ELIWLAD FLIES AGAIN? or A NEW SOLUTION TO AN ETYMOLOGICAL PROBLEM

       The God Lleu as a Death-Eagle

                      in an Oak Tree

Several years ago I floated a new theory in a book about a possible Ribchester-based Dark Age Arthur. That theory has two related arguments. One, that the kawyl of the MARWNAT VTHYR PEN stood for Sawyl, and second that Eliwlad son of Madog son of Uther was a Welsh substitution for the Irish ailithir epithet of Madog son of Sawyl Benisel.

Surprisingly, I had obtained a considerable body of scholarly support for the idea. But, ultimately, there were a couple of linguistic problems and I opted to abandon the theory. Chief among these problems was my inability to explain why the Welsh would have substituted gwlad for Irish tir when Welsh had its own tir with the same meaning. There was also the need to propose a very rare metathesis due to a copying error when going from Ir. Aili- to W. Eli-. The only way out of that predicament was to suggest a direct phonological development of Aili- to Eli-. This was deemed quite acceptable, but left us with the gwlad-for-tir issue.

It didn't help that I myself had come up with a couple of other very decent etymologies for Eliwlad (Eiliw-gwlad, Appearance-lord, or Eiliw/Eilyw-gwlad, Grief-lord).

To be honest, I deeply regreted the collapse of the Eliwlad theory. Instinctively, I felt it was right. I really didn't want to let go of it. This stubbornness on my part led me to consider another possibility for transmission of the name Eliwlad.

Supposing, I wondered, Eliwlad had somehow been brought over into Welsh entire from the Irish? This may seem like an obvious question - except that no aile + flaith compound was extant in Irish.  As I'd often done when I faced such a dilemma, I contacted Prof. Jurgen Uhlich of Trinity College London and put the question to him.

His complete analysis of the Eliwlad problem from the standpoint of a possible connection with Ir. Aile + flaith follows:

"Such a compound as aile-flaith could of course have existed. Its shape (and thus its reflection in spelling), however, would not feature aili-, as e.g. in ailithir, since while the schwa /ə/ of the second syllable in alilithir is followed by a palatal consonant and needs to be spelled <i> between two palatals, what follows in your example is neutral, hence *aile(f)laith or indeed, if with regular unstressed processing, *ailelaid.

The element <f> of flaith is merely due to etymological spelling, i.e. by recognising that the second element corresponds to the simplex flaith, with an actually pronounced /f/, whereas in second position in a compound like aile(f)laith, the f is non-existent after it (or more precisely its predecessor *u̯) had been lenited to zero, and the <f> was only spelled ‘back in’ for easier etymological recognition (cf. also GOI §231.7 for purely phonetic spellings without such an <f>).

As for your question of timing, the phonetically regular spellings would be Early Old Irish a(i)le(f)laith (including with or without <f>, which is immaterial), with -th still preserved at the end of an unstressed syllable, > Classical Old Irish aile(f)laid with regular voicing in that position (cf. the pattern of cath ‘battle’ and its compound cocad ‘war’, and see further McCone in Ériu 32 (1981), 29-44). In compounds, however, the same etymological ‘recognition’ as discussed for a different feature above could always spell the -th ‘back in’ by association with the simplex flaith, all the while still pronouncing it still with voiced /δ/, or indeed even with the voiceless /θ/ introduced from the simplex. Whatever about these considerations, neither a pronounced /δ/ nor a pronounced /θ/ would correspond to Welsh /d/ in a mechanical = purely phonological loanword, so once again one has to resort to etymological recognition equating not only aile- with eli- but also –(f)laith with native gwlad, which were then put together as the corresponding Welsh compound Eli-wlad, by a process referred to as ‘’loan-translation’. This could work at any period, as long as aile(f)laith/d was understood as aile + flaith."

Now, this revelation was pretty exciting to me. But it wouldn't do me much good unless I could arrive at a consensus view by involving other Celtic language specialists in the discussion.

So I sent Jurgen's treatment on Eliwlad and aile-flaith first to Prof. Peter Schrijver. His response?

"Yes, that might work."

While not exactly a ringing endorsement, I take this to be a victory as the top linguists are notoriously difficult to satisfy (language laws being by definition quite rigid in their applications). Furthermore, his opinion was followed by agreement on the possibility by Dr. Simon Rodway, Dr. Richard Coates, Prof. Ranko Matasovic and Dr. Alan James.

What was I to do now?

Well, the Uther = Sawyl equation was the only good identification I had for Arthur's father. Otherwise, I was forced to interpret Uther Pendragon as a conjured name (from the mil uathmar... chend of the Irish COMPERT MONGAN tale). And as Sawyl had Irish connections (his wife was a Dal Fiatach princess) and resided where Sarmatian cavalry had been in garrison for a couple centuries, I could easily account for the later Arthurs in the horse-loving, Irish-founded kingdoms of Dyfed and Dalriada.

It does not matter that Eliwlad is placed by the "Dialogue of Arthur and the Eagle" in Cornwall  in a 15th century MS. (which Sir Ifor Williams traced back to the 12th century), or that the dead eagle in an oak motif was taken from the Lleu story (the Madoc place-names in Cornwall match those in Nantlle). All of that would have been subsequent story and is not reflective of the earliest tradition. However, Eliwlad as instructor of Christian virtues to Arthur is perfectly paralleled by another medieval didactic poem in which a pilgrim (W. creiriwr; cf. Ir. ailithir) is instructed by Mary Magdalene of Brynbuga. And Prof. Stefan Zimmer has told me that an "other land" designation for Eliwlad may well have contributed to the notion that the other land in question was the Otherworld, i.e. Heaven. Hence Eliwlad appearing in spectral form.

There is an element of the Uther identification with Sawyl that I don't like, of course: the Sarmatian connection. I'm quite sure of that connection because I'd earlier proven that Sawyl Benisel belongs at Samlesbury in Lancashire near Ribchester. And that means the epithet Pendragon will once again be claimed as a reference to the Sarmatian draco.

As my readers know by now, I did a lot of research on the draco standard. My conclusion was that we utterly lacked textual or archaeological evidence for the existence of the Sarmatian draco. The Dacians had one, as did the Alans. So while it would be safe to assume the Sarmatians had their own flying dragon standard, we can't prove that they did. We can say that rather than adhering to strict Welsh poetic usage of the word dragon as a metaphor for "warrior", Sawyl might have been referred to as the "Chief Dragon" of an ethnic group whose totem was the dragon. 

We cannot relate Sawyl's epithet Benisel to Pendragon. Benisel or "Low-head" meant someone who was humble. Penuchel, sometimes substitited for Benisel, conversely meant "High-head", i.e. one who was prideful or arrogant.

Where does this all leave us?

I'm not sure yet. Certainly, if the preferred emendation of Sawyl for kawyl in the Uther elegy is correct - and that does make for a good contextual fit - and we can allow for Eliwlad to have come from Ir. aile-flaith as an epithet identical in meaning to Ailithir (an epithet originally attached to St. Matoc that was wrongly made into Matoc's/Madog's son), then we have a historical candidate for Uther Pendragon. 

I should remind my readers that Uther says in the elegy that God transforms him into a "second kawyl", or rather "second Sawyl." That is to say, a second Samuel, the first being, presumably, the Biblical prophet of that name. The /k/ of kawyl would be due to the common scribal error of eye-skip, the /k/ being dropped down from the /k/ of kawell in the previous line. Kawyl has no meaning in Welsh. It's unlikely to be an error for cannwyll, "candle', for reasons of scribal practice [1] and poetic convention [2], facts made plain by Dr. Simon Rodway and Prof. Marged Haycock (editor of the MARWNAT VTHYR HEN). 

[1]

"1) This requires positing an n-suspension. These do occur occasionally in medieval Welsh MSS, but they are very rare." [NOTE: cannwyll in 4 out of the 5 earliest attestations in GPC is spelled with only one /n/.]

2) The single l would mean suggesting an Old Welsh exemplar, for which there is no other clear evidence in the poem. Elsewhere the scribe has ll where needed, so if he was copying from an examplar with l for ll, then this would be the only occasion on which he didn’t correctly modernize.

Overall, emendation to Sawyl, while totally speculative, involves less issues (eye-skip to kawell), and eil Sawyl, ‘a second Samuel’ gives plausible sense." - SR

[2]

"Kawyl (or Sawyl) is not at the end of the line so is not in this case involved in any prescribed rhyme (tywyll, kawell are proest rhymes)." - MH This is significant because tywyll and cannwyll are found paired in the poetry, but a poet would not pair them unless both were at the ends of their respective lines.

Haycock was also kind enough to approve my reading of kafell ("temple, sanctuary") for kawell (which as it stands would appear to mean "basket"):

"The cafell variant of cawell < Late Latin cauella may already have developed before the first attestation in GPC."

She then concluded with a discussion of end-rhyme: 

"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. 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)."

Needless to say, kawell from cannwyll evinces the same /nn/ problem as we just encountered trying to derive cannwyll from kawyl.

All in all, after applying Occam's Razor to the rather opaque Uther elegy, we have these key lines:

Our God, Chief of the Sanctuary, transforms me;

It's I who's a second Samuel in the gloom

This may well be an allusion to the Biblical Samuel in the darkness of the sanctuary at Shiloh.