Documented: no Basque term Atlantica and no sunken-homeland myth appears in the standard ethnographic corpus, Barandiaran's collections, the Etniker archives. Basque mythology's documented chthonic figures are Mari, Sugaar, the lamia, basajaun, none tied to a drowned island. The Basque-Atlantis linkage is documented instead as 19th-20th century romanticism, from Fabre d'Olivet's Atlantis-as-Basque speculation through the modern video circuit, projecting Plato onto the Pyrenees.
The folk logic is understandable: isolate language, odd blood, ancient roots, proximity to the Tartessos coast and the drowned Doggerland-scale shelves of the Bay of Biscay, so the lost-civilization slot fills with the nearest mysterious people. The genuine Basque landscape memory is documented elsewhere: coastal legends of drowned towns do exist in Iberian lore, just not as Atlantica and not in the Basque corpus.
The claim: the Basques are the remnant whether or not they remember it, the language and blood doing the remembering the myths refuse to, Atlantica as the name the archive lends the gap. The counter is the archive's own discipline: a people whose documented survival mechanism is precision deserves better than a borrowed flood.
When a people this real gets assigned a homeland this unverifiable, who is the legend serving?