Churches as compass fossils

Claim (verbatim)

Churches as compass fossils. This joins medieval church archaeology to geomagnetism. Churches were meant to face east, yet surveyed orientations scatter by several degrees, and the scatter is usually written off as sloppy surveying. But once the magnetic compass entered the building trades — plausibly around 1200 — masons who set out east magnetically would err by exactly the local geomagnetic declination, which wanders over the centuries and is reconstructed by archaeomagnetism. The conjecture predicts that post-1200 church orientation errors trace the geomagnetic declination curve: buildings of the same date and region should share a systematic offset equal to the declination of their decade. The payoff runs both ways — the date at which orientation errors lock onto the curve dates compass adoption in the building trades, while thousands of dated churches become new data points refining the archaeomagnetic record itself.

Prediction clause (verbatim)

For each dated church in orientation databases vs paleo-declination, compute the orientation error from true east and pair it with the reconstructed geomagnetic declination at that place and construction date; regress orientation error on paleo-declination for post-1200 cohorts, with pre-1200 churches as the control. Primary clause: the post-1200 regression slope must be positive, significant, and within the range 0.5 to 1.5 — consistent with magnetic setting-out — while pre-1200 errors show no such relation; absence of a declination signal in post-1200 orientations falsifies the claim. The verdict follows the primary clause.

Kill-dataset (verbatim)

orientation databases vs paleo-declination.

Novelty / leakage triage

already answered in the literature

CORRECTED 2026-07-04 after independent verification: the original rationale attached an uncited '570 Danish Romanesque churches, under 5% compass-consistent' statistic that conflated two different studies and pointed the wrong direction for the Danish case. Corrected picture: the compass-orientation hypothesis is explicitly PROPOSED in the literature — Abrahamsen 1992 (Archaeometry 34(2), traced by the verification pass) argues a substantial minority (~25%) of 12th-century Danish churches were probably compass-oriented and uses declination to date them — and explicitly TESTED with negative results elsewhere: Arneitz et al. 2014 (GJI) find church orientations do not track the geomagnetic declination curve, and Ali & Cunich find little evidence of compass alignment in early English/Welsh churches. The connection (church orientation as a compass/declination fossil, usable for dating) is therefore thoroughly pre-existing with regionally CONTESTED results — a live scholarly question, not a new join. Verdict leaked stands; the earlier 'already largely refuted' framing was too strong and is withdrawn.

• 'Orientation of churches by magnetic compasses?', Geophysical Journal International 198(1) (2014) — Direct test; orientations do not track the declination curve
• Ali & Cunich, 'The Orientation of Churches: Some New Evidence' — English/Welsh orientation evidence; little compass-consistency
• Abrahamsen 1992, 'Evidence for church orientation by magnetic compass in twelfth-century Denmark', Archaeometry 34(2) — PRO side: ~25% of Danish Romanesque churches probably compass-oriented; added by the post-verification correction — this is the study the defective statistic garbled

Its literature citations feed the frontier as source leads (3 leads below the evidence/publication boundary, not yet reviewed).

Predictions

No prediction registered yet.