Himyar

The kingdom that had rain

Drafted from (ed.), Arabs and Empires before Islam (2015)formation

Himyar appears in the inscriptions in the last century BC, in the highlands south-west of Marib, with a capital at Zafar and a citadel called Raydan. It was one of several South Arabian kingdoms and initially not the most impressive.

What it had was different ground. Saba' lived on a dam catching a flood twice a year at the edge of a desert, and on a percentage of a caravan road. Himyar had the Yemeni highlands: terraces on mountainsides with reliable monsoon rainfall, which support a dense agricultural population without any hydraulic engineering at all, and a position facing west towards the Red Sea rather than north towards the caravan route.

In the first two centuries AD that stopped being a secondary advantage. The Roman shipping that went past Arabia on the monsoon, described in the Periplus of the Erythraean Sea, made the caravan road less necessary and the coast more so, and the kingdom with the coast and the agriculture began winning the competition with the kingdom that had the toll.

Three centuries of that competition are documented in the inscriptions and are genuinely confusing — alliances, campaigns, shifting titles, Aksumite interventions from across the Red Sea from the third century onwards. The outcome is legible in the titulature. By the late third century the ruler at Zafar styled himself king of Saba' and of dhu-Raydan; not long after, Hadramawt and Yamanat were added, and South Arabia was one kingdom for the first time.

And the Himyarite kings then went on using the Sabaean royal name for two and a half centuries after Saba' had ceased to exist. They ruled from Zafar, they were Himyarites, and they called themselves kings of Saba' first, because in the country they now held that was the name worth having.

Charibael

Drafted from Anonymous, The Periplus of the Erythraean Sea: Travel and Trade in the Indian Ocean by a Merchant of the First Century (1912)expansion

Somewhere around the middle of the first century AD a Greek-speaking merchant out of Egypt wrote down what he knew about the sailing routes of the Red Sea and the Indian Ocean — the ports in order, the distances, what sold at each, what to bring for the local ruler, and where the anchorage was bad. He was not writing history and he had no view about South Arabian politics. That is why his book is worth more here than most things that were.

His account of the Arabian shore runs like this. Coming down the gulf you reach Muza, a market-town established by law, crowded with Arab shipowners and seafaring men, busy with commerce, sending its own ships across to the African coast and out to Barygaza in India. Three days inland is a city called Sana, in a region called Mapharitis, where there is a vassal-chief named Cholaebus. Nine days further on is Saphar, the metropolis, where lives Charibael — lawful king of two tribes, the Homerites and those living next to them, called the Sabaites; and through continual embassies and gifts, a friend of the emperors.

Saphar is Zafar. Homerites is Himyar. Charibael is Karib'il.

What one sentence establishes

That single line does a great deal of work.

It shows that a Himyarite king was already styling himself over both Himyar and Saba' in the first century AD — two centuries before the titulature settles into the formula the inscriptions record, and a good deal earlier than the point at which the unification is conventionally dated. The claim was being made long before it was made good.

It shows the shape of the state. Zafar is the metropolis and the king is there; San'a is held by a subordinate chief; Muza on the coast is the commercial engine. That is a highland capital with a client in the interior and a port, which is exactly the arrangement that let the highland kingdom outgrow the one at Marib.

And it shows the reach. The merchant notes further along that the East African coast — Azania, the shore south of the Horn — is subject to Charibael and the chief of Mapharitis, and that ships from Muza trade at Socotra. A king in the Yemeni highlands held tributary rights across the water in Africa and had merchants working the island trade off Cape Guardafui.

The port that was destroyed

The strangest sentence in the passage is about somewhere else.

Down the coast past the strait lies a village the merchant calls Eudaemon Arabia — Aden. He explains the name. It was called fortunate because in the early days, when the voyage was not yet made from India to Egypt and nobody dared sail across the ocean from either end, everything came together there: it received the cargoes of both countries, the way Alexandria was then receiving the goods of the world. It was the transfer point of the whole trade.

Then, in the translation used here: but not long before our own time Charibael destroyed the place.

The name is not certain. What the manuscript actually says is that Caesar destroyed it, which cannot be right — no Roman emperor attacked Aden in the first century, and the author elsewhere calls his own sovereign by a different word. Editors have proposed several repairs; the one adopted here reads Charibael, and another candidate is the king of the frankincense country further east. So the sentence is certain in what it reports and disputed in whom it blames, and the polity that gained by it is not in doubt whoever swung the hammer.

What is plain from the rest of the book is the consequence. The shipping that had stopped at Aden now went past it to Muza, which had a bad harbour and a good roadstead and the king's protection, and the customs revenue moved with it. Aden's eclipse was not permanent — by the fourth century it was a working port again, and the Arab geographers write about it in the old glowing terms — but it lost the century in which the sea route to India was being established, and Muza had it instead.

Two hundred years later the kings at Zafar had the whole country. This is what that looked like on the way up: an embassy to a Roman emperor, a client chief holding San'a, tribute from the African shore, ships of its own working the route to India — and, down the coast, the wreck of the port that everyone used to stop at.

The gods stop being named

Drafted from Bowersock, Emblems of Antiquity (2013)peak

Ancient South Arabian religion is attested continuously, in tens of thousands of inscriptions, for well over a thousand years. Almaqah at Marib, Athtar, Shams, Wadd, Sayin at Hadramawt: named gods with temples, priesthoods, festivals, pilgrimages and an unbroken habit of dedicating buildings and campaigns and recoveries from illness to them.

Around 380 the royal inscriptions stop.

Not the inscriptions themselves — the same script, the same language, the same formulas for dating and for building works continue without interruption. What stops is the gods. In their place the texts invoke a single deity: Rahmanan, the Merciful, described as lord of heaven and earth, sometimes with the addition of "and his son" or "the lord of the Jews", with no image, no temple and no consort.

It is one of the most abrupt religious transitions documented anywhere in antiquity, and what makes it strange is how little else changed. There is no conquest, no new dynasty and no foreign occupation at the point it happens. A kingdom simply stopped naming its gods in public and started naming one, and kept doing so for a century and a half.

State conversions are usually loud. A king is baptised, an edict is issued, a rival is put down, an embassy arrives from whichever capital sponsored it, and the whole thing is narrated by somebody with an interest in narrating it. Himyar's is silent. There is no account of it from any side, no missionary, no convert-king with a name attached to the moment. It is visible only in what a state's own documents stopped saying — and the god it turned to is never given a name in them beyond an epithet.

What the epithet meant is the next question, and it is the more interesting one.

A Jewish kingdom in Arabia

Drafted from Bowersock, Emblems of Antiquity (2013)aside?What aside means

The monotheism Himyar adopted around 380 was, on the evidence, Judaism or something very close to it, and by the early sixth century the kingdom was explicitly and publicly Jewish.

The evidence is of several kinds and it accumulated slowly. Inscriptions that invoke Rahmanan and refer to Israel and to the people of Israel; a synagogue and Jewish burial inscriptions at Zafar and elsewhere; a royal formula asking for help "for his people Israel"; Hebrew and Aramaic words appearing in Sabaic texts. The last king, Yusuf As'ar Yath'ar, is described as a Jew by Christian sources that hated him, by later Arabic tradition that made him a legend, and — which is the point — by his own inscriptions.

Historians argue about how much of the population followed the court and about whether what the royal texts describe is rabbinic Judaism, a Judaising monotheism of local development, or something between. Bowersock's book is partly an argument for taking the Jewish character of the sixth-century kingdom at face value against a tradition of scholarly hedging. What is not seriously disputed is that the state was Jewish in its public formulas for something like the last two decades of its independence, and monotheist for a century and a half before that.

A Jewish state is a rare object anywhere between antiquity and the twentieth century, and it is worth being exact about what makes Himyar one. There were Jewish rulers elsewhere: the royal house of Adiabene, a Parthian vassal kingdom in northern Mesopotamia, converted in the first century and remained Jewish for at least two generations. That is a converted dynasty, and it is a different object from a kingdom whose public inscriptions change, whose old gods stop being named across several reigns, and whose confession becomes the axis of its foreign policy. Himyar is the second kind.

Two things follow that are worth stating carefully.

The first is what it was for. South Arabia in the fourth century sat between two expanding Christian powers — Rome across the Red Sea trade and Aksum directly across the water, which had adopted Christianity under Ezana at almost exactly the moment Himyar's inscriptions changed. A monotheism that was neither of theirs was a way of not being a client of either, and the choice reads as strategic as much as devotional — a confession that came with no foreign protector attached and no foreign patriarch to answer to.

The second is that it did not work. The confessional choice that was meant to keep two Christian empires at arm's length gave them, in the end, the reason to come.

Najran

Drafted from Bowersock, Emblems of Antiquity (2013)contraction

Najran is an oasis town on the northern edge of Himyarite territory and it was Christian — a substantial community with a church and a bishop, converted in the fifth century, and connected to the Christian world across the Red Sea and up through Syria.

In 522 or 523 Yusuf As'ar Yath'ar came to the throne in circumstances involving an Aksumite garrison he expelled, and moved against Najran. The town was besieged, surrendered on terms, and the terms were not kept. The leading men were killed; the accounts describe the burning of a church with people inside it, and mass executions in a trench.

The Christian sources are immediate and detailed. A letter of Simeon of Beth Arsham, written within months, circulated the story through the Syriac-speaking churches; the Book of the Himyarites preserves a fuller account; martyrologies followed. They are hostile witnesses reporting an atrocity, which is the usual epistemic position for events of this kind, and the archaeological and epigraphic record corroborates the essentials rather than the numbers.

The Qur'an refers to it, at Surat al-Buruj: the people of the ditch, the fire with its fuel, seated and watching what they did to the believers. The identification with Najran is the traditional one, and it is not the only reading; it is worth carrying as a tradition rather than as a settled fact about the text.

What the massacre did matters more than its scale, which will never be fixed. It made Himyar's confession a matter for Constantinople. Justin I had every reason to want a Christian power on the Red Sea route to India and now had a casus belli to give the Aksumite king; the letters that reached the empire were used, and were meant to be used, to produce an invasion.

It arrived in 525.

A slave of a Roman merchant

Drafted from Caesarea, History of the Wars, Books I-II: The Persian War (1914)contraction

Procopius was writing in Constantinople in the 540s about the eastern war, and he stopped to explain what had become of Himyar, because Justinian had tried to use it and failed.

His account of the conquest is brisk. The king of the Ethiopians, a Christian and a devoted one, found that a number of the Himyarites on the opposite mainland were oppressing the Christians there outrageously — many of the oppressors being Jews and many of them holding to the old faith. He collected ships and an army, beat them, killed the king and a great many others, installed a Christian Himyarite named Esimiphaeus in his place on condition of an annual tribute, and went home.

Then the sentence that decides the next forty years. Many in the Ethiopian army — slaves, and everyone readily disposed to crime — were quite unwilling to go back with the king, and stayed behind, because they wanted the land of the Himyarites; for it is an extremely goodly land.

Those men, with others, rose against Esimiphaeus not long after, shut him in a fortress, and made a king of their own. His name was Abraha. Procopius, who plainly found this remarkable enough to specify, records that Abraha was a Christian, and a slave of a Roman citizen engaged in the shipping business at Adulis.

The attempts to undo it went the same way as the conquest. The Ethiopian king, angry on Esimiphaeus's behalf, sent three thousand men under a relative of his; that army also decided it would rather stay in a goodly land, opened negotiations with Abraha without telling its commander, and then, when the fighting started, killed the commander and joined the other side. A second expedition was heavily beaten and went home. After that, says Procopius, the king of the Ethiopians became afraid and sent nothing more. Abraha eventually agreed to pay tribute to the next Ethiopian king, and by that means made his own position secure.

What Constantinople wanted out of all this is the part worth keeping, because it shows what Himyar looked like from a thousand miles away. Justinian sent an ambassador, Julianus, to both courts with a proposal built on their shared religion. The Ethiopians were to buy silk in India and sell it to the Romans, which would make them rich and stop Roman money going to Persia. The Himyarites were to install a fugitive named Qays as captain over the Ma'add and invade Persia overland with their own army and the Saracens.

Both kings promised. Neither did anything.

Procopius gives the reasons and they are entirely practical: the Persians had the Indian harbours sewn up, and the desert was too wide to cross in order to attack people tougher than themselves. Abraha, later, when his power was finally secure, promised Justinian many times that he would invade Persia, and once actually set out, and turned straight back.

That is the position of South Arabia at the end of its independence, described by a man in the imperial chancery who had no reason to flatter it. It had been conquered by one Christian power, governed by a runaway slave of a Roman shipowner, and was being solicited by a third as a strategic instrument against a fourth — and its own answer to the solicitation was that the desert was wide and the enemy was tougher than they were. There is no more accurate summary of what it had become, or of why the Sasanians took it a generation later without much difficulty.

Kaleb, Abraha, and the elephant

Drafted from Bowersock, Emblems of Antiquity (2013)contraction

Kaleb of Aksum crossed the Red Sea in 525 with Byzantine encouragement and, according to the sources, Byzantine ships; Yusuf was defeated and killed, and tradition has him riding his horse into the sea.

Abraha, who displaced Kaleb's client within a decade, then ruled Himyar for something like thirty years, and he is the best documented ruler of South Arabia in the whole period, because he cut long inscriptions and one of them survives at Marib.

That inscription records the repair of the dam, which is worth noting: the hydraulic system on the Saba' record was still being maintained in the 540s, a thousand years after it was begun, by an Ethiopian general ruling a formerly Jewish kingdom in the name of a Christian god.

He built a church at Sana'a that the Arabic sources describe as extraordinary, and this is where Himyar enters the tradition every Muslim reader knows. Abraha, in that tradition, marched north on Mecca to destroy the Ka'ba and divert the pilgrimage to his own church, with an elephant in his army; the elephant knelt and would not advance, and birds dropped stones on the host, and it was destroyed. The Qur'an's Surat al-Fil refers to it, the year is called the Year of the Elephant, and it is traditionally the year of Muhammad's birth.

Historians treat the campaign as probable and the date and the details as uncertain: an inscription of Abraha's records a northern expedition, and the traditional dating of the Year of the Elephant to about 570 is difficult to reconcile with the epigraphic chronology of his reign. It is not a question to be settled here. What is not in doubt is narrower: the last independent power in South Arabia was an Ethiopian Christian kingdom, and the Arabic tradition that Islam grew out of remembered it as the thing that came closest to destroying Mecca.

What was left

Drafted from (ed.), Arabs and Empires before Islam (2015)end

Around 570 the Sasanians took Yemen. The Arabic accounts have a Himyarite noble, Sayf ibn Dhi Yazan, going to Constantinople for help against the Ethiopians, being refused, going to Ctesiphon instead, and being given a force — reportedly of convicts — which took the country. Whatever the details, South Arabia was a Persian province by the 570s and remained one until the Muslim conquest.

There are therefore two dates for the end and they measure different things: 525, when the kingdom stopped being Himyarite, and about 570, when the country stopped being Aksumite. There is no third date on which South Arabia recovered its independence, because it did not.

The dam went in the same period. The last great breach of the Marib works is dated to the sixth century, and after it the irrigation was not restored; the population that had lived on it dispersed, and the migration of South Arabian tribes northwards into the peninsula is a large theme of early Arabic genealogical writing. The Qur'an's account of the flood of the dam belongs to this century and not to the kingdom at Marib, which had been gone for three hundred years by the time the water stopped.

What South Arabia was when Muhammad was born is worth setting out plainly. A formerly wealthy country whose irrigation had failed, ruled by a Persian governor, with a large Jewish population and a large Christian one and the old temples abandoned, remembering an Ethiopian invasion and a Jewish king and a thousand years of kingdoms whose inscriptions were still standing in the fields.

Yemen then spends three hundred years as a province of somewhere else, and the transition into the first of those centuries is characteristic of the whole period. There was no conquest of Yemen by the Muslim armies. There was the surrender of a Persian governor, which is the last act of Sasanian rule in the country and also the first act of Islamic rule in it, performed by the same man without leaving his post.

The Tubbas

Drafted from trans.), Yaman: Its Early Mediaeval History (1892)afterlife

The Arabic word for the great kings of pre-Islamic Yemen is tubba'. It is not a dynasty and not quite a title in the ordinary sense; it is what the Himyarite kings had been, collectively, once nobody living remembered them. And for the next thousand years it functioned in Yemen as the standard against which power was measured.

Follow it forward through the chronicles and it is everywhere.

Ibn Khaldun, describing the arrival of the Abbasid officer who founded a dynasty on the Yemeni coast in 819, writes that he took the Tihama, founded a city to live in, was acknowledged in Hadramawt and Shihr and the country of Kinda, and so acquired the power and dignity formerly enjoyed by the Tubbas. That is the measure applied to a governor sent from Baghdad by a caliph: not that he was the greatest of the caliph's officers, but that he had got hold of what the old kings had.

At San'a in the same period the ruling family were the Banu Ya'fur, whom the chroniclers call descendants of Himyar and one of the noblest houses to have survived the rule of the Tubbas. They held San'a, Bayhan, Najran and Jurash for a century and a half. When the coastal dynasty's greatest ruler was old and his provinces were slipping, the man at San'a who read his name in the Friday prayer and sent him nothing was a Himyarite, and the historian who records the insult records the pedigree in the same sentence, as though it explained something.

The genealogists went further than that. The Zaydi imams at Sa'da, who were Alids from Medina and whose whole claim rested on descent from the Prophet's family, are also attached to the posterity of the Tubbas in one of Ibn Khaldun's chapters — which, as his English editor drily notes, is difficult to derive from the pedigree Ibn Khaldun himself prints elsewhere. Two incompatible pedigrees for the same men, and the reason to want the second one is that in Yemen it helped.

What was actually left to see

The physical remains encouraged this. They were considerable and they were visible.

Zafar, where the kings had reigned, was still known in the twelfth century as the seat of empire of the Tubbas, and by then it was in ruins. At San'a there stood the palace called Ghumdan, reduced to a high mound, of which the historian Umara says flatly that no later king of Yemen ever built one like it, or so lofty. Towns up and down the country were said to date from the Tubbas, and the houses in some of them were built of their stone.

That is the ordinary condition of medieval Yemen: a landscape of famous ruins that everybody could point at, belonging to kings whose language could no longer be read, in a script that had gone out of use, and whose religion — a Jewish monotheism that ran the state for a century and a half — had been so completely replaced that the tradition remembering them barely mentions it.

What the word was for

It is easy to read all this as nostalgia and it is not quite that.

Yemen after Islam was a country of many simultaneous authorities — a caliph's governor on the coast, an Alid imam in the northern highlands, a Himyarite family at San'a, chiefs holding mountain fortresses that no lowland army could take. Nobody in it held everything, and everybody in it needed an argument for why they should. The Tubbas were the one thing all of them could point at: a period, safely closed, in which Yemen had been a single country ruled from inside itself rather than a province of somewhere else.

So the word did a job. To say a man had the dignity of the Tubbas was to say he had reassembled the country; to say a family had survived the Tubbas was to say it was older than the conquest and owed its standing to nobody. Both claims were about the present.

The kingdom that generated the word had been Jewish, wrote in a script that had gone out of use, spoke a language nobody could still read, and ended by being invaded from Africa. Almost none of that is in the tradition. What survived was the bare proposition that Yemen had once been great and had been great on its own — which is the most portable thing a vanished state can leave, and the hardest to disprove.