Kingdom of Aksum

A third power

Drafted from Munro-Hay, Aksum: An African Civilisation of Late Antiquity (1991)formation

Some time in the third century the Persian prophet Mani listed the great kingdoms of the world. There were four: Rome, Persia, Aksum, and China.

Three of those need no introduction. The fourth is a highland kingdom in what is now northern Ethiopia and Eritrea, and anybody meeting it for the first time is entitled to ask what it was doing in that company.

A port and a plateau

The answer is a combination almost nowhere else had.

Aksum sits at about two thousand metres on the northern Ethiopian plateau — good agricultural country, reliable rainfall, defensible, and cool enough to be healthy. Adulis, its port, sits on the Red Sea about a hundred and seventy kilometres away and eight hundred metres of descent below.

The plateau fed the state and the port connected it to everything. The Periplus of the Erythraean Sea, a Greek merchant's handbook of the first century, has Adulis exporting ivory, rhinoceros horn, tortoiseshell and obsidian, and importing cloth, glass, brass, iron and wine, with a note that the king — the Periplus calls him Zoskales — is careful about his money and knows Greek literature.

That is the earliest outside description of the kingdom, and it is already describing a state with a customs regime and a ruler who read Greek.

The route

What made a small highland kingdom one of Mani's four is where the goods were going.

The Red Sea was the Mediterranean's route to India. Roman ships from Egypt ran down it with the monsoon; African ivory, incense and gold came up it; and Aksum sat at the point where the African interior's goods met the sea lane. The kingdom lived on the toll, in the way Palmyra and the Nabataeans lived on theirs, and at a considerably larger scale than either.

The other half of the income was tribute. Ezana's campaign inscriptions, which are the closest thing this kingdom left to an administrative record, list what came back from expeditions against neighbouring peoples: cattle by the thousand, sheep, captives, and the submission of named rulers.

Two languages, in stone

The kingdom wrote in Ge'ez and in Greek and frequently in both at once. Ezana's inscriptions survive in versions in Ge'ez in its own script, Ge'ez in South Arabian letters, and Greek — the same text, three ways, on the same monument.

Greek because the customers were Greek-speaking. Ge'ez because the population was. A state that puts both on a royal monument is telling two audiences the same thing, and the sixth chapter is about what happened to the second of those scripts under Aksumite hands.

The stelae

Drafted from Phillipson, Foundations of an African Civilisation: Aksum and the Northern Horn, 1000 BC - AD 1300 (2012)expansion

At Aksum, in a field, there is a granite monolith lying broken in five pieces. It is about thirty-three metres long and weighs something in the order of five hundred tonnes. It was carved to look like a multi-storey building — a false door at the base, rows of beam-ends and windows up the shaft, a curved top — and it was quarried four kilometres away and dragged to the site.

It fell, almost certainly while being raised, and was never re-erected.

It is the largest single piece of stone anybody in the ancient world attempted to stand upright. The Egyptian obelisks are smaller. The unfinished obelisk at Aswan is bigger and was never moved.

What that is evidence of

The stelae are pre-Christian, they mark royal graves, and there are more than a hundred and forty of them of which the six or seven largest are the carved architectural ones. The one still standing is twenty-four metres. Another, taken to Rome by Mussolini's forces in 1937, was returned in 2005 and re-erected in 2008.

For a state that left no administrative documents, they are the institutional record. Quarrying, transporting and raising a five-hundred-tonne monolith requires a labour force that can be assembled and fed, an engineering tradition that has done it repeatedly at increasing scale, and a political authority that can commit resources on that order to a grave.

Ezana's inscriptions name regiments, which suggests a standing force rather than a levy, and that is an inference from a handful of names. The stelae are the harder evidence for the same underlying claim: whatever Aksum's arrangements were, the state could direct organised labour at scale, and it did so for six or seven generations running.

And what it is not evidence of

Not how the throne passed, which nobody knows. Not the fiscal machinery beyond the customs and the tribute. Not who was buried under most of them.

Phillipson is careful about the limits. The excavated tombs under the great stelae were robbed in antiquity. The names on the king lists cannot be securely matched to the monuments. What survives is the fact of the effort and almost nothing about the people who commanded it.

The turn

Then the stelae stop.

They stop in the fourth century, at about the point of Ezana's conversion, and the royal burials move to churches. That transition — from monolith to basilica, both of them the biggest structure anyone in the region was building — is the clearest single statement of what Ezana's conversion did to the state, and it is legible in the ground before any text has to be consulted.

Ezana

Drafted from Munro-Hay, Aksum: An African Civilisation of Late Antiquity (1991)peak

Ezana ruled somewhere around the middle of the fourth century, and for a kingdom with no administrative archive and no chronicle he is unusually well attested: he is the reign all three kinds of evidence converge on.

His inscriptions survive in Ge'ez, in Greek and in Ge'ez written with South Arabian letters. His coins survive in quantity. And the thing that happened during his reign can be dated from the money without anybody having written it down.

The change of symbol

The early Aksumite gold carries a disc and a crescent above the king's head — the pre-Christian symbol, on issues from the late third century onward.

From part-way through Ezana's reign the same coinage carries a cross.

The inscriptions change in the same reign and in the same direction. The earlier ones open by invoking Mahrem, the unconquered war god whom the Greek versions render as Ares, and describe the king as his son. The later ones invoke the Lord of Heaven, who has given the king power, and end with a formula about the might of the Lord of All.

The structure of the claim did not change at all. The king's right came from above before the conversion and after it; the formulae kept recognisably the same shape while the name inside them was replaced.

It is very rare to be able to watch a state change its god inside one reign, in two independent media, with the wording of the claim otherwise untouched.

Frumentius

The story of how it happened comes from Rufinus, writing in Latin around 400, who says he had it from the brother of the man involved.

Two young Syrians from Tyre, Frumentius and Aedesius, were on a ship that put in on the Red Sea coast; the crew were killed and the boys were taken to the Aksumite court, where they were kept and given positions. Frumentius eventually went to Alexandria, asked Athanasius to send a bishop for the Christians in Aksum, and was told to go and be one himself.

Whatever the story is worth, the institutional result is not in doubt and it is the most consequential thing the kingdom did. The Ethiopian church took its head from Alexandria and went on doing so for sixteen hundred years — the abun was an Egyptian monk appointed by the Coptic patriarch until 1959. That made it Miaphysite, aligned with Egypt and Armenia and Syria rather than with Constantinople and Rome, and it is why what Aksum carried outward was Miaphysite Christianity rather than Christianity in general.

The alignment is also why the church survived. When Egypt became Muslim in the 640s, the Alexandrian patriarchate continued under Muslim rule and went on consecrating bishops for a Christian kingdom in the Ethiopian highlands that no Muslim power ever conquered. Two Christian establishments cut off from the rest of Christendom, each supplying something the other needed, for a thousand years.

The order of events

One thing is worth stating plainly because the sequence is unusual.

Ezana's conversion is roughly contemporary with Constantine's and earlier than the Christianisation of most of the Roman west, earlier than Ireland, earlier than the Franks by a century and a half, and earlier than the Rus by six hundred and fifty years. Armenia has the earlier claim by a couple of decades and is always cited; Aksum is usually not in the list at all, which is a fact about who has written the lists rather than about the fourth century.

Gold, and a change of symbol

Drafted from Munro-Hay, Aksum: An African Civilisation of Late Antiquity (1991)aside?What aside means

Aksumite coinage is the best evidence the kingdom left about itself, and what it shows is a state explaining itself through a currency rather than through inscriptions.

It was struck in gold, silver and bronze from the late third century until some point in the seventh. The legends are in Greek on the early issues — the language of the Red Sea trade — and shift to Ge'ez later. The coins have been found in Arabia, in Egypt, and in India, at Karnataka and in a hoard in the south, which is the physical trace of the Red Sea and Indian Ocean route the kingdom lived on.

Why gold matters

Very few states minted a gold coinage. On the common contemporary reckoning there were three in the world the Mediterranean knew: Rome, Persia and Aksum.

That is not a statement about wealth alone. Issuing gold means having gold, having a trade that requires denominations that large, and having a stamp that merchants outside your territory will take at face value rather than by weight. The third condition is the political one and it is the reason Mani's list in the first chapter includes this kingdom.

The silver and bronze issues were for domestic use and several carry an unusual feature: inlaid gold on the design, applied to a base-metal coin, which is a technique nobody else in the period was using in quantity.

Dating a conversion from a die

The religious change is legible in the same series and this is the part that makes the coinage more than an economic indicator.

The early issues carry a disc and crescent above the king's head. From part-way through Ezana's reign they carry a cross. The inscriptions shift in the same reign from invoking Mahrem to invoking the Lord of Heaven, and the third chapter is about what that did.

Nobody in Aksum wrote the conversion down in a form that survives. It is dated from the money.

That is worth holding onto, because it is the condition of everything known about this kingdom. Most ancient states are reconstructed from paperwork — delivery receipts, provincial files, temple accounts. Aksum left none of it. What it left is coins, royal inscriptions and very large stones, and every statement anybody makes about how it worked is built out of those three things.

When it stops

The coinage ends in the seventh century, and the reason is the same reason everything else about the kingdom ends: the Red Sea reoriented after the Arab conquests, and the trade that had required gold denominations went elsewhere.

For a kingdom whose best evidence is its money, the end of the money is the end of the evidence. Aksum goes on for roughly another three hundred years, and almost nothing is known about any of them.

Across the sea

Drafted from Bowersock, Emblems of Antiquity (2013)peak

In 525 Kaleb of Aksum crossed the Red Sea with a fleet, destroyed the kingdom of Himyar in south-western Arabia, and killed its king.

Himyar was a Jewish kingdom in the Yemeni highlands, and it had been persecuting the Christian community at Najran. The invasion was mounted in response, with Byzantine encouragement and, on some accounts, Byzantine ships. It ended Himyar.

For about half a century afterwards an African kingdom held both shores of the southern Red Sea.

What the sources are

Unusually good, and unusually strange.

Bowersock's book reconstructs the war largely from two inscriptions carved on a throne at Adulis, a monument that no longer exists and is known only because a sixth-century Egyptian merchant called Cosmas Indicopleustes was asked by the local governor to copy the texts before an expedition and did so, and included them in a book about why the world is flat.

Beside that there is Syriac hagiography about the martyrs of Najran, Sabaean inscriptions from the Himyarite side, a Greek diplomatic account, and Ethiopian tradition. Five bodies of evidence in five languages about a war between two kingdoms that most narratives of late antiquity do not mention.

The high-water mark

That was the furthest Aksum ever reached and the last thing it did that the wider world troubled to write down.

An Aksumite viceroy named Abraha eventually made himself independent in Arabia and ruled there for decades; Islamic tradition remembers him for an expedition against Mecca with an elephant, in a year the tradition calls the Year of the Elephant. Whether that expedition happened as described is argued about and is not settled.

The occupation ended when a Sasanian expedition took Yemen in the 570s, and Aksum never crossed the water again.

The reign that mattered less

Kaleb took Aksum further than any other king and is not the reign the kingdom is remembered for, which is worth being explicit about.

Ezana's reign produced the conversion, the inscriptions, the change on the coinage and a church that is still standing. Kaleb's produced a successful overseas expedition that held for fifty years and left nothing behind on either shore: no Aksumite institution in Arabia, no lasting church, no dynasty. Abraha's viceroyalty went its own way and was then taken by Persia.

Reach and consequence come apart here about as cleanly as they ever do.

Kaleb is said to have abdicated, become a monk, and sent his crown to Jerusalem. That is the kind of detail the Ethiopian tradition preserves and the external sources do not, and it should be read as tradition.

The copy they kept

Drafted from Indicopleustes, The Christian Topography of Cosmas, an Egyptian Monk (1897)peak

Around 518 a merchant from Alexandria was at Adulis, the Aksumite port on the Red Sea, when the governor asked him to do a piece of clerical work. The king was preparing to invade Himyar and had written down for copies of two Greek inscriptions standing outside the town.

The merchant and a colleague named Menas went and copied them. One set went to the governor. They kept a duplicate, and twenty-five years later the merchant — by then a monk, writing a book to prove that the earth is a flat rectangle shaped like the Tabernacle — printed it. That is the only reason the text exists.

What was standing there

Cosmas describes the objects with a surveyor's care, because he had measured them.

At the entrance to the town, on the western road, was a marble chair about two and a half cubits high, of the shape he calls a bishop's throne, cut with Greek characters. Behind it stood a slab of basanite three cubits high, quadrangular, rising to a point at the top so that the sides sloped down like a lambda. The slab had fallen over behind the chair and its lower part was broken away. Both were covered in Greek.

On the back of the chair were carved figures of Hercules and Mercury. His companion Menas thought they stood for strength and wealth. Cosmas preferred to take Hercules as strength and Mercury as speech, on the authority of the Acts of the Apostles, where the people of Lystra call Barnabas Jupiter and Paul Mercury because Paul was the chief speaker.

The inscription on the slab was Ptolemaic — a king of Egypt recording an expedition into Asia with elephants from the Troglodytes and Ethiopia. The one on the chair was Aksumite, and it is the only extended contemporary account of an Aksumite king's campaigns that survives in any language.

Why the king wanted a copy

Because he was about to do the same thing, and the stone said it had been done before.

The king was Kaleb, whom Cosmas calls Elesbaan, and he was fitting out an expedition across the Gulf against the Jewish king of Himyar, who had been killing the Christians of Najran. That war is the high point of Aksumite power: an African kingdom crossing the Red Sea, destroying an Arabian one, and governing it through a client for half a century.

The governor of Adulis was ordered to send up copies of the inscriptions. The inference is plain enough and Cosmas does not draw it: a king preparing an overseas campaign wanted the text of the local monument recording the reach of earlier rulers, and the traders who happened to be in port were told to transcribe it.

The other thing he was told about

Cosmas also records how the kingdom paid for itself, and he is careful to say that this part is not his own observation.

Every other year the king of the Aksumites sent agents through the governor of Agau into the country called Sasu, on the edge of the gold-bearing land, to bargain for gold; some five hundred other traders travelled with them. They took oxen, lumps of salt and iron. At the mining district they camped and fenced the camp with a great hedge of thorns, slaughtered the oxen, cut them up, and laid the pieces of meat on top of the thorns along with the salt and the iron.

Then come the natives bringing gold in nuggets like peas, called tancharas, and lay one or two or more of these upon what pleases them — the pieces of flesh or the salt or the iron, and then they retire to some distance off. Then the owner of the meat approaches, and if he is satisfied he takes the gold away, and upon seeing this its owner comes and takes the flesh or the salt or the iron. If, however, he is not satisfied, he leaves the gold.

Business was done that way, Cosmas says, because the languages were different and interpreters were hardly to be found. The expedition stayed five days or thereabouts, and went home armed, because the tribes on the road knew they were carrying gold. Going and returning took six months.

Two kinds of evidence

Both halves of the chapter are what a merchant notices, and they are worth separating.

The inscriptions Cosmas measured and copied himself, and his transcript has been the basis of everything written about early Aksumite conquests ever since. The gold trade he had from the men who made the journey, and the silent exchange he describes is a story told about distant peoples in a great many places and periods, including by Herodotus about the Atlantic coast. It may be exactly right. It is not the same order of testimony.

What is not in doubt is the shape of the operation: a kingdom in the highlands, reaching the sea at one port, sending a state-organised caravan of five hundred men on a six-month round trip every second year to buy gold with salt, iron and beef, and using the proceeds to put an army across the Red Sea.

The script

Drafted from Phillipson, Foundations of an African Civilisation: Aksum and the Northern Horn, 1000 BC - AD 1300 (2012)aside?What aside means

Aksum's most durable act was not a conquest. It was a change made to a set of letters in the fourth century, and it is being used to type this sentence's Ethiopian equivalent today.

What was adapted, and by whom

The Ge'ez script descends from the South Arabian alphabet — the same family the Sabaean inscriptions across the Red Sea are written in. That derivation happened before Aksum's own history begins, and belongs to whoever carried the letters over the water.

What happened under Aksum, in the fourth century and apparently in the same generation as the conversion to Christianity, is that the script acquired systematic vowel marking. Each consonant sign was modified into a set of seven forms, one for each vowel, by adding strokes and loops and by changing the shape of the base letter.

That turned a consonantal alphabet into a syllabary. It is a different kind of act from the one that usually spreads a writing system: Hittite scribes took Mesopotamian cuneiform and wrote their own language in it, Urartian scribes did the same with Assyrian cuneiform, and the Cyrillic workshops took Greek letters for Slavonic. Each of those borrowed a system and left it standing. The Aksumite change rebuilt the system itself — the inherited consonantal skeleton is South Arabian, and the vowels are Aksum's.

What it is still doing

Ge'ez as a spoken language died. The script did not.

It is the script of Amharic, of Tigrinya, of Tigre, and of Ge'ez itself as a liturgical language, and it is read and written daily by tens of millions of people. It is one of a very small number of writing systems in use today that belongs neither to the Phoenician line nor to the Chinese one, and the shape it has — a consonant with its vowel built into the letterform — was given to it in this kingdom, in the fourth century, and has not been given up since.

The sea closed

Drafted from Phillipson, Foundations of an African Civilisation: Aksum and the Northern Horn, 1000 BC - AD 1300 (2012)contraction

Aksum was a port and a plateau, and in the seventh century it stopped being a port.

The Arab conquests of the 630s and 640s took Egypt, the Levant and eventually the whole northern Red Sea littoral. Aksum itself was not invaded — early Islamic tradition remembers the kingdom as the place that sheltered the Prophet's followers when they fled Mecca, and treats it as exempt from jihad on that account, which is a remarkable piece of diplomacy for a Christian state to have achieved in the seventh century.

But the trade Aksum had lived on now ran between Muslim ports under Muslim arrangements, and Aksum was not one of them. Adulis stopped being an international harbour. The coinage stopped in the same century. A kingdom that had lived on tolls and tribute lost the first of those two outright.

What was left was the plateau: an agricultural kingdom with no port, cut off from the Mediterranean world it had been a member of, and with a church whose head still came from Alexandria and now came from a Muslim-ruled Alexandria.

The three hundred years

Almost nothing is known about the last three centuries.

The causal chain above — the conquests reoriented the trade, and the loss of the trade killed the state — is a reconstruction rather than a report. Nobody in Aksum wrote it down. What can be observed is that the coins stop, the inscriptions stop, the capital declines archaeologically, and the political centre of the Ethiopian highlands moves south over the following centuries.

Phillipson's argument, and the reason he is cited here against the older account, is that this is a transformation rather than a collapse — that the continuity into medieval Christian Ethiopia is stronger than earlier historians allowed, and that the search for a moment of fall is the wrong shape of question.

940

Drafted from Phillipson, Foundations of an African Civilisation: Aksum and the Northern Horn, 1000 BC - AD 1300 (2012)end

Aksum has no fall. No conqueror is recorded, no battle ends it, and the date usually given is a convention standing in for a process nobody described.

Gudit

Ethiopian tradition supplies a moment anyway.

It names a queen — Gudit, or Yodit — said to have destroyed Aksum around the end of the tenth century, burned the churches and killed the royal line. She appears in later Ethiopian chronicles, and an Arabic letter of the period refers to a queen attacking the Christian kingdom. Almost nothing about her can be established, and the conventional end date comes from that neighbourhood of tradition rather than from anything datable.

So the year Aksum ends is a number drawn from a story, and it is better to say so than to print it as though it were the same kind of date as its beginning.

What came after

The Zagwe dynasty, from roughly the twelfth century, ruling from further south and building the rock-hewn churches at Lalibela. Then the Solomonic dynasty from 1270, which claimed descent from Solomon and Sheba and ruled, with interruptions, until 1974.

Whether either of them claimed to be continuing Aksum — as against ruling the same highlands and the same church — is a real question, and the sources here do not settle it. What is not in question is that the church, the script, the liturgical language and the highland Christian state all carried on without a break. The kingdom ends and the thing it was does not, which is the least satisfying kind of ending and the one the evidence supports.

The place of abode

Drafted from Nagast, The Queen of Sheba and Her Only Son Menyelek (1922)afterlife

Four hundred years after Aksum stopped being the capital of anything, a scribe called Isaac finished a book which asserted that God lived there.

The Kebra Nagast — the Glory of Kings — was assembled at Aksum in the early fourteenth century. Budge summarises what its compiler believed he was doing, from Isaac's own colophons, and the third of the three propositions is the one that concerns this kingdom:

That the God of Israel had transferred His place of abode on earth from Jerusalem to Aksum, the ecclesiastical and political capital of Ethiopia.

Isaac says he toiled at it for the glory of the kingdom of Ethiopia, and for the going forth of the heavenly Zion, and for the glory of the King of Ethiopia.

What the claim required

That the Ark of the Covenant had been taken from Jerusalem to Aksum by Menelik, Solomon's firstborn son by the queen of Sheba, and had stayed there.

Isaac is entirely untroubled by how Menelik obtained it. He glories in it, on the ground that the removal was manifestly God's will: God had judged that the Jews were no longer worthy custodians of the Ark in which His presence was, and the Ark itself wished to go. Ethiopia, in the words of the psalm, had stretched out her hands unto God, and He had gone to her.

That is a very large claim to make about a provincial town in the northern highlands, and it was made at the moment the town had least going for it politically.

When it was made, and why

The political centre of Christian Ethiopia had moved south out of Aksum centuries earlier. The Zagwe ruled from Lasta and built the churches at Lalibela. Then in 1270 a dynasty took the throne claiming descent from Solomon through Menelik, and it is that dynasty's need that the book serves — Isaac's own tradition held that God had permitted the Solomonic line to be broken and the Zagwe to reign, and had then restored it.

So a book compiled at Aksum argued that the legitimacy of a dynasty ruling from elsewhere depended on a relic held at Aksum, in a church at Aksum, under the care of the clergy of Aksum.

What that bought the town

Six hundred years of being the ceremonial centre of a state that was governed from somewhere else.

Ethiopian emperors went to Aksum to be crowned, at the church of Maryam Tsion, long after there was nothing else there. The chapel beside it is where the Ark is said to be kept, watched by a single guardian monk who does not leave and who alone may see it. The claim has never been withdrawn and is not subject to inspection.

A city that outlived its own state twice over

The stelae were already ancient and already fallen when Isaac was writing — a row of granite monoliths carved to look like storeyed buildings, raised by a pagan dynasty for its own dead, standing among the churches of the Christian kingdom that succeeded it.

Aksum's political existence lasted perhaps seven hundred years. Its existence as the place where Ethiopian kingship was authorised has now lasted a thousand more, on the strength of a book written by a scribe who believed that the God of Israel had moved house.