The Accidental Discovery

In 1963, a man in the Cappadocian town of Derinkuyu knocked down a wall in his basement during renovations. Behind it, he found a room. Behind that room, another. And behind that, a tunnel sloping downward into darkness. What he had stumbled into was the entrance to one of the largest underground cities ever constructed — a structure reaching roughly 85 meters beneath the earth’s surface, spanning eighteen levels, and capable of sheltering an estimated 20,000 people along with their livestock and food stores.

The story sounds almost too cinematic to be true. But the archaeological record is unambiguous: somebody, at some point in antiquity, carved an entire city out of volcanic tuff — complete with ventilation shafts, water wells, churches, stables, wine presses, and a defense system so sophisticated that a small group of people could seal off the entire complex from the inside using half-ton rolling stone doors. The question that gnaws at engineers and archaeologists alike isn’t whether it was built. It’s how.

The Rock That Made It Possible

You cannot understand Derinkuyu without understanding Cappadocian geology. The region sits on thick deposits of ignite tuff — compacted volcanic ash ejected by eruptions from Mount Erciyes and other volcanoes millions of years ago. This material has a peculiar and enormously useful property: when first exposed, it is soft enough to carve with hand tools, almost like cutting through dense cheese. But once exposed to air, it undergoes a chemical hardening process, oxidizing and losing moisture until it becomes remarkably rigid and load-bearing.

This is not a minor detail. It is the entire premise of the engineering. Without tuff, Derinkuyu could not exist. Limestone would have required vastly more labor. Granite would have been effectively impossible with Bronze or Iron Age tools. The geological accident of Cappadocia — a landscape shaped by volcanic eruptions, erosion, and time — handed its inhabitants a building material that was both forgiving to work and strong enough to support multi-story excavation without structural failure.

Ömür Harmanşah, an archaeologist at the University of Illinois at Chicago who has worked extensively on Anatolian subterranean spaces, has noted that the tuff’s properties created a kind of “natural invitation” to dig. The Cappadocians didn’t invent underground architecture out of abstract genius. They responded to what the rock offered them.

Breathing Underground: The Ventilation Problem

Here is where the engineering becomes genuinely impressive. Dig a hole in the ground and go deep enough, and you will suffocate. Carbon dioxide accumulates. Oxygen depletes. Temperature regulation fails. For a shelter housing thousands, this isn’t a theoretical concern — it’s the primary engineering constraint.

Derinkuyu’s builders solved it with a network of approximately 52 ventilation shafts, some of which extend from the surface all the way down to the lowest levels. The deepest known shaft runs roughly 55 meters. These shafts served a dual purpose: they moved air, and several of them doubled as wells, tapping into the underground water table.

The principle at work is stack ventilation — sometimes called the chimney effect. Warm air rises. If you create vertical channels connecting lower and upper levels, temperature differentials between the cool underground and the warmer surface generate natural convection currents that pull fresh air downward through some passages while stale air exits through others. The builders didn’t need to understand fluid dynamics in any formal sense. They needed to observe that certain configurations of tunnels and shafts produced airflow, and then replicate what worked.

But observation alone doesn’t explain the precision. Roberto Germanò, an Italian engineer who conducted structural analyses of Cappadocian underground sites in the early 2000s, pointed out that the shafts are not randomly placed. They are distributed to create circulation zones, ensuring that even the deepest inhabited levels receive adequate air exchange. The spacing suggests either empirical trial-and-error refined over generations, or a more deliberate understanding of airflow mechanics than we typically credit to ancient builders.

There’s also an underappreciated cleverness in how the ventilation shafts interact with the city’s layout. Corridors are narrow — rarely wider than a person’s shoulders. Rooms branch off at angles. This isn’t just defensive design (though it serves that purpose). Narrow passages with bends accelerate airflow through the Venturi effect, pulling air through the complex more efficiently than wide, straight tunnels would. Whether this was intentional or a happy byproduct of defensive architecture is genuinely unknown. Both explanations are plausible. Neither is proven.

Water: The Non-Negotiable Resource

A city that cannot breathe is a tomb. A city without water is a slower tomb. Derinkuyu’s water system is arguably more impressive than its ventilation, because it had to solve a harder problem: contamination.

Several of the deepest ventilation shafts also served as wells, reaching down to the water table. But crucially, not all of them did — and the wells that supplied drinking water could be controlled independently from the surface. During a siege, inhabitants could block access to their water supply from above while still drawing from it below. Some shafts that appeared to be wells from the surface were reportedly decoys or were deliberately kept separate from the drinking water system, preventing an invader from poisoning the supply.

Andrea De Pascale, who surveyed Cappadocian underground sites for the Italian geological survey, documented evidence that water channels within the complex were carved to direct flow away from habitation areas and food storage, reducing the risk of contamination from waste or standing water. Drainage channels in the floors of several levels appear designed to move water toward collection points or away from the complex entirely.

This is sanitary engineering, and it was practiced at Derinkuyu centuries — possibly millennia — before germ theory. The builders didn’t know about waterborne pathogens. But they understood, probably from hard experience, that mixing waste water with drinking water killed people. The architectural separation between water systems, waste areas, and living quarters reflects that knowledge encoded in stone.

Defense in Depth — Literally

Whoever designed Derinkuyu expected to be attacked. The entire structure reads as a military engineer’s paranoid fantasy, and that’s a compliment.

The entrances — and there are many, scattered across the town above — are narrow enough that only one person can pass through at a time. Corridors between levels are similarly constricted, often requiring a person to crouch or turn sideways. An invading force couldn’t bring numbers to bear. Ten thousand soldiers were no more effective than ten at a chokepoint seventy centimeters wide.

Then there are the rolling stone doors. These are massive circular stones, roughly 1 to 1.5 meters in diameter and weighing between 200 and 500 kilograms, carved from the same tuff as the surrounding rock. They were designed to be rolled into place from inside a room or corridor, blocking passage completely. A hole in the center of each stone — too small for a person to crawl through — allowed defenders to thrust spears at anyone attempting to break through. The stones could not be moved from the outside. The geometry was deliberate: they sat in carved grooves that locked them into position when closed, and the corridor floors were shaped to prevent them from being pushed or pulled from the invader’s side.

Each level could be sealed independently. This is what military engineers call defense in depth. If attackers breached the first level, the defenders simply retreated to the second and sealed it. Then the third. Then the fourth. Each retreat cost the attackers time, casualties, and morale while the defenders remained supplied with air, water, and food. An attacking army would find itself in a pitch-dark, oxygen-poor maze full of dead ends, false corridors, and stone doors that couldn’t be forced open — while the people they were trying to reach sat comfortably eighteen levels below, eating stored grain and drinking clean water.

There is also a tunnel, partially explored, that appears to connect Derinkuyu to another underground city at Kaymaklı, approximately 9 kilometers away. If this connection was fully functional — and its full extent hasn’t been conclusively mapped — it would have allowed the population to evacuate from one city to another entirely underground. The strategic implications of that are staggering.

Who Built It, and When?

This is where confidence gives way to educated guessing. The honest answer is: we don’t know for certain.

The most commonly cited attribution is to the Phrygians, an Iron Age people who inhabited central Anatolia from roughly the 8th to 7th centuries BCE. The Turkish Department of Culture officially associates the earliest construction with the Phrygians, and several archaeologists, including the late Turkish historian Semavi Eyice, supported this dating based on architectural similarities to other Phrygian rock-cut structures.

But the site was clearly expanded, modified, and used by successive civilizations over many centuries. Early Christians used the lower levels as refuges during Roman persecution — a cross-shaped church carved into the rock on one of the lower floors dates to approximately the 5th to 10th centuries CE. The Byzantines expanded the complex significantly, likely adding many of the defensive features and deepening the excavation. Arab-Byzantine wars between the 7th and 10th centuries would have provided ample motivation for a civilian population to maintain and improve an underground refuge.

Some scholars have pushed the origins back even further. A minority view, advanced by researchers including Aner Keinan-Schoonbaert of University College London’s Institute of Archaeology, suggests that earlier Hittite populations may have begun the initial excavation, placing the oldest layers in the second millennium BCE. The evidence for this is circumstantial — Hittite texts reference underground storage and refuge structures in Anatolia, and Hittite settlement patterns overlap with Cappadocian underground sites — but no Hittite artifacts have been definitively recovered from Derinkuyu’s lowest levels.

What seems most likely is that Derinkuyu wasn’t “built” in any singular sense. It accreted. One generation dug a few rooms for storage or shelter. The next expanded them. Over centuries, what started as root cellars and grain stores became a city. This is how most monumental ancient structures actually came into being — not through a grand plan executed all at once, but through incremental effort compounded over time. The Colosseum had an architect. Derinkuyu probably didn’t. It had hundreds of them, working across generations, each responding to the needs and threats of their own era.

What We Still Don’t Know

Only about half of Derinkuyu has been excavated and opened to the public. The lower levels remain partially explored. The tunnel to Kaymaklı has never been fully mapped. We have no written records from the builders describing their methods, their organizational structure, or how they managed the logistics of removing thousands of tons of excavated rock — a question that, honestly, doesn’t get asked enough. Where did all the stone go? Tuff is useful as building material, so presumably it was carried to the surface and used for above-ground construction, but the sheer volume involved implies a sustained, organized labor effort over a very long period.

We also don’t know how many people actually used the complex at its peak, or for how long at a stretch. Estimates of 20,000 are based on spatial analysis — total habitable volume divided by minimum space per person — but these are theoretical maximums, not census data. Living underground for weeks or months with thousands of others, even with ventilation and water, would have been psychologically brutal. The archaeological record tells us what the builders could do. It tells us almost nothing about what that experience was like.

The Question That Remains

Modern engineers, armed with computational fluid dynamics, ground-penetrating radar, and centuries of accumulated knowledge about structural loads, would approach a project like Derinkuyu with computer models and stress analyses before cutting a single stone. The people who actually built it had none of that. They had hand tools, oil lamps, and the accumulated knowledge of previous generations who had dug into the same rock and learned, sometimes fatally, what worked and what didn’t.

So here’s what I keep coming back to: we tend to treat ancient engineering achievements as puzzles to be solved — as if the real question is whether people in the past were capable of building such things without modern technology. But Derinkuyu doesn’t ask whether they could. It’s standing proof that they did. The better question, and the harder one, is what we’ve lost. Not technologically — we can tunnel through mountains with boring machines now. What I mean is the institutional knowledge, the generational patience, the willingness to undertake a project whose completion you would never see. Derinkuyu took centuries to build. What are we building today that we expect our great-great-grandchildren to finish?