- Language and Society among the Insular Celts AD 400-1000-
entirely genuine historical person (Magnus Maximus), but for the genealogists he was
the link between Britain and its Roman past: Gildas's tyrant who removed the
military force of Britain and so laid it open to Saxon attack was still the man 'who
killed Gratian the king of the Romans',lO but he was also made into a descendant of
Constantine the Great and the ancestor of the dynasty of Dyfed as well as of other
royal lines. Romanitas was now like an ancient title-deed: it was important that
descent from Christian Roman emperors could be displayed in case of need, but the
claim looked to another age and the title-deed usually gathered dust in the archive.
Yet if romanitas was not a present force, Roman descent was at least not Irish descent.
The line of Irish kings with their dofreth 'billeting' « Early Irish "damreth) which
they imposed upon the land of Dyfed had been pushed aside.ll Tenth-century Dyfed
was a British kingdom whose kings enjoyed the high dignity of descent from
Constantine and from his mother Helen, 'who went from Britain as far as Jerusalem
to seek the Cross of Christ, and she brought it with her to Constantinople, and there
it remains to the present day'.12 There had been a conscious rejection of the Irish
past of Dyfed and a conscious stress on a Roman past.
Already by the mid-tenth century, therefore, the surviving evidence of Irish
settlement consisted of the odd fossil - some place-names, some inscriptions in an
alphabet now probably unintelligible, some loan-words. It was known that there had
been Irish settlers, but now they were often not considered people with whom good
men consorted. A version of Rhigyfarch's Life of St David tells his readers that, in
order to found St David's, the saint had been obliged to deal with an Irish chieftain
Boia and his formidable wife.^13 The origin of the premier church of Wales lay in a
8 The latter sense is attested in the Continental and British Celtic nemeton, A. Holder, Alt-
celtischer Sprachschatz (1891-1913), s.v. nemeton; Bechbretha, ed. T. Charles-Edwards and
F. Kelly (Dublin, 1983), pp. 107-8. Compare also Nemetes, Nemetu, J. Whatmough, The
Dialects of Ancient Gaul (Cambridge, Mass., 1970), pp. 1093, 1290, Ednyfed < Iudd Nyfed
'lord of a sanctuary' or, perhaps, 'sacred lord'.
9 This may have been encouraged by biblical precedent, for example Israel, both the people
and the alias for the ancestor, the patriarch Jacob. Cf. Glywys, probably from Gleuenses,
the people of Gloucester and its region, which also appears as a personal name, ECMW no.
255 (and possibly no. 239). Demetus was likewise used as a personal name in ECMW no.
390, but the Dyfed supposed to be a person and the ancestor of the kings of Dyfed must be
a counterpart to the biblical habit of naming peoples by the ancestors of their kings.
10 Gildas, De Excidio Britanniae, c.13; EWGT, p. 10, § 4. Compare ECMW no. 182 and
EWGT, p. 2 (the Pillar of Eliseg), qui occidit regem Romanorum, and J.F. Matthews,
'Macsen, Maximus, and Constantine', Welsh History Review II (1982-3): 431-48.
I I Irish ddmrad < "-ddmreth was used for the company of a king or other great lord; but in
Dyfed, betraying the point of view of the subject, dofreth came to mean the obligation to
provide quarters for men, notably from the company of the king, who were billeted upon
his subjects.
12 EWGT 10, § 2.
13 The identification of Boia (or Baia) as an Irishman is found in the Vespasian Recension of
the Life and its Welsh offshoot, Buchedd Dewi: see Rhigyfarch, Life of St David, ed. J.W
James (Cardiff, 1967),9; Buchedd Dewi, ed. D.S. Evans (Cardiff, 1965), 7; but it is also
found in the poem by Gwynfardd Brycheiniog, Hen Gerddi Crefyddoi, ed. H. Lewis
(Cardiff, 1931), p. 50 (XVIII, 225).