Saffarid Dynasty

The coppersmith

Drafted from Bosworth, The History of the Saffarids of Sistan and the Maliks of Nimruz (247/861 to 949/1542-3) (1994)formation

Ya'qub ibn al-Layth was a saffar, a coppersmith, in Sistan. The nickname stuck to the dynasty, which is unusual — most dynasties are named for an ancestor or a place, not a trade — and it stuck because it was accurate and because his enemies found it useful.

Sistan in the 850s was held by the ayyars, irregular volunteer bands who had grown out of frontier warfare against the Kharijite rebels and the pagan territories to the east. Ya'qub rose through these. By 861 he had control of Zarang, and from there he did something that is easy to state and was genuinely new in the region: he built a standing army out of the frontier irregulars, paid it himself, and went looking for provinces.

He did not seek an appointment first. There is a well-known exchange, recorded in the Persian Tarikh-i Sistan, in which a caliphal envoy arrives with a document of investiture and Ya'qub shows him a sword and a plate of onions and coarse bread, and says roughly: this is my food, and if I lose I go back to it, and this is how I got what I have. Whether or not it happened, it is the right story for what he did.

Every other polity in this thread acquired legitimacy first and power afterwards, or acquired both at once. The Saffarids are the case that ran the sequence backwards, and it is worth watching what that cost them.

What Sistan had instead of a government

Drafted from Bosworth, The History of the Saffarids of Sistan and the Maliks of Nimruz (247/861 to 949/1542-3) (1994)formation

Sistan is the basin of the Helmand, on the edge of the Iranian world, and in the ninth century it was the kind of province a caliphate holds on paper. It had a governor and it had a frontier, and between the two it had something the settled provinces did not: a permanent condition of low-level war, and a social form that had grown out of it.

The ayyars are hard to translate and worth not translating. They were bands of armed young men — volunteers, irregulars, in the hostile sources brigands and in the friendly ones a brotherhood with a code of loyalty, hospitality and protection of the weak. They existed across the Iranian world in this period, in the towns as well as the frontier districts, and what made Sistan's different was that there was constant fighting available. The Kharijites had been established in the region for two centuries, and beyond it lay territory that was still not Muslim; a man with a following and no prospects could find employment.

Ya'qub came up through that. So did everyone he commanded.

What he did with that material is the reason anybody remembers him. He took a form of organisation that was episodic, local and unpaid — a band follows a leader while the leader is winning — and made it permanent: a standing force that he paid himself, out of what it took, on a schedule. When he beat the Kharijites in Badghis he did not disperse the survivors; he enrolled them as a formed contingent in his own army, the jaysh al-shurat, the army of the Kharijites, which then fought for him.

The result was not a mercenary army, and the distinction matters. Mercenaries can leave. What Ya'qub built was an army with no other employer, no tribal home to return to, and no land, whose entire existence was the campaign — and it is the reason a coppersmith from the far edge of the empire could take Nishapur inside twelve years and then turn west towards Baghdad.

Taking Khurasan, and asking afterwards

Drafted from Bosworth, Encyclopaedia Iranica · sourceexpansion

Between 867 and 873 Ya'qub took Herat, Kirman, Fars and finally Nishapur, ending the Tahirids. He sent the captured Tahirid governor to the caliph along with, in effect, a demand that the conquest be recognised.

Baghdad's position through all of this is the interesting one. The caliph al-Mu'tamid alternated between granting Ya'qub governorships he already held and publicly cursing him, depending on how close he was. In 876 Ya'qub marched on Iraq itself and was stopped at Dayr al-Aqul, fifty miles from Baghdad, by the caliph's brother al-Muwaffaq. He withdrew and died three years later, still, formally, a rebel.

He had held Khurasan for six years without ever holding a valid title to it, and had come within fifty miles of taking the institution that issued titles. His brother, in the next chapter, tried the opposite experiment.

The brother who took the paper

Drafted from Bosworth, The History of the Saffarids of Sistan and the Maliks of Nimruz (247/861 to 949/1542-3) (1994)peak

Ya'qub died in 879 having never accepted an appointment from anyone. His brother Amr, who succeeded him, accepted one within months, and held the Saffarid state for twenty-two years — longer than Ya'qub held it, over more ground, with less trouble.

The reversal is the whole interest of the reign. Amr took investiture as governor of Khurasan, had his name and the caliph's on the coinage in the correct order, sent the tribute, and received in return the thing Ya'qub had been offered a sword and a plate of onions to refuse. It worked. Baghdad, which had cursed Ya'qub from the pulpit, dealt with Amr as a legitimate governor for two decades, and the province was quieter under him than it had been under either of them.

He also governed, which Ya'qub had largely not had time to. The sources describe an administration in Sistan and Khurasan with a functioning fisc, attention to the irrigation the Helmand basin lives on, and a reputation for paying troops on time, which is worth more than it sounds: a century later the Samanid state would be brought down by its own commanders, and the immediate cause was arrears.

Then he made one mistake and it was terminal. In 900 he moved north against the Samanids, who held Transoxiana and had just been granted Khurasan by a caliph who had granted it to Amr first. He was beaten at Balkh, captured, and sent to Baghdad, where he was executed in 902.

What the caliph did next is the sentence the Saffarid page exists for. Al-Mu'tadid transferred the governorship of Khurasan to Isma'il the Samanid in the same year, on the same terms, with the same wording. The investiture Amr had reorganised his whole political life to obtain turned out to be transferable to whoever won the next battle, which is exactly what his brother had said it was worth.

So the dynasty ends up claiming its right to rule on two grounds at once, and almost nobody else does. Conquest, because Ya'qub took his province and told the caliph afterwards; and caliphal investiture, because Amr asked for it properly and got it. The two brothers ran the same sequence in opposite directions, and the province ended up in the same place either way.

A hundred thousand swords

Drafted from Khallikan, Ibn Khallikan's Biographical Dictionary (Wafayat al-A'yan) (1843)contraction

Amr asked for Transoxiana. It was already held, competently and quietly, by Isma'il ibn Ahmad the Samanid, and the caliph's grant of it was therefore a licence to go and take it rather than a province delivered. Amr understood this perfectly well at the moment he received the document.

Ibn Khallikan preserves the scene, from a historian of Khurasan writing not long after. The envoy came to Nishapur with the diploma and a set of presents Amr had already refused once, among them seven robes of honour. He put them on Amr one at a time, and at each robe Amr prostrated himself and prayed two prostrations. Then the diploma was handed over and Amr asked what it was. This, said the envoy, is the thing you asked for. Amr answered: what use is it to me? Isma'il ibn Ahmad will not give up that province unless I enforce the demand with a hundred thousand swords. You asked for it yourself, said the envoy; now get ready to go and govern in his place. Amr took the diploma, kissed it, set it down, sent the embassy seven hundred thousand dirhams, and began raising an army.

Isma'il wrote to him before it came to anything. You have been given the government of a vast country and I hold Transoxiana; be content with what you have and leave me this frontier province. Amr refused. Someone raised the Oxus and how hard it would be to cross, and he said that if he chose to dam it with bags of money and walk over on them he could.

The first army he sent did not reach the river. Isma'il crossed it instead and destroyed the force Amr had given to Muhammad ibn Bishr, in October 899, while Amr himself sat at Nishapur. There is a story attached that is really a comment: on the eve of the battle a Sistani officer found Ibn Bishr having his head shaved and asked whether he had got Isma'il's permission to do it, meaning that by fighting Isma'il at all he had already made the man master of his head. Ibn Bishr cursed him out of the tent. The next morning he was routed, taken, and beheaded, and Isma'il — who had to ask the prisoners to identify which head was the commander's — heard the remark repeated and was unsettled by it.

Amr went north himself the following spring and posted himself at Balkh. Isma'il called in every landholder under his authority, crossed to the western bank, and occupied the country round about instead of offering battle; Amr, finding himself half blockaded, is said to have regretted the whole enterprise and asked for a truce, which was refused. Almost no serious fighting took place. On the morning of the battle Amr learned that his own secretary had gone over to Isma'il in the night, taking a general and a large body of troops with him, and that most of his partisans had followed. He lost his nerve and rode. Told there was a short cut through a jungle, he sent the main body by the highway and went into the wood with a small escort; his horse got into marshy ground and fell, his companions rode on without noticing, and Isma'il's men came up and found him sitting on a horse that could not move.

The authorities do not agree on the day — one gives 20 March 900, another 21 April — and Ibn Khallikan reports both rather than choose. On the rest they agree. He was put in irons and taken to Samarkand, and Isma'il let him choose between staying there as his prisoner and being sent to the Commander of the Faithful. He chose Baghdad, and reached it on 23 April 901, ahead of the robes, the standard and the diadem that were travelling the other way to Isma'il.

Why do you recite what I cannot understand

Drafted from Bosworth, The History of the Saffarids of Sistan and the Maliks of Nimruz (247/861 to 949/1542-3) (1994)aside?What aside means

Ya'qub took Herat in 867, and his poets did what poets did: composed an ode in Arabic celebrating the conquest and recited it to him. He is reported to have asked why he was being read something he could not understand, and his secretary, Muhammad ibn Wasif of Sistan, went away and produced the verses in Persian instead.

That poem is conventionally the first New Persian court qasida — the beginning, in the tradition's own account of itself, of a literature that runs through Rudaki, Firdawsi, Nizami, Hafiz and everything after.

Take the anecdote at the level it deserves. It is almost certainly tidied. What it records accurately is a condition: a ruler at the head of a Persian-speaking army from a frontier province, with no Arabic education and no interest in acquiring one, at whose court Arabic was therefore useless. Persian court poetry began because the patron could not follow the alternative.

Set the three holders of Khurasan side by side and the sequence is not a progress. The Tahirids were Persian governors who cultivated Arabic assiduously, because their position depended on being the caliph's most reliable servants. The Saffarids had no such position to protect and no such education, and got the vernacular by default. The Samanids, a century later, turned it into a programme — commissioning, paying, building a library. What looks in retrospect like a Persian revival unfolding in three stages was one dynasty avoiding a language it needed, one dynasty unable to use it, and one dynasty deciding it was worth money.

The Saffarids also have something most dynasties of this period do not: a local chronicle. The Tarikh-i Sistan is an anonymous history of the province, written in Persian by somebody from it, and it is where the sword and the onions come from.

Every other surviving account of Ya'qub was written in Iraq, by men who regarded him as a bandit who had frightened the caliph. The Sistani version is not neutral either — it is proud of him. But two partial accounts from opposite directions are a great deal more than most rulers of the ninth century get, and almost everything interesting anyone can say about Ya'qub comes from having both.

A hundred years in Sistan

Drafted from Bosworth, The History of the Saffarids of Sistan and the Maliks of Nimruz (247/861 to 949/1542-3) (1994)end

After 901 the Saffarids stopped being a claimant to Khurasan and went back to being what they had started as: the power in Sistan. They stayed that, on and off, for another century, under Samanid and then Ghaznavid overlordship, as the maliks of Nimruz.

Two endings, both real

That is why the dynasty has two ending dates instead of one, and why they are a hundred and two years apart.

The imperial Saffarid state — the thing that took Nishapur, frightened Baghdad and marched on Iraq — was finished at Balkh in 900 and buried when the caliph handed Khurasan to the Samanids. Nothing after that has any resemblance to it. The family, on the other hand, went on governing the Helmand basin from Zaranj until Mahmud of Ghazna deposed Khalaf ibn Ahmad in 1003.

Both statements are true about different objects, and it is not obvious that either has a better claim to be the ending. The first is the end of a state; the second is the end of a family holding a province.

What that says about how a dynasty stops

The Saffarids are unusual in having the distinction so cleanly separated, but they are not unusual in having it.

Very few dynasties stop the way a battle stops. Most of them stop being one thing and go on being a smaller thing, sometimes for a very long time — a sovereign becomes a governor, a governor becomes a landholder, a landholder becomes a family that used to matter. What makes the case here legible is the gap: an empire that lasted forty years, followed by a hundred years of the same family in the same province, doing something that would never have been worth recording if the forty years had not happened.

Ya'qub's whole career was an argument that a man with an army needs nobody's permission. His descendants held Sistan for a century on somebody else's permission, which is the answer his brother had given him, and which was correct.

The moving sands

Drafted from Strange, The Lands of the Eastern Caliphate: Mesopotamia, Persia, and Central Asia, from the Moslem Conquest to the Time of Timur (1905)afterlife

The capital outlived the dynasty by nearly four hundred years, and the geographers who describe it are describing a place unlike anywhere else they went.

A city built against the wind

Zaranj sat at the edge of the Helmand delta with salt marshes and date palms on one side and desert on the other, and the wind there never stopped. It blew continually and violently, moving the sand about in a way the Arabic writers found genuinely dangerous — whole villages overwhelmed, cultivated districts buried. Ibn Hawkal gives a long account of how in the 970s the Great Mosque of Zaranj was choked with sand.

What the inhabitants did about the wind is the detail worth keeping. They used it. Sistan had windmills, which Le Strange notes were a feature peculiar to that country: the same ceaseless wind that was burying their fields turned their grain. Later a great dyke was thrown up to hold the sand off the city, which Mustawfi says was originally built by an ancient king and restored by another, though the names he gives belong to legend rather than to a reign.

Muqaddasi, in the tenth century, adds the riches and the learning of the inhabitants, the strongly fortified castle, and the two famous minarets of the Great Mosque.

One of the two had been built by Ya'qub the Coppersmith.

Which is a strange thing to be left

Consider what that means about the sources. Ya'qub is remembered in the Iraqi chronicles as a brigand who terrified the caliphate, and in his own province's history as a hero; and the item of his that a Palestinian geographer thought worth recording a century after his death is a minaret on a mosque.

He put it up in a city he had made the capital of an empire, and it stood there through the Samanid overlordship and the Ghaznavid one and the Seljuks, and it was still standing when Chinggis Khan sent his armies into Sistan in 1222 and, unusually, the capital came through more or less intact — under a Mongol governor afterwards, and described by Mustawfi in the fourteenth century as still very flourishing, with excellent fruit in gardens watered from a canal off the Helmand.

Timur

In 1383 Timur came to Sistan and took the fortress of Zarah. Then he came to the city, which by then was simply called Shahr-i Sistan.

It closed its gates and refused to surrender. After a short siege it was taken by storm, every inhabitant who could be found was killed, the walls were razed and the houses destroyed.

Le Strange's sentence is the last word on it: since that time Zaranj has come to be a nameless ruin.

So the sequence runs: a coppersmith from a frontier province takes an empire and loses it, his family holds the province for another century, the city he built in outlasts them by four hundred years and survives the Mongols, and then a different conqueror arrives and there is nothing on the map at all. The wind is presumably still blowing.