Samanid Empire

Four brothers and a governorship

Drafted from Bosworth, The New Islamic Dynasties: A Chronological and Genealogical Manual (1996)formation

The Samanids do not begin with a conquest. They begin with an administrative decision taken in Baghdad.

In 819 the caliph al-Ma'mun, having won a civil war largely on Khurasani strength, distributed governorships in Transoxiana to four brothers — the grandsons of a Balkh landowner remembered as Saman Khuda. Nuh took Samarkand, Ahmad took Farghana, Yahya took Shash, Ilyas took Herat. They were not kings. They were appointees of an appointee, holding provinces under the Tahirid governors of Khurasan, who in turn held Khurasan under the caliph.

What matters about this arrangement is the shape of it. The Samanids entered history already inside a chain of delegated authority, and every subsequent step of their rise was taken without ever formally leaving that chain. There is no Samanid declaration of independence, no year in which they renounce Baghdad. There is only a slow accumulation of the things independence is made of, while the language of subordination stays in place.

That is worth sitting with, because it recurs. Nearly every polity in this thread acquires power the same way: not by breaking a link but by quietly becoming heavier than the thing it is attached to.

By the 860s the Tahirid grip on Khurasan was failing, and the family's Samarkand line had absorbed the others. Nasr I held Transoxiana as a unit. The paperwork still said governor.

Isma'il moves the capital

Drafted from Bosworth, Encyclopaedia Iranica · sourceexpansion

The brothers' arrangement ended the way such arrangements usually do, between two members of the family. Nasr I ruled from Samarkand; his brother Isma'il held Bukhara for him. In 888 they fought, and Isma'il won.

Isma'il I is the point at which the Samanids stop being a provincial family and start being a state, and the clearest evidence of it is the move of the centre of gravity to Bukhara. Samarkand was the older and grander city. Bukhara was Isma'il's, and from 892 it was the capital — a working administrative centre rather than an inherited seat.

In 900 he faced the Saffarids, the other power that had grown out of the Tahirid collapse, and he beat them at Balkh. Amr ibn al-Layth was captured and sent to Baghdad. The caliph, who had recently been obliged to recognise Amr as governor of Khurasan, now transferred the recognition to Isma'il.

This is the moment the thread doubles back on itself. The Saffarids had taken Khurasan from the Tahirids by force and extracted an unwilling investiture for it. The Samanids took it from the Saffarids by force and received a willing one. The caliphal grant did not create either state and could not have prevented either. It followed the outcome and dressed it.

What Isma'il had by 900 was Transoxiana and Khurasan under one administration, a capital he had built up himself, and a piece of paper from Baghdad that said he had been given all of it.

Nasr II's Bukhara

Drafted from Bosworth, Encyclopaedia Iranica · sourcepeak

Under Nasr II, from 914 to 943, Bukhara did the thing that makes the Samanids matter more than their territory suggests.

The court patronised poetry in New Persian and kept its chancery working in both Persian and Arabic. Rudaki wrote there. Bal'ami produced a Persian version of Tabari's history, which is to say that the standard historical work of Islamic scholarship was made available to readers who did not read Arabic, at a Samanid court, on Samanid initiative. The language that would be the literary and administrative language of Iran, Central Asia and eventually much of northern India for the next thousand years was assembled into a working instrument here.

It is worth being exact about what this is and is not. The Samanids did not invent a script; they wrote Persian in the Perso-Arabic alphabet they had inherited, unchanged. So they left no descendant alphabet, which is the honest answer and sits oddly beside the size of what they actually did. Inventing a script and starting a literature are different achievements, and only one of them is easy to count.

The other export was religious. Bukhara sat on the frontier with the Turkish steppe, and Samanid frontier practice combined raiding, trade and missionary activity in a way that carried Sunni Islam north and east past the limits of Samanid administration. The Turkish groups that adopted it did not become Samanid subjects. Some of them, within two generations, would be the people who ended the dynasty.

The trade the whole system ran on

Drafted from (ed.), The Cambridge History of Iran, Volume 4: From the Arab Invasion to the Saljuqs (1975)peak

The Samanid state's principal export was people.

Transoxiana sat on the frontier between the settled Islamic world and the Turkish steppe, and the Samanids controlled the crossing. Turkish boys and young men — taken in steppe warfare, sold by their own or by neighbouring tribes, or acquired as tribute — were brought south into Samanid territory, where the trade was regulated and taxed: a licence for each slave leaving the realm, a duty collected at the frontier, and a market at Bukhara and Samarkand that supplied the entire Islamic world. A tenth-century geographer describes Turkish slaves arriving in such numbers that the price fell. It was, on the accounts, the largest single source of the state's income.

Three things follow, and they run a long way past this dynasty.

The first is that the Samanid slave army was not an innovation. It was a domestic application of an export business: the dynasty was selling ghulams to Baghdad, to Egypt and to the west, and buying some for itself was the cheapest option available and the one it understood best. The whole institution of the Turkish slave soldier, which the caliph al-Mu'tasim had built into the centre of the Abbasid state in the 830s, was fed for the following two centuries largely from this market.

The second is that the Samanids were arming the frontier they depended on. Turkish military skill entered the Islamic world through a market the Samanids ran, and the political consequence arrived on schedule: a Samanid slave-general took Ghazna in 962 and founded the Ghaznavids, and the Karakhanids who took Bukhara in 999 were the Karluk confederation from the same steppe, converted and organised. The two polities that ended this one both came out of the frontier its economy was built on.

The third is a caution about the map. Two and a half million square kilometres describes territory, and territory is not what the state was for. What it was for was to sit astride a set of routes — silver flowing north to the Baltic in quantities that Scandinavian coin hoards still record, slaves and furs flowing south — and to tax what crossed. A state like that is better described by a line than by an area, and drawing it as a shaded region on a map gets the emphasis exactly backwards.

What Bukhara was

Drafted from Frye, The Centers of Civilization Series (1965)peak

The Samanid court made New Persian a language of administration and literature. That is the largest thing this dynasty did and it is worth being exact about what it means.

Persian had never stopped being spoken. What it had stopped being, after the Arab conquest, was written: Middle Persian went out of official use, and for two centuries the language of administration, scholarship and poetry across Iran was Arabic. What happened at Bukhara in the tenth century was that a court began commissioning serious work in the vernacular, in the Arabic script, and paying for it.

Rudaki, who died in 941, is the figure the tradition puts first — a poet at Nasr II's court credited with working every major form of later Persian verse into shape, and remembered as the moment the language acquired a literature rather than a folklore. Bal'ami, the vizier, produced a Persian version of al-Tabari's enormous Arabic history, which put the standard account of the Islamic past into a language a Khurasani could read without Arabic. Daqiqi began versifying the Iranian national epic and was murdered before he got far; the poet who finished the Shahnama took his thousand surviving lines into it and presented the completed work to Mahmud of Ghazna, who by then ruled the country the Samanids had ruled.

Bukhara itself is worth the chapter. It was a city on an oasis fed by the Zarafshan, with an agrarian hinterland that supported it and a caravan trade that made it rich, and under the Samanids it acquired the apparatus of a capital: a bureaucracy of ten specialised divans, a library that was famous enough to be described by the people who used it, and a scholarly establishment that drew men from across the east. Ibn Sina — Avicenna — grew up outside it, was given access to the Samanid royal library as a teenager after treating a member of the family, and said afterwards that he found books there that nobody else had.

Set that beside what the dynasty is measured by anywhere else — square kilometres, years, a list of conquests — and the disproportion is the point. The Samanids held Khurasan and Transoxiana for a hundred and eighty years, which is respectable and not remarkable. What they are actually remembered for is that a language which had been out of written use for two centuries came back at their court, in their pay, and is still being written.

The year the army chose the amir

Drafted from (ed.), The Cambridge History of Iran, Volume 4: From the Arab Invasion to the Saljuqs (1975)contraction

Nasr II reigned for twenty-nine years and is the height of the dynasty by every ordinary measure. He was deposed by his own son in 943, at the demand of his own officers, over religion.

Isma'ili missionaries had been working in Khurasan and Transoxiana for some time — the Isma'ili da'wa, the same missionary apparatus that had just put a Fatimid caliph on a throne in Ifriqiya, operating exactly as that chapter says it did, inside a rival caliph's territory. They reached the Samanid court. How far Nasr went is the disputed part: the hostile tradition says he converted, sympathetic accounts say he protected and listened to them, and the sources are late enough and partisan enough that the distinction may not be recoverable.

What is not disputed is what the army did. The Turkish commanders and the Sunni religious establishment of Bukhara regarded an Isma'ili amir as intolerable, conspired to kill him, and were forestalled by his son Nuh, who took his father's abdication instead and then had the missionaries and their supporters at court executed.

Two things are settled in that transaction and both are bad for the dynasty.

The first is doctrinal and it held: after 943 the Samanid state was firmly, publicly Sunni and Hanafi, tied to an ulama in Bukhara that had just demonstrated it could veto a ruler. That is the same settlement the Abbasid page describes arriving a century earlier through the mihna, reached here in one season rather than fifteen years.

The second is the one that ends the polity. In 943 the guard discovered that it could remove an amir and install another, and be obeyed. Everything in the two chapters that follow — the slave-general who walked away with Ghazna in 962, the succession disputes the commanders arbitrated, the officers who simply did not turn out in 999 — is a consequence of a lesson the army learned here.

The instrument that walked away

Drafted from Bosworth, The Ghaznavids: Their Empire in Afghanistan and Eastern Iran, 994-1040 (1963)contraction

Samanid armies rested on Turkish military slaves, ghulams, bought young on the steppe frontier, converted, trained, and promoted into command. The system produced excellent soldiers with no local landholding, no tribal following and no claim of their own. Their standing was entirely a function of the amir's favour, which was precisely the point.

It worked until the favour became the only thing holding it together.

In 962 a Samanid Turkic slave-general named Alptigin, having backed the wrong candidate in a succession dispute at Bukhara, withdrew east and took Ghazna, a town on the Afghan side of the frontier. He did not rebel in any declarative sense. He held Ghazna, in the ordinary way a commander held a place, and the line he founded there outlasted the masters who had bought him.

Read that alongside the first chapter and the symmetry is exact. The Samanids became a state by being appointees who grew heavier than their appointers. Then their own appointee did it to them. The mechanism did not change; only the direction of view.

Through the 970s and 980s the amirs at Bukhara lost the ability to arbitrate between their own commanders. Real power moved to whichever general held the capital that season. By the time the external blow came, there was not much left inside to resist it.

999, and the six years after

Drafted from Bosworth, Encyclopaedia Iranica · sourceend

In 999 the Karakhanids took Bukhara. There was very little fighting. The Samanid army did not concentrate, the city did not hold out, and the last effective amir was taken.

South of the Oxus, Mahmud of Ghazna — the grandson-in-law of Alptigin's line, governing Khurasan in Samanid name — did not march north to save his nominal overlords. He kept Khurasan. The Karakhanids kept Transoxiana. The Oxus, which had been an internal river of the Samanid state, became the border between two successors, and stayed one for a long time.

Then there is Isma'il al-Muntasir, and he is the reason the dynasty has two ending dates rather than one. A Samanid prince who escaped Karakhanid custody, he spent from 1000 to 1005 raising forces, retaking cities briefly, losing them, crossing and recrossing the Oxus, and being defeated by Karakhanids and Ghaznavids in turn. He was killed in 1005 by Arab tribesmen he had taken shelter with.

Whether the dynasty ends in 999 or in 1005 depends on whether you are dating the end of the state or the end of the family's attempt to be one. The sources differ, and both dates are worth keeping. Six years of a man crossing and recrossing a river to retake cities he could not hold is neither a state nor nothing.

What passed forward was not the territory. It was the administrative Persian, the chancery practice, the court model and the slave-soldier system — all of it inherited more or less intact by Ghazna, which is where the thread goes next.

Nothing is left save Rudaki's lays

Drafted from Browne, A Literary History of Persia, Volume I: From the Earliest Times until Firdawsi (1902)afterlife

Not very long after the dynasty ended, a Persian poet of Gurgan named Sharif-i-Mujallidi wrote a couplet about what had survived of it. Browne quotes it:

From all the treasures hoarded by the Houses Of Sasan and of Saman, in our days Nothing survives except the song of Barbad, Nothing is left save Rudaki's sweet lays.

Two dynasties, four centuries apart, each of which governed most of Iran; and the only thing the poet can find left of either is the music they paid for. Barbad was the harper of Khusraw Parviz. Rudaki was the harper of Nasr II.

The song about the road home

The story that made Rudaki famous is the one everybody in the tradition tells, and the earliest version is Nizami-i-Arudi's.

Nasr II had gone to Herat for the summer and stayed four years. His army wanted to go home to Bukhara and could not say so. They promised Rudaki five thousand dinars if he could get the amir to move, and he wrote them a ballad — no rhetoric in it at all, which later critics found so plain that Dawlatshah records his surprise that words so simple could have such an effect, and adds that anyone reciting such a thing at court in his day would have been laughed at.

The Ju-yi Muliyan we call to mind, We long for those dear friends long left behind. The sands of Oxus, toilsome though they be, Beneath my feet were soft as silk to me. Glad at the friends' return, the Oxus deep Up to our girths in laughing waves shall leap. Long live Bukhara! Be thou of good cheer! Joyous towards thee hasteth our Amir!

When Rudaki reached that verse, the account says, the amir got down off his throne, took the horse of the sentry on duty, and rode for Bukhara without his boots — they were carried after him for two parasangs before he stopped to put them on, and he did not draw rein again until he was home. The army paid Rudaki ten thousand.

What kind of achievement this is

It is worth being exact about it, because "founded Persian literature" is the sort of phrase that slides past.

Rudaki was not, on this evidence, a man of letters in the later sense. He was a harper and a singer and an improviser, closer to the balladeers who still perform at Persian entertainments than to the elaborate court poets of three centuries afterwards — which is precisely why the tradition compares him to Barbad, who had done the same job for a Sasanian king, and who was remembered for being able to tell Khusraw Parviz things nobody else dared say by putting them in a song.

So the Samanid contribution was not the invention of a literary language by scholars. It was a court that would pay a man to sing in Persian, in front of everybody, about wanting to go home — and would go on paying, until a generation later a vizier was rendering Tabari into it and a poet at Tus was beginning the Shahnama.

The measure

The dynasty held Khurasan and Transoxiana for a hundred and eighty years, governed well, ran the largest slave market in Asia, and was ended by men its own trade had armed.

Within a lifetime of that, a poet looking for what remained of it could not find the territory, or the army, or the treasury worth mentioning, and named a song. He was right, and the song is still known.