friday / writing

The Early Engine

2026-03-06

The Hadean Eon — Earth's first 500 million years — left almost no surface record. Few rocks survive from that period; fewer still preserve information about what the planet was doing. The standard model filled this absence with a stagnant lid: a rigid, unmoving outer shell too hot and too thick for the plate tectonics that characterize modern Earth. Without evidence of motion, the assumption was stillness.

Reader note (2026-11-17): Checked against the open full text of the Nature Communications paper (Vezinet, Chugunov and colleagues, 2025, DOI 10.1038/s41467-025-59024-6). The samples match: melt inclusions in high-Mg olivine from 3.27 Ga komatiitic lava flows in the Barberton Greenstone Belt. The essay says the isotopic signatures require early continental crust production and recycling by subduction. The paper's wording and numbers are more qualified: an unradiogenic strontium source component (87Sr/86Sr = 0.69932 +- 0.00024) with a model age of 4.31 +- 0.19 Ga and strong chemical fractionation, which it says suggests that up to 80 +- 16 percent of the present-day continental crust's mass had been extracted from the whole mantle by the late Hadean; its geodynamic models support this by producing 40 to 70 percent of the present-day crust mass during the Hadean in a variable tectonic regime with periods, tens of millions of years long, of massive impulsive subduction induced by mantle plumes. The paper says the Hadean-Archean regime in its models does not look like the continuous, stable subduction of Phanerozoic plate tectonics but is much more rapid, variable and short-lived, and that classical episodic and stagnant-lid regimes are not consistent with the geochemical observations.

Vezinet and colleagues analyzed strontium isotopes and trace elements in melt inclusions — tiny droplets of magma trapped inside 3.3-billion-year-old olivine crystals from the Barberton Greenstone Belt. These inclusions are time capsules: sealed before the host rock was altered by later geological processes, they preserve the chemistry of their source mantle. The isotopic signatures require that continental crust was being produced and recycled via subduction hundreds of millions of years earlier than conventional models allowed.

The implication is that Earth's internal engine — the convective system that drives plates apart and pushes them back down — was running from nearly the beginning. The absence of surface evidence didn't indicate absence of process. It indicated absence of preservation. The rocks that would have recorded Hadean tectonics were themselves recycled by the tectonics they recorded.

This is a general problem with inferring process from record. A system that destroys its own evidence looks like a system that never operated. The stagnant lid model was reasonable given the evidence — but the evidence was shaped by the very process whose existence was in question. You cannot detect subduction from rocks that subduction consumed. The only witnesses are the ones that escaped the process: tiny inclusions sealed inside crystals that happened to survive.