entirely different.  That a strong case has been made for his presence there bolsters the idea that the name Artorius was preserved in the region.  This could only have happened, it seems to me, had Castus been known to the Sarmatians who served under him.  It is difficult to sustain an argument that the Artorius name was taken by the Ribchester folk from York, despite the acknowledged ties between the two places. Why would the partly Sarmatian-descended population of Ribchester care about a Sixth Legion prefect who has served in Britain before they even arrived?  Especially after a couple of centuries had elapsed!

Admittedly, I have been seeking a way out of having to embrace the Sarmatian element in Arthurian theory.  And this is precisely because I feel it has been misapplied and grossly overdone - to the point where everything has been made out to be Sarmatian (or the allied Alanic).  Still, as my late father was prone to saying, "Don't throw the baby out with the bath water." 

At this point I am holding onto Sawyl, and leaving my book THE BATTLE-LEADER OF RIBCHESTER out there.  Until and if someone comes forward with evidence or good argumentation to change my mind, I'm letting the matter rest.   

HOW TO GET CEAWLIN FROM COLINE: A LINGUISTIC ARGUMENT FOR CUINNID/CUNEDDA MAC CUILINN AS THE GEWISSEI CHIEFTAIN

-coline on the Wroxeter Stone

Ceaulin in the ANGLO-SAXON CHRONICLE

The following is taken from the WIKIPEDIA article on the West Saxon name 'Ceawlin:'

Since there is no obvious Old English or wider Germanic origin for Ceawlin,[5]: 4  commentators have frequently assumed that it must originate in the Celtic languages, like the name Cerdic borne by another early West Saxon king.[7]: 37 [8]: 513  However, no secure Celtic etymology has, as of 2019, been forthcoming:

In 1941, O. S. Anderson suggested that the names were both contractions of the Welsh name Cadwallon, with the addition of the diminutive suffix -īn in the case of Ceawlin.[4]: 64, 92 

While noting that no such name is found in the Celtic languages, Richard Coates cautiously suggested in 1989–90 that "It could be derived from a British *Cawolīnos or, better, a hypothetical Pr[oto]W[elsh] *Cawlīn", positing a relationship with the Welsh word caw ("skilled"), but lacking close parallels for the -līn element.[5]: 4 

Arguing that -lin was a diminutive suffix for names more widely in Old English, John Insley argued in 2019 that Ceawlin is a diminutive of Ceawa, but did not offer an etymology for that name.[6]

Unbeknownst to Insley, however, Ceawa had been etymologised by Gillis Kristensson as an Old English counterpart of Middle High German kouwe 'jaw, jawbone' (from West Germanic *kauwō-).[9] In this case, Ceawlin would be a name of Old English etymology.

I do not think any of these suggestions are correct.  In my book THE BEAR KING, I demonstrate that the great Cunedda of NW Wales was not British, and he did not hail from Manau Gododdin in the far North of Britain.  Instead, he is to be identified with a member of the Irish Ciannachta named Cuinnidh (and variations) Mac Cuilinn, 'son of Holly.'  He came from Drumanagh.  My argument for this identification is long and involved, but I think very strong.  

However, I did not stop there. It became apparent to me that Cerdic of Wessex, indisputably the British name Ceredig, was the prince of that name who was the son of Cunedda.  The connection was made through a variety of means, but of primary importance is the Wroxeter Stone pictured above.  I made the case for Cunorix of the stone being Cynric of the ANGLO-SAXON CHRONICLE, with his father Maquicoline being Ceawlin.  [Note there is a curious reversal of generations in the Gewissei of the Welsh sources and those of the English sources.]

Still, there remained one serious sticking point in what was otherwise a very comprehensive portrait of Irish federates or mercenaries fighting in southern England: I had to be able to show that Ceawlin could, indeed, come from Irish Coline.  

The solution to the problem came when I discussed the 5th century Coline with Professor Jurgen Uhlich of Trinity College, Dublin.  I alread knew from A LATE INSCRIPTION FROM WROXETER

by R. P. WRIGHT, F.S.A. and PROFESSOR K. H. JACKSON that Coline

"does not yet show the vowel change which would have been expected to turn it into Culini at some time in the later part of the fifth century"

Uhlich began by explaining to me that the form Cuilin was not germane to our discussion, as this change occurred

"Sometime in the 8th century, at a rough guess, i.e. when the habit of marking even internal consonants for colour gradually caught on. To be sure, this extra i is merely a ‘glide’ letter, with no additional pronunciation involved whatsoever."

We need to remember that Bede wrote c. 730, while the ASC was written in the 9th century.

Continuing his explanation of how this name could have changed through the centuries, Uhlich told me that

"au for u is merely a hypercorrect spelling in backward imitation of the 8th-century change of short au > u (such as in Cú Chaulin(n) > Cú Chul(a)inn, with the second word continuing what in Ogam spelling is attested as CALUNO-). This mechanical hypercorrection, however, went so far as to include even long ú (where a pre-stage *áu is impossible to assume), so you even find, say, cáu for cú, etc. As it happens, I have dealt with this in Ériu 46 (1995), §16."

His study can be found here:

https://www.jstor.org/stable/30007872?read-now=1#page_scan_tab_contents

I then noticed the spellings for cawl in the BOSWORTH AND TOLLER ANGLO-SAXON DICTIONARY, which included ceawl:

https://bosworthtoller.com/005894

cawl

Noun [ masculine ]

 

cawl, caul, ceawl, ceaul, es; m. A basket; sporta, corbis, cophĭnus = κόφινος

 Linked entries

v.  caul ceaol ceawl cel ceofl ceol ceoul ceowl cewl.

I then proposed to Prof. Jurgen that a Caulin spelling for Culin was known by the West Saxons. Pronunciation is not important, as a scribe faced with Caulin could easily have 'Anglicized' that spelling to Ceawlin. According to Richard Coates, Ceaw- would be pronounced like Welsh caw, which is roughly the "chow" sound we need. I think Cau- could have been construed as having the same pronunctiation - as we can see in the entry for cawl above, where the spellings ceawl and caul are both found. 

It is true (in Coates' words) that "only WSax has the palatal affricate 'ch', but again, we are talking about a written form, not a spoken one.  Actual Irish phonology is not at issue here.  A WSax scribe with the name in front of him spelled C- would still have copied it as a C-, even if he pronounced it Ch-. 

Professor Jurgen's response?

"In short, your scenario could work technically, i.e. an AS scribe reading the hypercorrect spelling with <au> somewhere and then transposing this mechanically into an OE one."

I, personally, am more than content with this method of suggesting that the West Saxon Ceawlin might represent the secondary name of Cunedda.

AN UNEXPECTED REVELATION: WHY THE ALLOWED RETENTION OF A SINGLE WORD IN THE UTHER ELEGY CHANGES EVERYTHING

Tywyll, Pen Kawell and Kawyl in the 'Marwnat Vthyr Pen'

The Discussion of 'proest' in Jenny Rowland's EARLY WELSH SAGA POETRY, p. 335

Several years ago I thought I had made an amazing discovery: a single word in the ancient Welsh poem 'Marwnat Vthyr Pen' ("The Death-Song of Uther Pen[dragon]) had allowed me to definitively identify Arthur's father with Ceawlin of the ANGLO-SAXON CHRONICLE.  While the idea to some seemed outlandish, even crazy, I had before this shown that there was a strong likelihood the Arthurian battles as found listed in the HISTORIA BRITTONUM were, in fact, Welsh renderings of the Gewissei battles of the ASC.  Many other correspondences were forthcoming, and all of them seemed to be confirmed by my interpretation of the Uther elegy.

Alas, because I was assured by three preeminent Celticists that kawell couldn't stand as is, that the word must be emended to fit the end-rhyme scheme of the poem, I had to settle for W. kannwyll in this position for the relevant line.  Such a emendation forced me to abandon the notion that Uther = Ceawlin, and I was spun off in different directions for all subsequent future research and theorizing.  

Well, as it turns out, the scholars in question were mistaken.  I've only come to realize this recently, at the end of a long road of ongoing communication with other academincs specializing in early Welsh poetry.  My journey of rediscovery began with the following message from Dr. David Callander:

"I have not studied this specific example in detail, but of possible relevance to your enquiry is the Welsh poetic technique known as proest. With this, vowels of the same length can "rhyme" so e.g. "yll" and "ell". I believe that this is discussed as part of the metrical study in the introduction to Jenny Rowland, Early Welsh Saga Poetry."

I had not heard of proest before, and thought that pursuing that technique might bear fruit.  So, once again, I began writing queries.

I started with Dr. Simon Rodway.  In a two part question, I asked first if proest might be present in the Uther elegy, and second if kawell could be left when operating under such a principle.  He responded, respectively:

"This is a good point."  

and

"Perhaps."

I then heard back from Dr. Ben Guy on this issue:

"I'm glad to hear that you're using Marged Haycock's edition in her Legendary Poems from the Book of Taliesin, which is the best general guide on the poem. Regarding your specific question, it would be fine to keep tywyll together with kawell from the point of view of the metre (Sawyl is less important in this sense since it's not an end-rhyme) - the final consonant is the same in both (-ll) but the vowel varies (-y-, -e-), so it's 'proest' rhyme rather than full rhyme, but that's fine in poetry like this."

This gave me a glimmer of hope, and I once again reached out to Prof. Marged Haycock herself.  In the past, she had failed to respond to my query.  This time, however, she was kind enough to get back to me in a timely fashion:

"Kawyl is not at the end of the line so is not in this case involved in any prescribed rhyme  (tywyll, kawell are proest rhymes). The tywyll and kawell do not form a full rhyme, and are not strictly speaking a normal type of proest (which is a half rhyme where the vowels vary, but consonants match). This is because tywyll contains a diphthong -wy- whereas kawell just has simple vowel. However, some -wy- sounds can morph into a clear -y, in which case tywyll might form a sort of proest with kawell. It seems near enough to pass muster."

This statement is utterly authoritative.  While it is important to add that she went on to say -

"However, the slight irregularity might suggest to some an emendation to a word ending in -wyll, or alternatively the fem. form of tywyll which is tawell (that would give you full rhyme)."

- the important thing to recognize is that kawell is acceptable.  I would, of course, go well beyond that conclusion.  Why?

Because if we save kawell, and instead emend the following kawyl to kannwyll, we 1) follow the law of Occam's Razor, changing only one word instead of two (as otherwise kawyl has to be altered to Sawyl by proposing a copying error known as eye-skip, where the scribe accidentally transfers the k from kawell on the previous line to sawyl on the following line), and 2) emending kawyl to kannwyll allows us to have Uther transformed into a star (cannwyll has the transf. meaning of 'star"), and this would account for the Galfridian dragon-star which is said to be Uther himself.  

[Geoffrey of Monmouth used the gorlassar epithet which Uther applies to himself in the elegy to conjure an entirely separate personage - Gorlois, Duke of Cornwall - but then acknowledges that Gorlois is Uther by having the latter magically transform into the former!]

Reading the poem in this way would allow us to reject the emendation Sawyl entirely.  We would no longer have to seek some connection between Uther and Sawyl, whether this connection involves St. Illtud or any other figure, including Sawyl on the North.  In short, with Sawyl missing from the poem, and Uther as a star being present there instead, we are free to utilize kawell as a Welsh reference to Ceawlin.

Because of this, I will shortly be reissuing my book THE BEAR KING.  At this point, I do not anticipate having to do any future work on Arthur, as every important aspect of his story that could be treated of is dealt with in detail in this volume.