One manuscript, written around 1325 and now in Aberystwyth, contains roughly sixty poems attributed to a poet named Taliesin. Perhaps a dozen of them were composed by a man who lived around 580 and worked for a king in what is now the Scottish border country. The rest were written by other people over the following seven hundred years, in his voice, and some of them claim he was present at the Crucifixion and the fall of Lucifer.
Sorting one from the other is the central problem of early Welsh literature, and it is not fully solved.
The historical poet
The Historia Brittonum, compiled in the ninth century, names five poets who were famous in Welsh verse in the sixth: Talhaearn, Aneirin, Taliesin, Blwchfardd and Cian. Two of those names have a surviving body of work attached, and Taliesin's is the larger.
The poems that scholarship treats as genuinely early are praise-poems and elegies for Urien of Rheged and his son Owain — a kingdom of the Old North, centred somewhere around the Solway Firth, fighting Northumbrian Angles. They are technical, professional, specific: they name battles, count cattle given as reward, and mourn a patron with real force. Ifor Williams's edition in 1960 made the case for isolating this core, and it remains the standard framework even where individual attributions are argued.
These are documents of a British-speaking kingdom in northern England and southern Scotland, written in an ancestor of Welsh, several centuries before anything survives from Wales itself.
The other Taliesin
The rest of the manuscript is a different creature. There are religious poems, prophetic poems foretelling the return of British rule, poems of learned riddling, and a group of extraordinary compositions in which the speaker claims to have been everything.
He has been a blue salmon, a stag, a roebuck on the mountain, a stock, a spade, an axe in the hand, a grain that grew on a hill, a word in a book. He was at the fall of Lucifer, in the ark with Noah, at the court of Don before the birth of Gwydion. He knows why the sea is salt, why the night is dark, and what supports the world.
Cad Goddeu, the Battle of the Trees, is the most famous of these — a poem in which Gwydion enchants the trees of the forest into an army, each species described in a line or two, alder in the front rank and oak causing the ground to shake.
These are not the utterances of a druid. They are the work of medieval Welsh poets in a learned tradition, performing omniscience as a professional credential and drawing on biblical, classical and native material at once. Whether some of the imagery preserves genuinely older ideas is arguable; that the poems as we have them are medieval is not.
Gwion Bach, the weir, and the radiant brow
The story that explains all of this is a late one. Ystoria Taliesin, whose full form is sixteenth-century, has him born twice: first as Gwion Bach, a servant who accidentally takes the three drops brewed by Ceridwen, and then — after the chase through animal shapes and his consumption as a grain of wheat — reborn from her and set adrift in a leather bag.
Elffin ap Gwyddno finds the bag at a salmon weir on the eve of May, expecting his annual catch and getting a child. Uncovering the forehead, he exclaims at the radiance of the brow, and the infant answers in verse. Tal is brow, iesin is radiant, and the name is explained.
The remainder of the tale is a professional revenge narrative. Elffin, imprisoned by Maelgwn Gwynedd for boasting about his wife and his poet, is freed when Taliesin walks into the king's hall and reduces the twenty-four court bards to blowing on their lips and saying nothing but blerwm, blerwm, then raises a wind that shakes the fortress until Elffin's chains fall off. A poet who can silence a court and unmake a king's confidence is the fantasy of every professional in that trade.
Robert Graves and the Battle of the Trees
The White Goddess of 1948 is built substantially on Graves's handling of Cad Goddeu, which he treated as a scrambled cipher to be reassembled — recovering from it a tree alphabet, a calendar and a lost matriarchal religion.
He did not read medieval Welsh, worked from a Victorian translation of a difficult poem, and reordered lines to produce the result he wanted. The book is a remarkable piece of imaginative writing and it is not evidence about anything Welsh. A very large proportion of what circulates today as ancient Celtic tree lore descends from those pages. Graves was candid that his method was poetic; his readers have been less so.
From Rheged to Wisconsin
His modern afterlife is unexpectedly wide. Arthurian fiction adopted him early and often, from Stephen Lawhead's novels to the Dark Is Rising sequence, generally as the archetype of the bard-magician. His name is now in use as a Welsh given name.
And Frank Lloyd Wright, whose mother's family was Welsh, named his Wisconsin home Taliesin and his Arizona winter camp Taliesin West — so the most consequential American architect of the twentieth century worked and taught under the name of a sixth-century praise-poet from the Solway. The buildings are far better known than the poems.
Was he a druid?
No, and the distinction is worth keeping sharp. He was a court poet in a Christian British kingdom, employed to praise a king, and the historical poems contain nothing pagan. Caesar does report that Gaulish Druids taught a doctrine of souls passing into other bodies, and the shape-shifting poems have often been pointed to as an echo of it — but those poems are medieval, they were written by Christians, and their transformations are a rhetorical device claiming universal knowledge rather than a statement of belief about reincarnation.
How can we tell which poems are early?
Chiefly by language and by content. The historical poems use older grammatical and metrical forms and refer to people, places and conflicts that fit sixth-century northern Britain; the later ones show developments in the language, mention events and texts from after that period, and belong to genres — prophecy, biblical learning — that emerge later. The method is well established and the boundaries are still argued at the edges, which is why counts of the "genuine" poems vary between editors.
Is Rheged's location known?
Not precisely, which is a genuine frustration. It was a British kingdom of the Old North, generally placed around the Solway Firth and Cumbria, with Carlisle a strong candidate for a centre; a major early medieval settlement excavated at Trusty's Hill in Galloway has been proposed as another. The poems are among the best evidence for its existence, which puts a kingdom's historicity partly in the hands of literary criticism.