Carved from the Living Rock: The Sunken Churches of Lalibela

The cross on Saint George’s roof lay below Francisco Álvares. Visiting Lalibela in the 1520s, the Portuguese priest found the church locked and walked around it instead. He recorded its cross-shaped plan and another cross carved into the roof. The surrounding ground stood higher than the building.^[1] The church had been revealed by taking rock away, leaving the roof, walls and interior within what remained.^[5]

That reversal gives Lalibela its first surprise. A building usually rises above its foundations; here, workers cut down into the hill. But the question of how to make such a space leads to another: when did each space become a church? Archaeological studies suggest a complex reshaped through several phases, while religious tradition associates its creation with King Lalibela.^[9] The rooms also have another history, continuing through the people who pray, teach and care for them.

A locked door left an unusually close account

Álvares paid attention to things he could examine from outside: the proportions of the cross-shaped church, its enclosure, tombs and a basin of water. He also wrote down what people told him, including a belief that water taken from the basin could treat intermittent fevers.^[1] The account preserves that belief; it does not establish a medical effect. Observations and reported explanations sit beside one another in his narrative.

The church is now generally called Bete Giyorgis. An entrance trench approaches the court around it, where the carved body stands apart from the surrounding rock.^[3] Its outline is comparatively easy to follow because it is isolated. Other churches at Lalibela occupy more closely connected spaces, with passages, courts and adjoining rooms complicating the relationship between one excavation and the next.

The doors, windows and carved details give Saint George’s the appearance of a carefully designed building, even though its material was never assembled into walls. Some exterior openings are false windows.^[6] The carvers could give the surface a familiar architectural vocabulary without needing every feature to perform the job it would have in a structure built from separate pieces.

Álvares eventually confessed his fear that readers would not believe the account. He insisted that what he had recorded was true and that he had omitted much more.^[1] The disbelief he anticipated helps explain the precision of his description. He wanted the reader to picture an arrangement of solid ground and carved space that an ordinary building comparison could not quite convey.

The king’s story answers a different question

Gebre Mesqel Lalibela, the Zagwe ruler associated with the churches, occupies both historical and sacred accounts.^[2] In the story Álvares heard, the king commissioned the work and it was completed in twenty-four years. The priest presented that timetable as information given to him, rather than a calculation he could verify.^[1] The attribution survived as part of the place’s religious identity.

In the tradition David Phillipson retells, angels took over when the human labourers stopped for the day. The work continued through the night while people rested.^[3] The story expresses the sacred scale of the achievement. It need not be turned into a literal estimate of how many people excavated the churches, or how much rock each worker removed in a day.

Historical investigation asks which spaces were cut first, how they were altered and what their original functions may have been. Tradition can connect a whole sacred landscape with a revered ruler; archaeological evidence can distinguish stages within it.^[9] Those activities do not have identical purposes. A building phase is a narrower claim than the meaning of the complex for the people who worship there.

The hill supplied both material and weakness

At Bete Gabriel-Rufael, a geological and architectural study led by Federico Sani examined the rock as part of a preservation problem. The team identified a softer scoriaceous basalt horizon between more massive basalt layers.^[8] The builders had worked within a volcanic sequence whose different materials offered different possibilities. The places where carving was feasible helped determine where a room could be made.

The same rock contains fractures. Some arose through geological processes; others were associated with excavation. The study described blocks that could behave independently, creating risks to the church’s stability.^[8] Removing material changes what the remaining rock must support. An excavation can free a wall or a roof from the surrounding mass while leaving it vulnerable in a new way.

The team installed gauges to monitor fractures and used modelling methods associated with underground excavation.^[8] The work treated the church as both architecture and rock mass. A crack cannot be understood only as a line spoiling a carved surface. It may indicate movement within the material that keeps the space open.

Water adds a second problem. Drainage channels are part of the church landscape, and their condition affects the movement of rain around the monuments.^[16] A court cut into the ground has to let water escape as well as let people enter. Keeping the church accessible therefore includes attention to passages and drains whose importance may be less immediately obvious than the façade.

Two northern churches turn the interior into the event

Bete Medhane Alem has five aisles, a broad interior cut within one body of rock. Its scale has led to its description as the largest monolithic church, although that superlative is commonly qualified as a belief.^[10] The aisles make the volume easier to grasp. The maker’s task included leaving material in the right places to divide and support the space.

At Bete Maryam, carved and painted decoration gives another character to the interior. Phillipson places it among the highly developed examples of Ethiopian rock-hewn church architecture.^[9] Its walls and forms belong to a building tradition as well as to an excavation technique. Cutting stone did not remove the desire for an organised, decorated place of worship.

Nearby, Golgotha carries a different kind of significance. It is associated with King Lalibela’s burial, but Phillipson cautions that the history of that attribution is uncertain.^[3] Pilgrimage and historical dating meet around the same place without supplying the same answer. Reverence can be well established even where an archaeologist cannot confirm the origin of a tomb identification.

The names of these churches also describe relationships within Christian worship: the Saviour, Mary and Golgotha.^[4] They organise the complex through dedication and sacred association. The passages connect rooms made in different phases into a landscape that believers can use together. An archaeological sequence and a devotional route can run through the same set of courts.

An older room can acquire a sanctuary

Phillipson’s argument for several phases depends on relationships between the excavations. A passage can be interrupted; a new chamber can join an older room. In his account, two early structures with defensive functions were later converted for worship by adding sanctuary spaces.^[9] The evidence concerns change within the rock, not simply a stylistic resemblance to an earlier building.

David Keys reported these proposals in 2004, bringing attention to the possibility that some of the structures substantially predated their use as churches.^[11] The report is part of the history of the debate, not an independent date stamped onto the stone. More secure than a single neat timetable is the recognition that different parts of the site preserve evidence of alteration.

A 2010 study led by François-Xavier Fauvelle-Aymar set out the difficulty of sequencing rock-cut architecture. Excavation removes some of the layers archaeologists usually rely on. The team looked instead at cuts, openings, spoil and the traces of sediments to establish relationships between phases.^[12] What is missing can matter, provided there is evidence for how it was removed.

Later research led by Marie-Laure Derat places the churches within an earlier local history of living and working in caves, especially at Washa Mika’el. The study argues for continuity and transformation through Christianisation, rather than an entirely new tradition appearing without predecessors.^[13] The ground had a human history before every surviving chamber acquired its present sacred function.

The result leaves room for disagreement about absolute dates and the length of particular phases. Phillipson associates the final major development with King Lalibela while arguing that the wider process lasted much longer.^[9] That is an interpretation grounded in archaeological relationships. It should remain distinct from both a universal scholarly consensus and the single campaign described in royal tradition.

Jerusalem’s names become a route in Ethiopia

The watercourse called Yordanos, or Jordan, gives the complex a connection with the Holy Land. Other names include Golgotha and the Tomb of Adam. Phillipson relates this sacred geography to the role attributed to King Lalibela.^[9] The connection is expressed through places people can approach and pass between. A distant landscape becomes present through naming and movement.

Samuel Walker’s account reconstructs pilgrimage relationships among these places, including holy-water sites and passages between churches.^[7] Such a reconstruction is an interpretation of how the landscape acquired significance. It brings attention to the traveller within the complex, rather than treating each church as a detached sculptural object.

The Jerusalem association cannot by itself date every cut. Names can be attached to older spaces as their meanings change.^[9] Nor does it reduce Ethiopian Christianity to a copy of something elsewhere. The rooms, rock-working methods and continuing custodians belong to a local history through which those wider Christian connections took shape.

Stone beams remember buildings made another way

At Bete Gabriel, Phillipson identifies a ceiling recut to represent capitals and beams. The room had already existed before that decorative architectural change.^[9] The carvers made the roof appear to contain elements that an assembled building would need for support. Here the apparent beam was part of the same rock as its surroundings.

Bete Amanuel also carries features associated with earlier Ethiopian building forms. Framed openings and imitations of timber architecture connect rock carving to a wider inherited vocabulary.^[11] A technique can change while a form remains recognisable. The stone does not need to behave like wood for a doorway or beam-like feature to recall a familiar kind of building.

Those continuities are one reason the site cannot be explained by amazement alone. The makers were choosing forms, organising rooms and adapting earlier spaces. Derat and colleagues place related transformations within the region’s pre-existing cave traditions.^[13] The achievement becomes more specific when it is attached to people with inherited skills, rather than left as an inexplicable miracle of engineering.

A church’s centre is not its outer shape

Ethiopian Orthodox worship gives the church an interior focus that architecture alone cannot explain. The Church’s account of worship places the consecrated tabot at the centre of its sanctity; services depend on its presence.^[17] A cross-shaped roof or a large monolithic body is therefore only part of what makes the place a church for its community.

The site remains home to priests and monks and receives pilgrims for the great feasts.^[16] Genna, the Nativity celebration, belongs to Lalibela’s pilgrimage calendar, and Walker describes observances associated with Bete Maryam’s court.^[7] The open spaces between the churches have roles alongside the interiors. A court can be both a product of removing rock and a place where people gather for worship.

Timkat, Ethiopian Epiphany, makes the tabot’s role public through procession. UNESCO’s account describes clergy carrying the tabot, prayers and hymn services through the night, and its return to the church.^[18] The ritual depends on people knowing what to carry, sing and read. The traditions are maintained through practice, not contained in stone alone.

Teaching is therefore part of conservation too. A UNESCO assistance project approved in 2023 aimed to revive Lalibela’s church schools after disruption linked to conflict and the pandemic. Its stated purpose included sustaining religious practice and the knowledge held by church communities.^[19] The project records an intention and a programme of support; it does not by itself establish that every planned recovery has been completed.

A laser scan records the cut but cannot keep the rain out

The Zamani Project documented Lalibela through field campaigns between 2005 and 2009. Its laser scans produced models, plans and sections of the churches.^[10] The record makes relationships between surfaces available for study in another form. A carved junction can be measured and compared without every researcher standing in the same narrow passage.

The physical churches still need protection. Rain shelters introduced another set of structures over some of them, intended to reduce water entering the rock. Their impact and continued use have been contested.^[14] Protection has to answer more than the question of whether an object stays dry. Supports, roof forms and the setting also affect how the church can be approached, seen and used.

In 2017, Priest Solomon Tolcha called attention to risks associated with the shelters and the need for action. The report records discussion of their removal and the resources required.^[15] It should be read as a dated intervention by a church official, rather than proof that the proposed work subsequently happened. His concern placed local religious responsibility within the conservation debate.

A 2022 report discussed a later proposal for a canopy with a bamboo-like appearance.^[14] A proposal is another stage in that debate. The geological study’s fracture gauges, the scanned models and the shelter designs each address a different problem. None removes the need for decisions by the custodians, specialists and public bodies responsible for the living churches.

The route leads back into use

Lalibela’s church schools preserve forms of learning needed for religious life, while the rooms and passages supply places for that life to continue.^[19] Maintaining a wall and maintaining a practice require different work, but they meet at the same doorway. Protection that kept the structure intact while losing its custodians would preserve only part of the place.

At Timkat, the tabot’s procession returns to the church.^[18] That return gives the carved room a purpose beyond the moment when it was made. Álvares recorded a door he could not enter and a cross below the rim of the ground.^[1] For the people responsible for the church, the door must also open for the next service.