Welsh Journals

and fix his place in his community.1 Clearly, since few communities have a static social or economic life, it will be necessary to allow an area of ambiguity 12 in each pedigree structure so that necessary adjustments may be made conforming to, and conferring legality on, contempory changes. Other changes which mar the chrono- logical validity of a pedigree may also occur whereby several members bearing the same name may be recorded as one single entry, and this entry may appear at an arbitrary point on the whole. Equally, unless the community holds to the principle of son suc- ceeding father, a king-list which emerges from it will not represent a dynasty in the modern European sense, even though the names on it may be connected by the conventional phrase son of'. It is not certainly known what rules governed the succession to each of the medieval Celtic kingdoms of Britain during the earliest period. Therefore it is difficult to distinguish what information the royal Welsh pedigree lists may be giving. However, once the obvious accretions have been removed, this does not mean that the lists of remaining names are quite arbitrary, being unconnected with each other or with the area to which they are attached, even though they cannot with assurance be made up into a family tree of the modem European kind. Primitive memory is tenacious of the names of rulers, or of some rulers, even though their chronological order may become confused and their blood relationships altered or invented. These are the names that seem to be recorded in the early Welsh lists. The non-genealogical traditions which gather around names that appear on a king-list may be stripped down to reveal a sub- stratum of historical truth. This in turn may offer clues about the careers of men who afterwards were represented as successive generations of one royal family, and who therefore are likely to have had some kind of real connexion with each other and with the group that the genealogical tradition believed them to rule, even if the king-list represents only a short-hand and conventionally expressed version of these connexions. In other words, if listed names were later considered to be connected with each other and with a specific social group, in this case the men of Dumnonia, and historical information about some of the names which appear on the list may be deduced from other sources, then it is probable that conclusions may be drawn about the history of the group and its rulers as a whole. Of all the surviving king-lists, that of Dumnonia contains the largest number of names that figure very prominently in traditions that are not genealogical, but rather poetic and hagiographical. 1 E. Peters, J. Roy. Anthrop. Inst., 1960, 29f. 2 E. Peters, J. Roy, Anthrop Inst., 1960, 41.