Ancient Engineering · Lalibela, Amhara Region, Ethiopia

Rock-Hewn Churches of Lalibela

Eleven churches carved downward into living rock — a New Jerusalem that tradition says was finished at night by angels.

Mainstream: c. 1181–1221 AD (reign of Gebre Mesqel Lalibela), with earlier phases possibly from the 8th–11th centuriesAlternative: Built in a single generation with angelic help (Ethiopian tradition), or with 12th-century Crusader/Templar involvement (Hancock and others)12.03°, 39.04°

Engineering Belief Myth

At a glance

Rock-Hewn Churches of Lalibela — Photo: Bernard Gagnon · CC BY-SA 3.0

At roughly 2,500 metres altitude in the Ethiopian highlands, the town of Lalibela contains eleven medieval churches that were not built but excavated — cut downward into the red volcanic scoria until entire buildings, some free-standing on all four sides, emerged from the living rock. The most famous, Bete Giyorgis (the Church of St George), sits at the bottom of a 12-metre pit in the shape of a perfect Greek cross, connected to the surface world by a sunken trench. Bete Medhane Alem, in the northern cluster, is generally reckoned the largest monolithic church on Earth. The churches are linked by a warren of trenches, tunnels and drainage channels, and the site is laid out with deliberately Biblical geography: a channel called the River Jordan divides the two main clusters, and features carry names such as the Tomb of Adam and Golgotha. Unlike most great archaeological sites, Lalibela was never abandoned — it remains one of the holiest places of the Ethiopian Orthodox Tewahedo Church, thronged by tens of thousands of white-robed pilgrims every Ethiopian Christmas (Genna) in January.

See it on the globe →

The mainstream view

What archaeology says

Ethiopian royal tradition and hagiography attribute the complex to King Gebre Mesqel Lalibela of the Zagwe dynasty (r. c. 1181–1221), who is said to have created a substitute Jerusalem for pilgrims after Saladin's capture of the real one in 1187. Historians accept a strong Zagwe connection — land grants and church documents tie the king to the site — but modern archaeology has complicated the picture of a single heroic building campaign. David Phillipson of Cambridge proposed that several structures, notably Bete Merkorios and Bete Gabriel-Rufael, began life centuries earlier — perhaps in the 7th to 8th centuries AD — as fortifications or elite residences of the late Aksumite period, and were only later converted into churches, with the whole complex developing in four or five phases.

Since 2009, and intensively through the 2019–2025 'Sustainable Lalibela' programme, a French-Ethiopian mission led by Marie-Laure Derat and Claire Bosc-Tiessé (CNRS, with the French Centre for Ethiopian Studies and Ethiopia's heritage authority ARCCH) has excavated around the churches for the first time. Their work has revealed that a powerful elite occupied and engineered the hill by around the 10th century, before the traditional construction date, and that the churches were repeatedly recut, enlarged and remodelled over centuries — a long, layered history rather than a single reign's miracle. The team also documented the enormous spoil heaps of quarried rock, a straightforwardly human signature of the work.

None of this diminishes the achievement: carving basilicas downward means every mistake is permanent, and the masons had to plan windows, arches, vaults and drainage before the first cut. The consensus view is that Lalibela is the culmination of an indigenous Ethiopian rock-cutting tradition, visible earlier in the hundreds of rock-hewn churches of Tigray, not an import.

Key evidence cited

Land grants and church documents linking the Zagwe king Gebre Mesqel Lalibela to the site
Phillipson's stratigraphic sequencing showing four to five carving phases, the earliest possibly late Aksumite (7th–8th century)
French-Ethiopian mission evidence (Derat and Bosc-Tiessé) of elite occupation and engineering around the 10th century
Aksumite architectural vocabulary — blind windows, monkey-head beams — copied from earlier Ethiopian buildings
A regional tradition of hundreds of rock-hewn churches in Tigray demonstrating the local skill base

The alternative view

What the skeptics propose

The oldest alternative account is the Ethiopian Church's own: the Gadla Lalibela (Life of Lalibela) records that the king was taken up to heaven, shown a city of rock-cut churches, and commanded to reproduce it — and that when the human workforce downed tools each evening, angels continued the work through the night, accomplishing in roughly 24 years what should have taken lifetimes. For the priests of Lalibela this is not folklore but history, and they point out that no graves, no masons' names and no construction records were ever found for so vast a project — though the French mission's spoil-heap studies now push back on the idea that the work left no trace.

A second strand looks to medieval Europe. The Portuguese priest Francisco Álvares, the first European to describe Lalibela in the 1520s, wrote that he feared no one would believe his account; later writers speculated that white craftsmen must have been involved. Graham Hancock, in The Sign and the Seal (1992), argued that Knights Templar reached Ethiopia in the late 12th century — following, in his telling, the trail of the Ark of the Covenant — and lent their engineering skill to King Lalibela's project, noting the coincidence of dates with the Templars' heyday and the crusader-era loss of Jerusalem.

Historians reply that there is no documentary, architectural or archaeological evidence of any Templar presence in Ethiopia, that the churches' style is demonstrably Aksumite (blind windows, monkey-head beam ends, stepped plinths copied from ancient Ethiopian buildings, not European ones), and that the earlier-origins findings of Phillipson and the Derat–Bosc-Tiessé mission make foreign crusader help even less necessary: Ethiopians had been cutting monumental spaces from rock for centuries before any Templar existed. The angel tradition, meanwhile, is read by scholars as a theological statement of the site's holiness — and perhaps a memory of just how impossibly fast the final phase seemed to those who inherited it.

Key evidence cited

The Gadla Lalibela's account of angels completing the churches by night, treated as history by the Ethiopian Church
The absence of any contemporary construction records, masons' names or builders' graves
Hancock's argument that the traditional dates coincide with the Templar era and the 1187 loss of Jerusalem
Álvares's 16th-century testimony that the churches seemed beyond belief, fuelling foreign-builder speculation
The sheer engineering audacity of planning entire basilicas downward with no possibility of correcting errors

Genuinely open questions

How much of the complex existed before King Lalibela — and were the earliest cuttings churches, palaces or fortifications?
How was so vast an excavation organised and provisioned in the highlands, and where did the specialist masons train?
Can the phases identified archaeologically be tied to specific reigns, given the near-total absence of written construction records?

Worth knowing

Bete Giyorgis was reputedly the last church carved — tradition says St George appeared on horseback, furious that none of the churches honoured him, and the hoofprints of his horse are still shown to pilgrims in the rock of the access trench.