In the history of the various Celtic peoples there is evidence for the existence of more than one Heroic Age. In the first place we have some traces of heroic poetry among the ancient Gauls, though unfortunately it has entirely perished, together with all their vernacular records. Then again Ireland has preserved a great body of heroic literature, referring to a remote period. But, whatever may have been their original form, these stories have come down to us for the most part only in prose. Moreover, the subject is beset with so many difficulties, both historical and linguistic, that it cannot be approached with safety except by an expert. We shall have to confine our attention therefore to the heroic poetry of the Cymry.

Even here the difficulties to be encountered are sufficiently serious. Yet it is fairly clear that many of these poems deal with events which are referred by the chronicles to the sixth century and the first half of the seventh. We may note especially four groups of poems: (1) a few concerned with the exploits of Arthur and his heroes; (2) a somewhat larger number referring to princes named Gwallawg, Urien and Rhydderch; (3) certain poems dealing with Gododin and Catraeth; (4) a few relating to Cadwallon. There are also two or three others, such as the elegy over Cynddylan, which seem to refer to the same period, though they cannot be classed under any of these headings. A number of the above poems, especially those included in the second and third groups, are attributed to two poets named Taliessin and Aneirin.

There is no doubt that Cadwallon is the well known Welsh king who overthrew the Northumbrian king Edwin in 633 and was himself destroyed by Oswald in the following year. But the poems of the second group seem also to have a historical basis. In the Historia Brittonum (Harleian text), § 63, we hear of four kings, Urbgen et Riderch Hen et Guallanc et Morcant, who fought against the Northumbrian king Theodric and his successors, Frithuwald and Hussa. Urbgen is said to have besieged his opponents for three days in Lindisfarne (Metcaud), and to have perished eventually through the jealousy of Morcant. In another passage (§ 62) we hear of Neirin et Taliessin among other poets who composed British poetry in the time of a king named Dutigirn[157], who is made contemporary with Ida, the father of Theodric. The origin of these entries is unfortunately obscure. They are incorporated with an English genealogical document, dating apparently from the end of the eighth century; but they may quite possibly be derived from earlier sources. In any case they cannot well be later than the first quarter of the ninth century.

All the four kings mentioned in § 63 are known to us also from genealogies of the tenth century. But one of them, Riderch Hen, is named in Adamnan's Life of St Columba, which is of course altogether independent of Welsh tradition. It is there stated (I 15[158]) that Rodercus, king of Dumbarton, consulted St Columba as to his fate and received the answer that he would die a peaceful death in his house—a prophecy which was subsequently fulfilled. As St Columba died probably in 597, the date given for Riderch Hen agrees well enough with what is stated in the Historia Brittonum; for according to the most trustworthy records Theodric and his two immediate successors reigned from about 572 to 592 or 593. We have no ground for doubting that the references to Urbgen and the rest are based on equally good tradition.

For the characters mentioned in the Gododin poems no such evidence is available. The poems themselves however contain references to the death of Dyvynwal Vrych—doubtless the Dalriadic king Domnall Brecc, who according to the Irish annals was killed by the Britons three years after the fall of Oswald, i.e. about the year 645. Unfortunately it is not made clear in what relationship this event stands to the main action of the poems[159]. The few identifications which can be made from the genealogies seem to be compatible with a date somewhere about this time.

There can scarcely be any doubt that we are justified in regarding the period covered by the poems as a kind of Heroic Age. The story of Arthur was the one most elaborated in later times—a process which must have begun before the ninth century, when it was incorporated in the Historia Brittonum (§ 56). How much historical truth this story contains we do not know, though the chronicles refer Arthur to the early part of the sixth century, i.e. the beginning of our period. To a certain extent however the other stories have experienced the same process; for, apart from the question when the poems under discussion were composed, there are others of undoubtedly later date, such as the Englynion y Beddau, in which the same heroes are mentioned. Moreover it must not be assumed that these were the only persons of the same period whose praises were celebrated in poetry. Maelgwn, king of Gwynedd, who died about 548, figures prominently in traditions of later times, and it is likely enough that the beginnings of these stories were due to a similar cause. On the other hand there is apparently no evidence for the composition of poems dealing with the exploits of princes who lived in the latter part of the seventh century or for several centuries later.

As to the origin of the poems discussed above we have no evidence which can be called conclusive. Some of the poems themselves claim to be the work of the poets Aneirin and Taliessin. At the present time however it seems to be generally held[160] that they are scholastic products of a later age, based upon chronicles and composed with the intention of glorifying the ancestry of some distinguished Welsh families. In support of this view it is contended that the language of the poems cannot possibly represent the form of Welsh spoken in the sixth or seventh centuries. But this argument is scarcely decisive; for it is clear that parts of the poems have undergone a certain amount of modernisation, and it is by no means incredible that the same process may have been in operation for centuries. I find it difficult to believe that in order to celebrate the ancestry of certain families poems would be composed recording an action which is represented as an overwhelming disaster for the British forces, such as that described in the Gododin poems. And can it really be proved that many of the heroes celebrated in these poems were claimed as ancestors by families in Wales?

At all events it is clear that the author of the entries in the Historia Brittonum knew of ancient British poets. Neirin and Taliessin are only two out of five names recorded by him, the rest apparently being altogether unknown from other sources[161]. It is true that he places all these persons about a generation before Urien and Rhydderch Hen, but chronological accuracy is hardly to be expected in references to such a remote period. Certainly we should note that the last British king mentioned in these entries is a certain Cadafael (Catgabail Catguommed), who is said to have escaped alone from the battle in which Penda was killed (A.D. 655). The wording of this passage (§ 65) seems rather to suggest acquaintance with a poem on the subject.

We have seen now that the persons and events celebrated in the poems belong to a period extending roughly from 500 to 650 and that there are ancient records of famous British poets, who are referred to the earlier part of this period and of whom two are traditionally associated with the poems. Next we must notice that the poems themselves, even in the corrupt and often unintelligible form in which they are preserved, plainly show all the marks of Stage I of our scheme. Their characteristics are those of court-poetry, but they never attain to true narrative. For analogies to the poems which deal with Urien we can hardly do better than turn to the court-poetry of the Viking Age (cf. pp. 15 f., 91); for the Gododin poems perhaps the best parallels are to be found in the Battle of Maldon and the Story of Igor's Expedition, though in both the latter the narrative element is much more fully developed. Taking all the evidence into account, it seems to me that the choice lies between two alternatives: either the poems really are works of Stage I, which have survived from the Heroic Age itself, or they are exceedingly clever imitations of such works, composed at a time when the latter were still in existence. The decision between these two alternatives must of course be left to experts. In the meantime we shall be ready to admit that the poems are extremely corrupt, and that in some cases perhaps we have nothing more than disjecta membra of earlier pieces. But until conclusive evidence is brought forward it seems to me highly improbable that the original poems have been entirely lost[162].

The history of Cumbrian heroic poetry is easily intelligible when taken in connection with the national history. It is probable enough that on the withdrawal of the legions large portions of the country were occupied by northern Britons from beyond the Wall—who are represented in tradition by Cunedda and his sons[163]. These were doubtless affected by Roman influence to a very much less degree than the inhabitants of the province; and the growth of heroic poetry during the sixth century may be satisfactorily accounted for by the wealth and prosperity which would naturally accrue to their princely families during the early days of their dominion. It was probably not until the time of Aethelfrith that their power was really broken. But the following half century was a most disastrous period, ending in the permanent obliteration of their nationality—in a political sense—throughout the whole region between the Dee and the Clyde. Any records which survived must owe their preservation, ultimately at least, to refugees who escaped into the Welsh highlands.

Now we can see why so many place-names (Reged, Catraeth and dozens of others) are incapable of identification. But further, the conditions in Wales, the poorest and probably the most backward part of the province, were doubtless highly unfavourable to the development of heroic poetry. Under such conditions we should expect that narrative poems would rapidly disintegrate into the semblance of folk-tales; and it may be that here we have the explanation of the medieval stories of Arthur. At all events they show a certain resemblance to the Russian stories of Vladimir and his heroes. But Arthur, as we have seen, belongs to the beginning of the Heroic Age. We have no evidence for the composition of poems, which can properly be called narrative, dealing with the later heroes. Here the original type of poem seems to have become stereotyped—presumably because conditions favourable to development were no longer present. The only marvel is that so much has been preserved.