Two words doing far too much work
“Celtic music” is a phrase of roughly the same age as the phonograph. It is a marketing category, a scholarly convenience and a genuine cultural inheritance all at once, and the trouble begins as soon as you try to make it be only one of those things. The instruments it refers to are mostly medieval or later. The languages it is sung in are ancient. The repertoire it draws on was largely written down between 1700 and 1900. The audience it addresses was assembled in the 1970s.
None of that makes it fake. It makes it a tradition rather than a relic, which is a considerably more interesting thing to be. But a performer who wants to work honestly inside it has to know which parts of the story are evidence, which are inference and which are nineteenth-century romance wearing a very convincing cloak. This essay is my attempt at that inventory, written from the position of someone who makes new work inside the tradition rather than curating it.
The archaeological Celts: Hallstatt and La Tène
The people the Greeks called Keltoi and the Romans called Galli are visible to us first through objects rather than words. From roughly 800 BCE, a material culture archaeologists name after the Austrian salt-mining site of Hallstatt spreads across central Europe: iron weapons, wagon burials, a distinctive way of ornamenting bronze. From about 450 BCE it is succeeded by the La Tène style, named for a site on Lake Neuchâtel in Switzerland — and it is La Tène that gives us the flowing, asymmetrical, endlessly interlacing ornament most people picture when they hear the word “Celtic”.
Those cultures left musical instruments, or at least the hard parts of them. The most spectacular is the carnyx, a tall bronze war-trumpet held vertically with a bell shaped like a boar or serpent head, held high above the heads of the fighting men so the sound carried over the ranks. Polybius, describing the Battle of Telamon in 225 BCE, complains about the noise of the horns and the shouting as if it were itself a weapon. Reconstructions made from the Deskford carnyx found in Banffshire, Scotland, produce something between a trombone and an animal — and when you hear one played you understand immediately why a Roman historian bothered to mention it.
What the carnyx tells us, and what it does not
It tells us that these were people with sophisticated metalworking, a taste for theatrical sound and the organisation to deploy it. It tells us nothing about scales, modes, song forms or what anybody sang at home. The gap between “they had loud instruments” and “here is their music” is roughly two thousand years wide, and no amount of wanting closes it.
This is the first and most important honesty: we have no Iron Age Celtic melodies. Not one. Anyone who tells you a tune is “two thousand years old” is telling you something they cannot know. What we have instead is a long, unbroken habit of music-making in the Celtic-speaking world, and habits leave deeper traces than tunes do.
The bardic orders: music as an institution
By the early medieval period the picture sharpens considerably, because the Celtic-speaking societies of Ireland, Wales and Scotland had turned poetry and music into a profession with grades, training, legal status and pay scales — and lawyers write things down.
Irish law tracts distinguish carefully between ranks of poet, the highest being the ollamh, a title that required some twelve years of training and carried a legal honour-price comparable to that of a minor king. The fili composed praise, satire and genealogy; the bard sat lower in the hierarchy. Welsh tradition has the parallel figure of the pencerdd, the chief of song, and the tenth-century laws attributed to Hywel Dda specify what a court harper was owed and what happened to a man who damaged a harp.
Three things follow from this, and all three matter to anyone performing today:
- Music was memory technology. Genealogies, land claims, legal precedents and origin stories were held in verse because verse survives transmission better than prose. The seriousness with which the tradition treats lineage is not sentimentality; it is the residue of an actual filing system.
- The word came first. In the bardic hierarchy the composer of text outranked the instrumentalist. The tune served the metre. This is why so much traditional Celtic song sits uncomfortably in strict bar lines — the phrase length is following the poetry, not a metronome.
- It was a patronage art. When the patrons went, the institution went. The collapse of the Gaelic aristocratic order in Ireland after 1601 and in Scotland after 1746 did not end the music, but it ended the profession that had sustained its highest register, and the repertoire that survived is disproportionately the domestic and the popular rather than the courtly.
The harpers and the last of the old repertoire
The harp is the one instrument with a continuous documented line from the medieval court to the present, and its history is the history of the whole tradition in miniature. The Irish cláirseach and the Scottish clàrsach were wire-strung — brass, and later sometimes silver or gold — played with the fingernails, producing a ringing, bell-like sustain utterly unlike the gut- or nylon-strung harps most players use now. Three medieval instruments survive: the Trinity College harp in Dublin (the one on the coins and the stout bottles), and the Queen Mary and Lamont harps in Edinburgh.
The harpers were the last professionals of the old order, and the most famous of them, Turlough O’Carolan (1670–1738), is also the clearest illustration of what the tradition actually is. Blinded by smallpox at eighteen, trained as a harper, he travelled Ireland for nearly fifty years composing planxties for patrons — and his music is saturated with the Italian baroque he heard in the great houses. Corelli and Vivaldi are audible in his phrases. He is simultaneously the most quintessentially Irish composer of his century and an enthusiastic importer of foreign fashion. That is not a contradiction in the tradition. That is the tradition.
The endpoint is precisely dateable. In July 1792, at the Belfast Harp Festival, ten of the last traditional harpers were assembled to play, and a nineteen-year-old organist named Edward Bunting was engaged to write down what they played. He came away with a notebook that became the foundation of Irish music collecting — and with the uncomfortable knowledge that he was transcribing an instrumental technique already almost extinct. Bunting himself noted that the harpers were mostly elderly and that nobody was learning the old fingering.
Collectors, romantics and the invention of a category
What happens next is the most consequential century in the story, and it is largely a story about editing. From Bunting onwards, the surviving music passed through the hands of collectors who were educated in European art music, wrote in piano-friendly keys, and frequently could not resist improving what they heard.
Thomas Moore set Bunting’s airs to new English lyrics in his wildly successful Irish Melodies, and in doing so gave a great many tunes the words most people now think they always had. In Scotland, James Macpherson’s Ossian poems — presented in the 1760s as translations of an ancient Gaelic epic and almost certainly substantially his own composition — set half of literary Europe alight and taught the Continent to hear the Gaelic world as misty, melancholy and sublime. Mendelssohn wrote the Hebrides overture inside that weather.
Later collectors pushed back towards fidelity. Francis O’Neill, a Chicago police chief born in County Cork, published in 1903 a collection of more than eighteen hundred Irish tunes that is still the working reference; Marjory Kennedy-Fraser’s Songs of the Hebrides did the opposite, arranging Gaelic song for drawing-room piano so thoroughly that scholars still argue about what is underneath. The Alan Lomax field recordings of the 1950s finally let the tradition be heard rather than read.
The honest position for a performer
Almost every “traditional” setting you will sing is an edited artefact of the last three hundred years. This is not a scandal; it is simply the material condition of the work. The useful response is not to chase an unrecoverable authenticity but to know which layer you are standing on — Bunting’s, Moore’s, O’Neill’s, or your own — and to say so.
The languages, and why they change the music
Six Celtic languages survive in some form, and they divide into two branches whose differences are audible in every phrase sung in them.
The Goidelic or Q-Celtic branch — Irish, Scottish Gaelic and Manx — kept the Proto-Celtic kw sound as a k. The Brythonic or P-Celtic branch — Welsh, Cornish and Breton — turned it into a p. The textbook illustration is the word for “son”: Irish mac, Welsh map, now ap in Welsh patronymics. It sounds like a footnote for philologists. It is not, because the two branches also diverged in stress, in vowel length and in consonant texture, and singers live entirely inside those three things.
- Stress placement. Irish and Scottish Gaelic stress the first syllable of a word with near-total consistency. Welsh stresses the penultimate. That single difference produces completely different melodic shapes: Gaelic song tends to land on the downbeat and decay, while Welsh song leans forward into the second-to-last syllable and gives the language its characteristic lilt.
- Broad and slender consonants. Irish and Gaelic consonants exist in two forms — velarised (“broad”, coloured by a, o, u) and palatalised (“slender”, coloured by e, i). The distinction is not an accent; it carries meaning, and a sung line that ignores it is genuinely harder for a native speaker to understand.
- Pure long vowels. English diphthongises almost everything — our “oh” is really “oh-oo”. Gaelic long vowels hold their shape from onset to release. This is why Gaelic song sits so well on a long sustained note, and why English-language singers approaching the repertoire often sound as though they are sliding.
- Consonant clusters and mutation. Welsh and Breton mutate initial consonants grammatically, so the same word begins differently depending on what precedes it. A singer working from a printed lyric without the grammar will produce forms that are simply wrong.
Then there is sean-nós — “old style” — the unaccompanied Irish-language singing tradition, and the closest thing the Celtic world has to a conservatory of ornament. Its defining habits are melodic variation from verse to verse, a free rhythm that follows the text rather than a pulse, nasal and glottal articulation used deliberately, and a complete absence of the dynamic swell that Western art singing is built on. Its regional dialects differ sharply: Connemara is the most ornamented, Munster the most melodically elaborate, Ulster the sparest and closest to speech.
Why this matters even if you sing in English
The ornament, the phrase lengths and the cadence habits of the whole tradition were shaped by these languages. A singer who has spent a month with sean-nós recordings will phrase an English-language ballad differently — not because they are imitating anything, but because they have absorbed where the tradition thinks a line should breathe.
The revival, and why 1970 matters
The tradition never stopped in the places where it lived. What changed in the twentieth century is that it acquired an international audience and, with it, a shelf in a record shop.
The ceili band boom and the founding of Comhaltas Ceoltóirí Éireann in 1951 institutionalised session playing. Seán Ó Riada’s ensemble Ceoltóirí Chualann in the early 1960s took traditional material and arranged it with a deliberate, almost classical sense of texture; out of it came The Chieftains. Planxty and The Bothy Band brought in bouzouki, rock-band energy and arrangement thinking. In Brittany, Alan Stivell’s 1972 Renaissance de la Harpe Celtique did for the Breton harp what the revival had done for the fiddle, and made “Pan-Celtic” an idea you could buy on vinyl.
By the 1990s the category had gone fully commercial. Riverdance in 1994 exported a choreographed version of the tradition to stadiums. Enya, working in a studio full of overdubbed vocal layers, and Loreena McKennitt, following the Celtic migration route east through Europe into Asia Minor, created a sound that was not folk, not classical and not pop, and that gave many listeners — me among them — a door in.
It is worth saying plainly what that sound owed to the recording studio. The stacked, reverberant, wordless vocal that defines a great deal of contemporary Celtic music is not a folk technique. It is a production technique, made possible by multitrack tape and then by plate reverb, and it is about as traditional as a synthesiser. It is also beautiful, and it has become its own convention with its own disciplines.
Six honest propositions to work from
This is where I have landed after two decades inside the material. I offer them as working propositions rather than conclusions.
- There is no Celtic scale. The modes used across Ireland, Scotland, Wales, Brittany, Cornwall, Galicia and the Isle of Man are ordinary European church modes — Ionian, Dorian, Mixolydian, Aeolian — often with gapped, pentatonic melodic habits. What marks a tune as belonging is ornament, phrasing and cadence habit, not exotic intervals.
- The drone is structural, not decorative. Pipes, hurdy-gurdy and wire-strung harp all supply sustained pitch. Harmony in this tradition is something that happens over a fixed centre, which is why functional chord progressions so often sound wrong on traditional material.
- Ornament is grammar. Cuts, rolls, crans and grace notes are not decoration added to a melody; on the pipes and whistle they are how articulation is achieved at all, because there is no tonguing and no dynamic control. Sung ornament inherits that logic.
- The languages are not interchangeable. Irish, Scottish Gaelic, Manx, Welsh, Cornish and Breton divide into two branches with genuinely different sound-worlds. Singing them without the stress patterns is like singing Italian with German vowels — intelligible, and wrong.
- The tradition has always imported. O’Carolan took the Italian baroque; the Breton revival took the American folk boom; the 1990s took the recording studio. Fusion is not a departure from the tradition. It is the mechanism by which it has stayed alive.
- Say where you stand. If you have written a new melody, say so. If you have taken a Bunting setting and rebuilt it, say that too. The audience for this music is, in my experience, entirely happy with invention and entirely unforgiving of pretending.
Everything I write comes out of that last point. The songs on The Ancestry are new; the history they sit inside is real; and the join between the two is something I would rather show than hide.
Celtic music history Hallstatt La Tène bardic tradition Carolan folk revival Pan-Celtic
Written in Moorhead, Minnesota. Corrections and counter-arguments are welcome at [email protected].
