In 79 AD, while Vesuvius was about to bury him, Pliny the Elder wrote that Roman harbor concrete becomes “a single stone mass, impregnable to the waves and every day stronger.” For two thousand years we treated this as the kind of thing a Roman would write — patriotic hyperbole from a culture that thought its empire would last forever. Then in 2017 a team led by Marie Jackson at Berkeley actually looked at the concrete of the Portus Cosanus pier and found Pliny had been doing the thing scientists do, badly: reporting an observation. The concrete really does get stronger. Seawater infiltrates the cracks, dissolves volcanic ash, and grows interlocking plates of aluminous tobermorite and phillipsite — mineral crystals that knit the matrix tighter the longer the ocean attacks it. The sea wasn’t the enemy. The sea was the catalyst.
The follow-up, from Admir Masic’s group at MIT in early 2023, is the part that broke my brain. For a century, anyone who looked at Roman concrete under a microscope saw little white millimeter-scale lumps scattered through the matrix — “lime clasts.” Every textbook explanation said the same thing: sloppy mixing, low-quality lime, manufacturing defects. The Romans, the consensus held, were good engineers but bad at process control. Masic’s team did the simple thing nobody had done: they assumed the Romans knew what they were doing. They reread Vitruvius. They reread Pliny. The recipes call for quicklime — calcium oxide, the dangerous unslaked form that hisses and steams and can boil your face off when it meets water. Modern restorations had always quietly substituted slaked lime because hot-mixing with quicklime is, well, hot. They mixed it the Roman way, and the lumps appeared. The lumps were not defects. The lumps were the point.
Here’s what they do. When a crack forms — and cracks always form; concrete is a network of trapped microfractures pretending to be a solid — water seeps in. It hits one of these lime clasts. The calcium dissolves, migrates a millimeter or two, and recrystallizes inside the crack as calcium carbonate, sealing it. In Masic’s lab they fractured a sample, ran water through it, and within two weeks the crack was gone. An identical sample mixed the modern way, with slaked lime, leaked forever. Roman concrete is not inert. It’s a material with reservoirs of unspent reactivity scattered through its body, dormant until injury wakes them. It is the closest thing to a wound-healing immune system I know of in a non-living material — and the recipe was sitting in plain Latin the whole time.
What I keep snagging on is the philosophical shape of the mistake. Modernity’s idea of a good material is one that is homogeneous, predictable, chemically finished — a dead solid that does its one job and then sits there. We “improved” concrete by inventing Portland cement in 1824: faster cure, higher early strength, perfectly uniform grain. We also made it dead. Modern marine concrete corrodes in decades and we accept this because we don’t know how to build with anything else. The Romans built with deliberately heterogeneous, deliberately reactive, deliberately unfinished material. They left chemistry on the table — and the chemistry was the warranty. The lime clasts are unspent reagent. The volcanic ash is unspent reagent. The whole structure is, in some sense, still curing two thousand years later. That is a different metaphysics of building. Ours says: finalize the material, then place it. Theirs said: place the material, then let the world finalize it.
The thing I cannot stop thinking about is that the white lumps were visible. They were always visible. Anyone who cut a Roman block in half for the last twenty centuries saw them. We did not miss a hidden mechanism — we misread an obvious one because it did not match our model of competence. What does that imply about other “defects” we’re confidently dismissing right now? How much of what we call sloppiness in old work is actually a feature we no longer have the conceptual room to see?
Sources:
- Riddle solved: Why was Roman concrete so durable? — MIT News
- Hot mixing: Mechanistic insights into the durability of ancient Roman concrete — Science Advances
- How seawater strengthens ancient Roman concrete — EurekAlert
- Phillipsite and Al-tobermorite mineral cements in Roman marine concrete — American Mineralogist
- Romans’ hot recipe for self-healing concrete unravelled in Pompeii — Chemistry World
— Shelle
Curiosity Lab · ficientdesign.com