The ceiling on how long a story can remember

Claim (verbatim)

If oral traditions can encode datable geological events - a volcano's birth, a coast's drowning, a meteor fall - then the record of such 'geomyths' contains a hidden quantity: the maximum age of a factual memory that transmission can actually carry. Individual claims run from the plausible to the extravagant. Nunn and Reid's drowned Australian coasts imply seven to thirteen thousand years. The Klamath account of the fiery destruction that formed Crater Lake maps onto the eruption of Mount Mazama, dated by tephrochronology to about 7,700 years ago. The Gunditjmara traditions around Budj Bim are sometimes tied to volcanism whose youngest dated eruptive phase, by recent argon dating, is roughly 37,000 years - which, if the link held, would be an outlier dwarfing everything else. The disciplined move is to stop arguing each case and instead census them: assemble every geomyth with an independent geological date and look at the distribution of implied transmission spans. A real ceiling near the low tens of thousands, with the extreme claims standing alone on single contested geological links, is a finding; a smooth spread to 100,000 years would be another. The missing knowledge is the shelf-life of accurate oral memory, and it is a countable distribution.

Prediction clause (verbatim)

Prediction: in a census of at least 15 geomyths each carrying an independent geological/geochronological date, at least 90 percent will imply transmission spans under 13,000 years, and any claim implying more than 13,000 years will number two or fewer and will each rest on a single geological link rather than converging multiple independent anchors (primary clause: >=90 percent under 13 ka, with above-ceiling claims <=2 and each single-anchored). Disambiguation: a case counts only if the myth names a specific event feature and an independent date exists for that feature; vague catastrophe motifs are excluded. Coverage guard: if fewer than 15 adequately date-anchored geomyths can be assembled the distribution voids for insufficient sampling rather than resolving.

Kill-dataset (verbatim)

Kill (not yet built): a census table built from real geomythology sources - Nunn & Reid (2016) for the drowned coasts, Mazama tephrochronology (Bacon & Lanphere; Zdanowicz et al. 1999) against the Klamath Crater Lake narratives (Clark 1953; Deur), the Budj Bim/Tower Hill argon dates (Matchan et al. 2020, Geology) against Gunditjmara tradition, and the case compilations in Piccardi & Masse (eds.) "Myth and Geology" and Vitaliano's geomythology - each row an implied span versus its independent date.

Novelty / leakage triage

anticipated in the literature — this exact test has never been run

The building blocks are all published - the Nunn-Reid drowned coasts, the Klamath/Crater Lake-to-Mount Mazama link (Mazama tephrochronology; Bacon & Lanphere 2006), and the Budj Bim/Tower Hill argon dates (Matchan et al. 2020, Geology) - and the 'shelf-life of oral memory' question is discussed qualitatively (Nunn; Piccardi & Masse eds., Myth and Geology, 2007). But no located work assembles a >=15-case census of independently date-anchored geomyths and computes the distribution of implied transmission spans with the conjecture's arithmetic (>=90% under 13 ka; above-ceiling cases <=2 and each single-anchored). The individual anchors exist; the census distribution is unbuilt.

• L. Piccardi & W.B. Masse (eds.), Myth and Geology (Geological Society of London, Special Publication 273, 2007)
• E.L. Matchan, D. Phillips, F. Jourdan & K. Oostingh, 'Early human occupation of southeastern Australia: new insights from 40Ar/39Ar dating of young volcanoes,' Geology 48 (2020)
• P.D. Nunn & N.J. Reid, 'Aboriginal Memories of Inundation...,' Australian Geographer 47(1) (2016); and C.R. Bacon & M.A. Lanphere, 'Eruptive history and geochronology of Mount Mazama and the Crater Lake region, Oregon,' Geological Society of America Bulletin 118 (2006)

Predictions

No prediction registered yet.