Cross-domain evidence that oral traditions encode measurable physical information
The Emergent Precision program tests a specific hypothesis: that oral traditions preserve accurate information about the physical world, and that this accuracy correlates with the observability of the phenomena being described. Traditions about highly observable events (flood directions, coastal changes) should be more accurate than traditions about less observable phenomena (invisible diseases, subsurface geology). This prediction — the observability gradient — is confirmed across six independent knowledge domains.
In the original 6-domain Emergent Precision analysis, accuracy correlates with observability at r = 0.899 (p = 0.015, blind-scored). This was predicted before testing and confirmed across independently collected datasets. The expanded 41-domain dataset sustains the gradient at r = 0.527 (p = 0.0004), with blind-scored subset r = 0.893 (n=7, p = 0.007). See the gradient page.
11/11 correct, Cohen’s h = 2.498
Significant accuracy
Match archaeological timelines
Correct identification of volcanic origins
Weak positive signal
Consistent with chance (h = −0.080)
In 2026, pre-registered predictions about submerged stepping stones in Torres Strait were tested against 30-meter resolution bathymetric data. The SSE bearing from Mabuiag Island — described in a tradition recorded by Haddon in 1904 — resolved as the shallowest corridor of 8 tested bearings, with 6 stepping-stone features confirmed along the predicted azimuth. These features have been submerged for approximately 8,000 years — 320 generations of oral transmission.
29 pre-registered predictions · 14 confirmed · 2 falsified · 3 resolution upgrades · 0 downgrades
Standard cultural transmission models predict information decay to noise within 500–1,000 years. The flood traditions retain 82% accuracy after 7,000+ years. The gap between predicted and observed retention is four orders of magnitude. The observability gradient provides the mechanism: environmental feedback acts as an error-correction system, suppressing transmission noise for observable phenomena while allowing drift for unobservable ones.
See the papers page for current status across the full DTRI portfolio.