KENFG/1

Westwood/1879, 19: `This stone stands upon the grass-sward at the side of the road, supposed to be identical with the Via Julia Maritima, between Kenfig (Kenfegge or Kenfyg) and Margam, near the Pyle Station, without any protection. It is called by the peasantry of the neighbourhood Bêdh Morgan Morganwg, i . e. the sepulchre of Prince Morgan...This stone was first noticed by Bishop Francis Godwin, Bishop of Llandaff, in a letter to Camden, preserved among the Cottonian MSS. in the British Museum (Julius F. vi. fol. 282), and published by Camden, ed. vi. p. 499.[1]...The inscription,' says Gibson (Camden, ii. p. 23), is now in the same place and is called by the common people Bêdh Morgan Morganwg, viz, the sepulchre of Prince Morgan, who was slain, as they would have it, 800 years before Christ's nativity, which whatever gave occasion to it, is doubtless an erroneous tradition, but antiquaries know full well that these characters and formes of letters be of a farre later date, it being no other than the tomb stone of one Pompeius Carantorius.

[1] Besides the figure of this stone given in the various editions of Camden's Britannia, others have been published by Strange in the Archaeologia, vol. vi. 1782, p. 17, and in Gent. Mag. vol. lV. 1785, p. 502, tab. ann. fig. 2; and by Donovan, Tour in South Wales, vol. ii. p. 30, fig. opposite p. 24'.

Rhys/1899, 136: `N.B. -- The stone, where it now stands by the roadside, is liable to be damaged, as it is a target for boys to throw stones at, and to be maltreated in other ways. It would be a very good thing if it could be placed with the other epigraphic treasures in the church'.

Macalister in Anon/1928, 374: `It has...long been known as a landmark beside the road from Margam to Pyle. It should not be left there much longer, however, for in these days of motor lorries an unfortunate side-slip might utterly destroy it'.

Macalister/1945, 385: `Formerly on the west side of the road from Pyle to Margam, and about half-way between those places. In Lhuyd's time it lay as a bridge over a ditch: Gibson describes it as lying by the wayside between Margam and Kynfyg: and it was erected by the roadside by `the Rev. Mr. Williams of Margam' according to Gough, where it was standing in 1761. Now moved for safety to the lapidary museum at Margam'.

Nash-Williams/1950, 132: `KENFIG...Water Street (near Eglwys Nunydd)'.

RCAHMW/1976, 38: `Its original location near Eglwys Nynnid was attested in 1578,[1] but it would seem to have been set up (or re-set) beside the road known as Water Street in the mid-18th century.[2] It is now in Margam Museum.

[1] Merrick, Morg. Arch., p. 5; cf. Lhuyd, Parochialia, III, p. 225. For the place-name, see Arch. Camb., CXVIII (1969), pp. 144-5. No. 921 is also from this spot.

[2] Camden, Britannia (ed. Gough), II, p. 495'.

RCAHMW/1976, 38: `Rhaetic sandstone'.

Westwood/1879, 19: `It is 4 feet 4 inches high on the back side of the stone, the top sloping to the front, which is 4 feet high ; it is 20 inches broad and 15 inches thick'.

Nash-Williams/1950, 132: `Roughly quadrangular pillar-stone: 60+" h. x 11--15" w. x 16--21" t. Ogam and Latin inscriptions'.

RCAHMW/1976, 38: `A roughly squared pillar-stone...with Ogam and Latin inscriptions. It is 1.35m above ground at its highest edge, 35cm wide on the inscribed face, and 52cm tapering to 45cm in thickness...A bench-mark has been cut on the adjacent right-hand face'.

Nash-Williams/1950, 132: `The Ogam inscription (incomplete) is incised along both angles of the face reading upwards on l. and downwards on r.'.

RCAHMW/1976, 38: `The Ogam characters on the upper left angle of the face read upwards, those on the right edge downwards'.

Brash/1869, 152--153: `on the two angles of the front face of the stone are several Ogham characters. Those on the left angle are as follow:--

They are situated on the top of the stone, where there is a considerable fracture or flake off at the angle, consequently the inscription is imperfect. The diagonal direction of some of the scores of the first two characters have been remarked on, but there is nothing peculiar in it beyond some freak or inadvertence in the engraving, as I have seen them similarly marked on other monuments. The second inscription on the right angle is as follows:---

H Q D A H C D NG G O A

From the long spaces between several of the characters it is quite evident that this inscription is imperfect; that several of the letters have been obliterated, principally vowels, which being usually small circular or oval dots on the angle are generally the first to be defaced either by violence or weather. From the skeleton of the Ogham which remains, it is, however, quite evident that this is not a bilingual inscription, as any filling-up of the missing letters could not produce the equivalent of the Roman inscription. It is also worthy of remark in this, as in all similar cases, that the inscriptions are always reverse, the Roman reading from top to bottom, the Ogham from bottom to top. It is therefore evident that they are by different hands and of different dates. To my mind, the evidence of this worn and mutilated Ogham pillar-stone is that it was appropriated as the monument of a Romanised Briton after having long performed a similar office for some invading Gaedhal'.

Anon/1869, 444: `But the most singular circumstance is, that in addition to the Roman inscription there is one of well defined Ogham characters. It has not, however (as far as ascertained), been translated'.

Westwood/1879, 20--22: `In the small figure and description of this stone which I published in the Archaeologia Cambrensis, 1846, p. 182, I directed attention to the small incised strokes in groups along the two edges of the front of the stone, those on the left side near the top being three parallel, three radiating,[2] and three radiating, and the groups on the right angle from the top having respectively 2, 5, 2, 5, 5, and 5 strokes. These I had no hesitation in regarding as Oghams. A more careful examination which I subsequently made of the stone, verified by additional rubbings, proved that I had overlooked some strokes, as seen in the result of this examination given in Pl. XII; although, from the evanescent character of many of the strokes by time and ill-usage, it is possible that I have still overlooked some of these marks; indeed, Mr. R. R. Brash, in an article `On the Ogham inscribed Stones of Wales' (Arch. Cambr. 1869, pp. 148--167), thus describes and figures these Oghams :--- [Here Westwood quotes Brash given above]

Mr. Rhys (`The Early Inscribed Stones of Wales,' p. 8). speaking of this stone, adopts a very different and ingenious reading of the Oghams: `The Celtic characters are very hard to read, owing to their having been extensively worn off. With great deference to archaeologists, I venture to suggest that the following letters are to be traced on the stone: Pompei...oral...smeq . ll . n. The first part of this would be Pompei Carantoral, and the termination al would be our adjectival -ol or awl, rendering the Latin ius of Carantorius. Here a character something like the Eisteddfodic [fig /|\] was extemporised to represent p, and when the scribe, if we may so term him, came to make m, that was done by making a long stroke across the angle of the stone as usual, but instead of making /|\ for p in this instance, he left out the first line of it and placed the other two lines to lean against the m, thus forming a conjoint character for mp which greatly puzzled me'.

[2] The three radiating lines have much puzzled subsequent writers. Mr. Wendele (Arch. Camb. 1846, p. 413) compared them to the bird's-claw Ogham of Killarney, whilst the Rev. John Williams (ibid. p.415), with reference to the supposed Bardic alphabet, gives a legend of Einigan Gawr [sic] (the giant) and his three pillars of light, adding, `The Bardic symbol is formed of three radiating lines [fig], which it is said are intended to represent the three diverging rays of light which Einigan saw, and it is remarkable that these three lines contain all the elements of the Bardic alphabet, as there is not a single letter in it that is not formed of some of these lines. Now are those scores on the left angle of the Kenfegge stone, Glamorganshire, the last and most genuine home of Bardo-Druidism, anything more than representations of the Bardic symbol? ... If these arrow-heads had been alone on the stone I should have considered them most certainly as nothing but the Druidical rays, but then come the other forms on the other angle, to create a difficulty which I cannot get over consistently with this theory. The character [fig]' he adds, `is one way in which the name of God is written in the Bardic mystery'.

As part of the above entry in Lapidarium Walliae, Westwood also includes a long footnote explaining the Ogham alphabet.

Rhys/1899, 133--135: `The Ogam is almost hopeless as a whole, though a part of it is of the utmost interest. It begins on the left angle near the top, and it must have read round the top. and down the right-hand angle to the ground, unless we are to assume that the top was left without any writing:---

POP I or E A or O

In fact, the Ogams do not reach to the actual top of the stone on either edge, but the spaces thus seemingly blank are so battered and irregular that it is impossible to say whether they were originally covered with writing or not. The reading on the left edge is Pop[i^e][a^o], which ends with six vowel notches, without any indicacation [sic.] how they are to be divided. As the three first Ogams doubtless mean Pop, the notches must have made either ia or eo; for Popia or Popias would probably be the form in early Goidelic of the Latin nominative Pompeius; but if one should prefer to read Popeo, we should have to regard this as the Goidelic genitive of Pompeius, treated in Goidelic as a name of the u declension. Such a combination as mp, nc, or nt has never been found in Ogam: compare also FECERVT for fecerunt on the old inscribed stone at St. Ninian's, Whithorn, in Galloway (The Academy, 1896, i. 201). In 1897 I found the symbol here used for p on a stone now at Donard, in County Wicklow. The name in which it occurs is Iacinipoi, with an element poi, which I find well established by Father Barry and Mr. R. A. S. Macalister, though they would interpret the symbol differently.

The Ogams on the right-hand edge baffle me, as the vowels are almost all gone, and possibly some consonants as well. The readings I guessed in 1897 may be represented thus:--

R(O)L (IO) N M (U)C (I) LL(U) SO

or

MAN& A ? S OE S (E) N A

I was inclined to prefer Rolion Mucoelluna, supposing Mucoelluna to resolve itself into Mucoe Elluna, with Elluna to be identified with the Iluna so read by Father Barry on a stone at Rathcanning, in County Cork. The first Ogam may be an imperfect ///// r, or else /./// m a ng: it is difficult to decide. The gap where I suggest io is, I fear, too wide to be filled by those vowels. On the other hand, there seems to be insufficient room later for the oe of mucoe. In writing out my notes of 1897, it occurred to me that, by reading upwards, one seemed to obtain the elements of a spelling De]ceddas maqui Dara, Dora, or Doro. But on re-examining the stone in 1898 I found at once that Deceddas was out of the question; but my more recent guesses differ somewhat from the foregoing:

R o l [i o] n Maqui Llena.

mang a oc U

This would admit of our supposing the whole to have originally run somewhat as follows: Popia[s maqua(s) or maqua(s) maqui-]Rol...n maqui Llena, `Pompey son [of or son of the son] of R... n son of L.

There is one thing to which I wish to call particular notice, namely, that the crooked and uneven nature of the angle of the stone near its top on the left hand, where one has to begin the reading with the name Popia (or Popeo), proves that word to have been cut last by the inscriber, and that he began, in fact, at the end. I have known similar instances in Ireland, such as one at Kinnard East, in County Kerry, reading Sangti Llotiti avi Srusa, `S. L. descendant of Sru'. The thing is easily explained: a druid or a scholar of some kind cuts the Ogam on a stick, and hands it to a workman to cut on a stone. Now if the workman could not read, he could not tell which way the Ogam ran. So it was only a chance whether he began at the end or at the beginning; but there was one thing which he usually did: he selected a pretty regular part of the angle to begin his scoring, whereas he ended where he could. In this instance he ended in a bend of the edge of the stone, where nobody can be conceived choosing to begin'.

Macalister, in Anon/1928, 375--377: `Of the Ogham on the left-hand angle only the tips of the scores remain. The angle has been broken off by some vandal -- a good illustration of the intentional destruction of the Ogham script in Wales, to which I have already called attention.

It is, however, easy to restore the missing characters. The inscription must have been [figure] PAMPES. The dotted line shows the present fractured edge of the stone. In the acommpanying photograph a restoration of the broken angle with plasticine is illustrated.

The letter `P' is not a Goidelic letter, and in the original Ogham alphabet there is no provision for representing it. When it became necessary to write `P', the character for the dipthong `IA' was adapted. This character is difficult to cut upon the stone, for it consists of two pairs of crossed lines; it was therefore simplified into single crossed lines, and as such we find it on the Turpill stone at Crickhowel, and also on a stone on Valencia Island. The Kenfig stone gives us a compromise between the orthodox cross of four lines and the stonecutter's shorthand cross of two lines; it is a cross of three lines. This result is important, for it gets rid of the radiating groups of three lines which have had to be somehow identified with the `P's', but for which no precedent existed: and also shows that the orthodox form was quite familiar, though it has never been found in any inscription. The tips of the `s' were taken by previous decipherers for vowel points.

In the plasticine restoration I write the inscription as POPES; this is what would be expected, as the combination `MP' would sink into `P' on Goidelic tongues. But on second thoughts, I should now prefer to restore PAMPES as given above; the name is so exceedingly foreign that it would hardly be assimilated to such an extent...The other Ogham inscription, reading downward on the present right-hand angle, represents, in my opinion, the original appropriation of the stone. In other words, it originally read upward on a left-hand angle; but the stone was stolen to serve as a memorial for Pumpeius, and was turned upside down before receiving its new inscription. At the same time it was shortened for ease of transport, by knocking off the butt end.

The inscription is battered and difficult to read, but yields to patience. It runs:--

ROLACUN MAQ ILLUNA

the first two words being abbreviations for ROLACUNAS MAQI. The only letter that cannot be seen is the `C', which is broken away from the angle; it must have been there, however, as there is just space for it in the fractured surface'.

Macalister/1945, 386--387, regarding the left and right edges as separate inscriptions: `[Right-hand edge] After the first L there is a blank, with no consonants on the B side and no injury there. The H-surface is chipped away: the blank space would contain about 10 scores. There is a spall-matrix which would just hold a C, in the proper place for completing the common final CUNA (though the A is not here expressed). The only restoration of the name that would fit is ROLACUN or ROLOCUN. There is no room for the final A, which we should expect between the N and the M. After the M there is a space, rather wide for an A. It would hold a U, and we think of MUQOI for MUCOI. But this (despite the precedent of MOQOI at Monataggart (117) is too improbable, and there is no room for the vowels OI after the Q. The space here is spalled. It would hold an I, but no more. On the whole, despite the wide space after the M, the most probable restoration is MAQ. The doubled LL (which most certainly is not an S) might, but probably does not, begin the third word: we, therefore, treat the I as the initial, curtailing the second word to MAQ. After the double LL there is space for U, hardly for E. The whole reading is, therefore, as given above...[the] Ogham, which formerly read thus:--

PAMPES

now, however, [is] reduced by a fracture of the angle to the lower halves of the P's and the tips of the S scores. The P's are represented by crosses of three lines on the B side, evidently a simplification of the IA sign, which it would be difficult to cut without fracturing the surface. The photograph below shows a restoration of the broken edge in plasticine, the surviving parts of the scores being chalked...The broken angle has certainly been fractured with intention. When I made the restoration I thought that the reading should be POPES, the form in which the foreign name POMPEIUS would most likely be represented by a Goidel. But on second thoughts I concluded, rightly or wrongly, that the MP of so outlandish a name would be more likely to stand, and I, therefore, now prefer the restoration PAMPES. The modification does not impair the illustrative value of the photograph, which I have retained.

For the treatment of the vowel compare RASTECE = Rusticae at Llanerfyl (421)'.

RCAHMW/1976, 38: `The left-hand angle gives: P [O or A] P followed by an uncertain number of vowel notches, probably representing POPIAS, a personal name.[3] The right-hand angle may be read: R(O)L[...]N[...]M[...IQ[ ]LLUNA, presumably part of a name. The gaps from fractures are each large enough to have carried more than one letter and not necessarily vowel notches only; the gap between M and Q is far more than would be filled solely with A to make the common MAQI. The U is indicated by the unusual `forfid' form of a loop rather than notches (cf C.I.I.C. 240)...It is unlikely that Ogam characters formerly existed along the upper edge of the face but the two angles nonetheless form a single inscription.

[3] Arch. Camb., 1899, pp. 132-5 (Rhys)'.

RCAHMW/1976, 38: `much of the inscription is made uncertain by subsequent damage to the angles'.

Macalister/1945, 387: `on the narrower face on the stone'.

Nash-Williams/1950, 132: `The Latin inscription (Fig. 144) is in two lines reading vertically downwards'.

RCAHMW/1976, 38: `The Latin inscription set in two lines down the length of the face'.

It may be questioned whether in PVMPEIVS CARANTORIVS we have the epitaph of one person or two. Two names in the nominative, unconnected by any indication of relationship, meet us again at Llanddewi Brefi, in the inscription DALLVS DUMELVS. At Llangwarren, Pembrokeshire, there are two names similarly independent of one another, TEGERNACI DOBAGNI; and here again, one of those names is echoed in the associated Ogham, but not the other. The lettering on the Langwarren stone seems to indicate that the two names were cut by different hands, and presumably at different times, pointing to the conclusion that in that case two persons, perhaps two brothers, are commemorated. But neither the Llandewi Brefi nor the Kenfig inscriptions show any such want of uniformity in the lettering. The question may be left unanswered'.

Macalister/1945, 387: `That these two names in the nominative case represent two persons is suggested by the rendering of one of them in the Ogham'.

RCAHMW/1976, 38: `While the Roman name Pumpeius (for Pompeius) is probably rendered in the Ogams by an early Irish equivalent such as Popias, there seems to be no correlation in the remainder of the two inscriptions'.

Westwood/1879, 19--20: `Roman characters, PVMPEIVS CARANTORIVS; the only doubt being as to whether the second and third letters of the top line, which are conjoined, should be VM or VN. I prefer the former reading, because if read VN the N would be reversed, whereas it is of the proper form in the second line; whilst the transformation of the second stroke of the V into the first stroke of a conjoined M is of common occurrence. The letters were also read as VM in Camden, as mentioned below...`The Welsh Britans (as the Right Reverend the Bishop of Landaff who sent me the copy [of this inscription] informed me),' says Camden, `by adding and changing some letters thus read and make this interpretatio: Pim bis an ear Antopius [given in Gough's Camden Purnpbus ear a'n lopius], i.e. the five fingers of friends or neighbours killed us'. Holland's Camden, 1610, p. 645...As for the word Pumpeius for Pompeius, we have already observed (Lhwyd, Arch. Britan. vol. i. p. 17, col. 2) that in old inscriptions the letter V is frequently used for O [Gibson's Camden ii, p. 23]'.

Rhys/1899, 132--133: `That is to say, Pumpeius (more usually Pompeius) Carantorius. The first letters, P and C, have been considerably damaged, probably by boys throwing stones at the old monument, for it stands unprotected by the, road-side. The VN seem to form a conjoint character for VM, though they can hardly be said to be completely joined, but no more are the two limbs of the V; otherwise we should have to read V and {N} (or N reversed). The P has a rather small top: the E is of the rounded minuscule kind, and the S in both instances is well developed somewhat below the line of the other letters. The first A shows a bulging out at the ends of the horizontal bar, while the other A is more regular and somewhat broader at the top. The N is rather straggling and carelessly formed, while the tail of the H in both instances tends to be horizontal. Lastly, the O is circular, and, as frequently happens, it is rather smaller than the other letters'.

Macalister/1945, 387: `The VM in the first name is ligatured, and might pardonably be misread VN'.

Nash-Williams/1950, 132: `Roman capitals, picked and partly deepened by cutting, with half-uncial E and one ligature. The R's have the open bow and short, nearly horizontal tail. The S's are elongated'.

RCAHMW/1976, 38: `The Latin inscription set in two lines down the length of the face reads PVMPEIVS / CARANTORIVS, formed of fairly regular but debased Roman capitals with a single ligature; the E is half-uncial, the S's elongated, and both R's have an open loop with short tail. These features indicate a mid-6th century date'.