- Chapter Thirty-Six -
The relationship of language to both nationality and status underwent major
changes in the early medieval period, just as the languages themselves were trans-
formed in the late Roman and post-Roman periods. I shall take the kingdom of Dyfed
in the south-west of Wales as a specimen case, for it is easy to show that, early in the
period, Dyfed was a land of three languages, while by the end it was, for most
purposes, a land of only one. The early medieval kingdom of Dyfed was heir to the
Romano-British civitas of the Demetae: reflecting this inheritance, the early inscrip-
tions of the area are predominantly in Latin. Moreover, the character of the Latin
used in the inscriptions demonstrates that it was then a spoken language, not merely a
language of the quill and the chisel. On the other hand, from at least the sixth century
until the ninth Dyfed was ruled by a dynasty for which Irish ancestry was claimed.^4
In much of the kingdom, in the county of Pembroke and the western part of Car-
marthenshire, an Irish presence in the early Middle Ages is confirmed by inscriptions
written in the ogam alphabet and in an early form of the Irish language.^5 The inscrip-
tions show therefore that Dyfed was far from being just a British kingdom.
Yet by the mid-tenth century another view of the past had been developed for a
new royal kindred of Dyfed. It shows how the Irish presence in Dyfed had come to
an end, but also how a British identity was partly defined in terms of an expulsion of
Irish settlers in western Britain. Early in the tenth century Dyfed was drawn into the
conglomeration of lands ruled by the descendants of Merfyn Frych (ab. 844) and his
son, Rhodri Mawr (ab. 877).6 Both Merfyn and Rhodri were kings of Gwynedd and
founders of the dynasty called by modern scholars the 'Second Dynasty of
Gwynedd'. This dynasty, however, soon extended its power beyond Gwynedd into
Powys and later into Ceredigion and Dyfed. The earliest surviving collection of
Welsh genealogies takes as its initial focus Owain ap Hywel ap Cadell ap Rhodri, the
king of Dyfed.? First it gives his patriline and then the descent of his mother from the
old kings of Dyfed who ruled before the sons of Rhodri gained control over
the south-west. So much is the normal genealogical testimonial of the time: first the
agnatic descent and then prestigious connections through females. What is more
surprising is that the line of the old kings of Dyfed is traced back, not to any Irish set-
tlers, but to an evidently fictitious 'Nyfed son of Dyfed son of Maxim Wledig'. The
personal name Nyfed perhaps echoes the Irish use of nemed 'sacred' for a person of
high status as well as for a sanctuary.8 Dyfed is the name of the people transmogrified
into the name of a distant ancestor.^9 Nyfed, therefore, was probably chosen simply to
rhyme with Dyfed. Maxim Wledig (later Maxen Wledig), on the other hand, is an
4 Tairired na nDesse, ed. and transl. K. Meyer, 'The expulsion of the Dessi', Y Cymmrodor
14 (1901): 112-13, § II; T. OCathasaigh, 'The Deissi and Dyfed', Eigse 20 (1984): 1-33-
M. Richards, 'Irish settlements in south-west Wales: a topographical approach', Journal of
the Royal Society of Antiquaries of Ireland 90 (1960): 133--62, discusses the place-name as
well as the epigraphical evidence.
6 J.E. Lloyd, A History of Wales from the Earliest Times to the Edwardian Conquest, 3rd
edn (London 1939), 333.
7 ob. 988 , Annales Cambriae (MSS B, C) (ed. J. Williams ab !the! (Rolls Series; London,
1860), 21). The Harleian Genealogies of the tenth century are edited by P.C. Bartrum,
Early Welsh Genealogical Tracts (henceforward EWG]) (Cardiff, 1966),9-13.