Ethiopia's 1,700-Year-Old Stone Towers

Published on October 19, 2026 at 2:53 PM

The 500-Ton Stone That Fell: Engineering the Aksumite Stelae

Did you know that a 1,700-year-old Ethiopian monument spent decades in Rome... and that it was carved to look like a multi-storey building?

The word usually attached to these monuments is "obelisk," and that label is the problem. The monuments at Aksum [AK-soom; older Western usage spells it Axum, AX-um], in the Tigray [TEE-gray] region of northern Ethiopia [ee-thee-OH-pee-ah], are called stelae [STEE-lee]. They are a different kind of object, with a different form and a different function.

The scale is the first claim worth stating. The largest of them, the Great Stela, was over 30 metres long and weighed roughly 500 tons or more. It is probably the largest single block that any ancient culture tried to erect. It did not stay up. It lies broken where it fell, which makes it useful to engineers, because failures show where the limits were.

This article reads the Aksumite [AK-soo-mite] stelae as engineering. Each section states one rule the monuments teach.

---

1. Rule One: A Structure Can Carry Its Meaning in Its Shape

What the stelae are

The stelae are monolithic stone monuments, carved and erected around the 3rd to 4th centuries AD. They appear to have marked elite or royal tombs, within a funerary landscape. UNESCO inscribed Aksum as a World Heritage Site in 1980.

The form is a building

The finest stelae are not plain pillars. They are carved to represent multi-storey structures, with false doors and windows. The Metropolitan Museum of Art describes them this way.

Think about what that means. The Aksumites took one block of stone and cut into its surface the look of a tall building. Whether that carving copies a particular timber-and-stone building technique, such as projecting beam-ends, has not been confirmed in the sources used here, so this article does not claim it.

The rule: a monument's shape can carry information as well as its inscriptions or its position. Here the shape states a building in stone, and the stone is a weather-resistant one.

Why the label "obelisk" misleads

"Obelisk" is a convenience. Aksumite stelae have a different form and function from Egyptian-style obelisks, and they are carved to resemble buildings. Their origin is debated. Rodolfo Fattovich [roh-DOL-foh fah-TOH-vich] explored Near Eastern and African megalithic traditions in a 1987 paper on the question. Calling them obelisks answers that debate before it starts.

---

2. Rule Two: Match the Material to the Timescale

The rock

The Met says the stelae were carved mainly from solid blocks of nepheline syenite, a weather-resistant rock that looks like granite. That detail matters. A monument meant to stand for centuries needs stone that does not degrade in the open air.

A label that needs checking

Sources disagree on the name. Many popular accounts say "granite." Some sources call the stone of the Obelisk of Axum "phonolite." The Met says nepheline syenite. This is not a trivial dispute. These are different rocks, and the label is a petrological statement. The petrology literature has not been checked for this article, so "a weather-resistant, granite-like rock" is the safe wording until it is.

The rule: durability and lifespan go together. A stone that has survived roughly 1,700 years in the open air is a durable one. Whether the Aksumites chose it for that reason is an inference, not a documented fact.

---

3. Head-to-Head: Three Stelae, Three Outcomes

Three of the best-known stelae make a natural comparison.

They are close in date and very different in result.

What the numbers show

Stela 1 is about a third longer than Stela 2 and roughly three times heavier. Stela 2 stands about 3 metres taller than Stela 3.

That gap between "a third longer" and "three times heavier" is the first engineering lesson of the table. As a body grows, its mass climbs faster than its length. The stelae are not identical in proportion, so treat this as an illustration, not a calculation. It is consistent with a very large stela being harder to raise, and Stela 1 is the only one of the three that lies fallen. It does not prove why it fell.

One caution on Stela 3: the name "King Ezana's [eh-ZAH-nah] Stele" is a convention. Its link to the 4th-century king is not confirmed by an inscription.

---

4. Rule Three: Erection Is the Hardest Load Case

The Great Stela

UNESCO gives the length of Stela 1 as about 33 metres. It lies where it fell, "perhaps during the process of erection." The Met gives about 32 metres and over 550 tons. Across sources the range is 32–33 metres and roughly 500–570 tons, so the safest wording is "over 30 metres" and "roughly 500+ tons."

The leading scholarly view is that it fell during or soon after erection. That is an inference, not a documented event.

The tomb underneath

A 3rd-century tomb called Nefas Mawcha [NEH-fahs MOW-chah] is linked to Stela 1. The Met describes its massive slab roof as lying beneath or beside the Great Stela, and sources differ on which. The spelling and pronunciation of the tomb name also vary. Fattovich (1995), as cited in *Nyame Akuma*, links the Mausoleum to Stela 1.

Local legend says the Devil dropped the slab. Whatever its origin, the legend suggests that the scale of the stonework impressed later observers.

A medieval collapse, with caveats

An Italian Foreign Ministry document reports that a stela collapsed in an earthquake about 1,000 years ago. Which stela, and whether the earthquake cause is supported archaeologically, remain open. If the report is right, the point for engineers is that a tall, slender monolith can fail from forces other than its own weight.

The rule: raising a monolith may be the moment it is most exposed. If the leading view is right, the Great Stela probably shows what happens when the margin runs out.

How the Aksumites raised the stelae is not documented. Any account of ramps, ropes or levers would be invention, so none is given here.

---

5. Rule Four: The Foundation Is Half the Monument

What the 1999 excavation found

In 1999, excavations at the original site of Stela 2 uncovered its foundation and base plates. This was before its re-erection in 2008. Bertrand Poissonnier [bair-TRAHN pwah-so-NYAY] discussed the excavations in a 2012 paper.

The finding is specific. The stela was fastened upright by notched slabs, the "base plates," that partly clamped it. That suggests the stone was not simply stood in a hole, and that a seating was designed to hold the base in place.

The groundwater question

A 2018 paper by Hagos, in *Annales d'Ethiopie*, citing a 2013 UNESCO report, says rising underground water levels at Aksum were attributed to the reinstallation of Stela 2. This is a contested causal claim from a management paper, and the 2013 report has not been checked here, so treat it as an attribution, not a settled finding.

Whatever the final answer, the question itself is the lesson. Tall stone structures depend on what is under them: the foundation design and the hydrology of the site.

The rule: a tall structure is only as stable as the ground and the seating beneath it. The 1999 excavation suggests the Aksumites paid attention to the seating, and the 2018 paper raises the question of whether a modern project had to ask about the ground as well.

---

6. Rule Five: Dating Is Context, Not a Stamp

UNESCO says the surviving stelae date between the 3rd and 4th centuries AD. Phillipson's work, as cited by Poissonnier, dates them only indirectly. David Phillipson [DAY-vid FIL-ip-son]'s 3rd–4th century range rests on associated tombs and pottery. The sources used here report no direct radiometric date on the stelae themselves.

The function carries the same caution. The stelae are widely interpreted as markers of royal or elite graves, and some may also have had commemorative roles. The identity of the individuals buried beneath specific stelae is largely unknown.

The rule: when the object cannot be dated directly, the date is only as strong as the context around it. Honest dating states which evidence it rests on.

---

7. Rule Six: A Monolith Can Be Cut, If the Joints Are Engineered

Removal

In 1937, the Italian Fascist occupation regime removed Stela 2 to Rome. It was cut into three pieces for transport, and it stood in the Piazza di Porta Capena [pee-AHT-sah dee POR-tah kah-PAY-nah].

The cutting was a transport decision, not a bombing and not a collapse. The stela was cut, moved, and set up again in a foreign city.

The long delay

In Article 37 of the 1947 Peace Treaty, Italy agreed to return items looted after 3 October 1935. The return took decades. Disputes over shipping costs played a role in the breakdown of earlier plans. A joint Italian–Ethiopian commission on the return dates to 1997.

From the 1947 promise to the 2005 return is 58 years. From the 1937 removal to the 2005 return is 68 years.

Return and reinstallation

Italy decided to return the stela in April 2005, and the pieces were flown back that year. UNESCO coordinated the reinstallation, and the third and last block was mounted on 31 July 2008, 71 years after the stela was taken. Italy paid for the dismantling, transport and reinstallation. The inauguration ceremony followed in early September 2008.

UNESCO's design for the reinstallation came from an Italian engineer. A secondary summary by Confinity names Giorgio Croci [JOR-joh CRO-chee] as the team leader and says the sections were joined with aramid fibre. Confinity is not a primary source, and neither the name, the role nor the aramid detail has been confirmed against UNESCO reports, so treat them as reported, not established.

The rule: cutting a monolith creates joints, and the reassembly has to be designed around them. The 2005–2008 project is documented as a UNESCO-coordinated effort to do exactly that.

---

8. The Modern Echo

Three modern connections are supported, and each has limits.

Heritage conservation engineering. The 2005–2008 dismantling, transport and reinstallation of a 160-tonne monolith in three pieces is a documented, large-scale conservation project coordinated by UNESCO. The fibre-reinforced joining method rests on Confinity alone.

Foundations and site hydrology. The groundwater question in section 5 echoes the Aksumite base plates. A foundation has to be designed for the ground it sits in, then and now. The causal claim behind that question is contested.

Raising large monoliths. Comparative work discusses how modern techniques for raising broken obelisks echo historical practice, including the Axum stelae. This comes from a single unreviewed Academia.edu paper, and a published version has yet to be found.

One claim must not be made. Nothing here shows that the Aksumites invented a technique used today. How they raised the stelae is not documented, so the echo is one of shared problems, not of inherited methods.

---

9. The Myth to Retire: "No Explanation, So Lost Technology"

A familiar claim says the Aksumites left no explanation for the stelae, so the explanation must be aliens or a lost technology.

There is no basis for it. The tombs, the quarry traces and the later archaeology all fit a local tradition. The excavated base plates at Stela 2 point to practical foundation design. The fallen Great Stela is consistent with a limit being tested and exceeded, which is what happens in a human engineering tradition and not in a mysterious one. The stelae are carved to resemble buildings, which ties them to the architecture they imitate. And Fattovich's 1987 work explores Near Eastern and African megalithic traditions, which is a scholarly debate about origins, not a gap to be filled with fantasy.

Three nearby myths fail for the same reason: the stela was not returned immediately after WWII (it came back in 2005–2008), Ezana did not demonstrably build all of the stelae, and the claim that the Great Stela stood 100 feet high is unsupported, because it never stood, or at best did not remain standing.

The honest gaps are real: who was buried beneath each stela, the exact dates, and how the stones were raised. They are gaps in the record, not evidence of anything external. The most defensible position is the one the evidence supports: a local engineering tradition, working at the edge of what a stone block can do.

---

Want the Full Build?

The Historeelz Blueprint on the Aksumite Stelae goes deeper into the structure, with the comparison data, the foundation detail and the full removal-and-return timeline laid out for study. It is the next step if this article raised more questions than it answered.

---

References and Further Reading

  1. UNESCO World Heritage Centre, "Aksum," whc.unesco.org/en/list/15
  2. UNESCO World Heritage Centre, "Mission accomplished: Aksum Obelisk successfully reinstalled," whc.unesco.org/en/news/456
  3. UNESCO World Heritage Centre, "Aksum obelisk returns to Ethiopia," whc.unesco.org/en/news/116
  4. UNESCO World Heritage Centre, "UNESCO Presents Design for Reinstallation of Aksum Obelisk," whc.unesco.org/en/news/228
  5. UN News, "UNESCO teams successfully reinstall Ethiopian obelisk," 2008
  6. Metropolitan Museum of Art, "The Monumental Stelae of Aksum (3rd–4th Century)"
  7. Metropolitan Museum of Art, "Monumental Architecture of the Aksumite Empire"
  8. Poissonnier, B. (2012), "The Giant Stelae of Aksum in the Light of the 1999 Excavations," *Journal des africanistes / Palethnologie*
  9. Fattovich, R. (1987), "Some remarks on the origins of the Aksumite Stelae," *Annales d'Ethiopie*
  10. Hagos, T. (2018), "Current Management Challenges at the World Heritage…," *Annales d'Ethiopie*
  11. Italian Ministry of Foreign Affairs, "AXUM" (PDF)
  12. *Nyame Akuma* 46, Ethiopia article (citing Fattovich 1995 on the Mausoleum)
  13. The Art Newspaper (1992), "Ethiopian football fans demand return of Axum obelisk"
  14. LA Times (2002), "Ethiopia Wants Its Pillaged Pillar"
  15. NBC News (2005), "Italy to return ancient obelisk to Ethiopia"
  16. Confinity, "Aksum" (secondary source; reinstallation detail only)
  17. "The technology for the raising of the broken obelisks…" (Academia.edu; unreviewed)
  18. Further reading, not consulted directly: Phillipson, D., *Ancient Ethiopia* (1998) and *Archaeology at Aksum, Ethiopia, 1993–97* (2000)
  19. Further reading, not consulted directly: Munro-Hay, S., *Aksum: An African Civilisation of Late Antiquity* (1991)

Add comment

Comments

There are no comments yet.