Rocks Are Earth's Receipts

Two principles do most of the heavy lifting. The first is superposition: in undisturbed layers, the stuff underneath got there first. The second is radiometric decay: unstable isotopes convert to stable daughter products at fixed rates, giving multiple independent clocks with different useful ranges — carbon-14 for tens of thousands of years, potassium-argon and uranium-lead for millions to billions.

Crucially, these methods cross-check each other. A volcanic ash layer dated by argon-argon can sit directly above a fossil bed, below a radiocarbon-dated hearth, inside a magnetically reversed interval that matches the global seafloor record. When four independent systems agree, the number is not a guess. When they disagree, that disagreement is itself data, and it is usually where the interesting paper comes from.

Dating a rock is not the same as dating an event. A radiocarbon date on a charcoal fragment tells you when that tree stopped taking up carbon — not when someone burned it, and definitely not when they built the wall next to it. Old wood, reused timber, contaminated samples and disturbed layers all inject error that has nothing to do with the physics.

Geologists handle this with error bars and stated assumptions. That habit is why the field is trustworthy, and also why press coverage of it is so often wrong: a date reported as "9,500 ± 100 years, calibrated" becomes "scientists say exactly 9,500 years" by the second retelling.

Also unresolved: the precise driving mechanics of mantle convection, the frequency distribution of large impacts over deep time, and whether several "extinction events" are single catastrophes or compressed sequences that our dating resolution simply cannot separate. At 66 million years out, a ten-thousand-year gap can look instantaneous.