A Sogdian envoy sold Byzantium a road around Persia in 568 (and a war)

Before: an empire that bought silk through its enemy

In 565, the year Justinian died and his nephew Justin II took the throne, the Eastern Roman empire wore silk it could not make in quantity and could not buy except through Persia. The Fifty-Year Peace negotiated at Dara in 562 had fixed the terms: commerce between the two empires was confined to designated frontier towns, Nisibis and Dara among them, where Persian customs officers took their cut, and the Romans paid Khosrow I an annual subsidy of 30,000 gold solidi for the privilege of peace.¹²⁰ The silkworm eggs that monks had carried to Justinian around 552 had produced imperial workshops, not independence; Byzantine sericulture stayed small and technically inferior for centuries, and the empire went on importing raw silk from the east in quantity. The 540s wars had taught Constantinople what an embargo felt like: Justinian had tried to cap the price of raw silk at fifteen solidi a pound and had tried to route purchases through the Aksumite kingdom on the Red Sea, and both schemes had failed because Persian buyers reached the Indian ports first.²⁰²¹

So the problem Justin inherited was not ignorance of where silk came from. It was geography. Every land route from the Sogdian and Chinese sources ran through the Sasanian realm, and the only known alternative, the sea route to India, delivered the cargo into Persian hands at the ports. The receiving culture of this record is therefore not a people without silk but a court that had spent a generation trying to buy it from someone other than its principal enemy, and had not yet found that someone.

What Constantinople knew of the steppe east of the Don

Byzantium's picture of the lands beyond the Caucasus was old and thin. The Avars had appeared on the Roman frontier in 558, fugitives who, in Evagrius's phrase, "had fled en masse from their neighbours the Türks after being ill-treated by them", and Justinian had taken them into service against the Huns of the Pontic steppe.⁵¹⁹ The word "Türk" itself entered Greek through them. Theophanes of Byzantium, writing at the end of the sixth century, still had to gloss it: "The Turks, formerly called Massagetae, and by the Persians Kirmikhiones, who live to the east of the Tanais".³ The Massagetae were Herodotus's people, dead a thousand years; the label tells us the court had no vocabulary of its own for a steppe empire that had, in the previous decade, destroyed the Rouran in Mongolia, crushed the Hephthalites in Central Asia and planted its western wing on the Volga.

There had been one contact. In July 563, Theophanes Confessor records, "envoys arrived from Askel, king of the Hermichiones, who dwell inland of the barbarian nation near the Ocean".⁸ Theophanes of Byzantium adds the substance: the Türks sent gifts to the emperor "beseeching him not to receive the Avars", and the emperor, "owing to the agreement he had made with the Turks", later refused the Avars' request to settle in Pannonia.³ C. A. Macartney doubted in 1944 that these Kermichiones were Türks at all; Denis Sinor, reviewing the evidence half a century later, concluded that they were, and that "Askel is the original form of the name of the first tribe of the confederation called by the Chinese Nu-shih-pi", the westernmost grouping of the Western Türks.²⁵¹⁶ Li Qiang and Stefanos Kordosis follow Sinor: the 563 embassy was a Türk embassy, sent by a federated tribe while the qaghan was busy with the Hephthalite war, and it set the first term of every later negotiation. The Avars were the Türks' runaway slaves, and Byzantium was not to deal with them.¹⁹

What the court did not have in 563 was a road. No Roman embassy had ever gone east of the Caucasus to the steppe. John of Ephesus, a contemporary, says so plainly of Zemarchus's mission: "never before had a Roman embassy been sent to these numerous and powerful tribes".⁴ The receiving culture had an ally's name and a grievance in common; it had no map.

The Sogdians: a merchant people who had just changed masters

The senders had a map and no state. The Sogdians were the Iranian-speaking townspeople of the Zarafshan valley, Samarkand, Bukhara and Panjikent, and the diaspora that had grown out of those towns along the caravan roads. The Sogdian "Ancient Letters" found near Dunhuang, written around 313, already show a commercial network with agents in the Gansu corridor writing home to Samarkand about prices and the sack of Luoyang.¹¹ By the sixth century that network reached from the Chinese capitals to the Caspian. Étienne de la Vaissière's reconstruction of it, the standard account, describes a people whose political history consisted of being ruled by others, the Hephthalites from about 479, and whose economic history consisted of ruling the road.¹¹

Between about 557 and 563 the Hephthalite kingdom was destroyed by a coalition of Sasanian Iran and the Türk qaghanate; the two powers divided the spoils at the Oxus, and Sogdiana north of the river passed to the Türks.¹⁵¹⁷ The Türks did what the Hephthalites had done before them: they used the Sogdians. Within a generation Sogdians were serving the qaghans as scribes, ministers and ambassadors. The earliest monument of the Türk empire, the Bugut stele in Mongolia erected around 584 for the qaghan Taspar, is written in Sogdian, not Turkic.²⁴ The tomb of the Sogdian merchant An Jia at Xi'an, dated 579, shows him feasting with a Türk chieftain.¹¹ And the Zhou shu records that when the Western Wei court first reached out to the Türks in 545, the envoy it sent was a Sogdian, An Nuopantuo.¹⁴

Here is the calibration the reader needs. The Sogdians of the 560s were a trading diaspora that had just acquired the most powerful patron in Asia, and their first idea for what to do with him was commercial. They wanted to sell raw silk inside Persia. What they got instead, and what they then carried to Byzantium, is this record.

The transmission: silk burned, envoys poisoned, a letter in Scythian

Ctesiphon: Katulph's advice and the fire

The source is Menander Protector, a court historian of Maurice's reign who had access to the diplomatic archives and whose account of these embassies survives in the Constantinian excerpts on embassies.¹ His fragment 10.1 opens with the Sogdians asking Sizabul, the Türk ruler whom the Orkhon inscriptions call Istämi, "to send an embassy to the Persians, to request that the Sogdians be allowed to travel there and sell raw silk to the Medes". Sizabul agreed; the embassy was Sogdian, "whose leader was Maniakh".¹

Khosrow stalled. Menander's narrative of the Persian court is specific enough to be an insider's: the king "put off his reply until the next day and kept postponing it", because free passage for Sogdian traders meant free passage for Türk agents into Persia. When the envoys pressed, he summoned a council, and the decisive voice was a renegade. Katulph was a Hephthalite who, "because the king had raped his wife, had betrayed his own tribe to the Turks", and who had then changed sides again and "joined the Medes". His advice is preserved verbatim in Menander's Greek and in Henry Yule's 1866 translation: the king should "on no account let the silk have free passage, but have a price put upon it, buy it up, and have it burnt in the presence of the ambassadors. It would thus be seen that though he would do no injustice, he would have nothing to do with the silk of the Turks."² So, Menander says, "the silk was burned, and the Sogdians returned to their homeland not at all pleased with what had happened".¹

The burning was not a tantrum. It was a legal act, paid for, witnessed, and designed, as Katulph said, so that no injustice could be alleged. Its message was the Sasanian silk policy stated in a single gesture: Persia would not become the retail outlet of the Türk qaghanate, and it would rather destroy the goods than let the sellers in.

The second embassy and the poison

Sizabul tried once more, and this time sent Türks. Khosrow, "after discussion with the Persian high officials and with Katulph, decided that because of the untrustworthy nature of the Scythians it was completely against Persian interest to establish friendly relations with the Turks". The decision was to make sure no third embassy came. "The majority of the Turkish envoys, all but three or four, were murdered by a deadly poison mixed in with their food."¹ The cover story, circulated among the Persians, was climate: the Türks had died "of the suffocating dry heat of the Persian climate; for their own country was subject to frequent falls of snow, and they could not exist except in a cold climate".² The survivors reached home and reported both the deaths and the rumour. Sizabul, "a shrewd and intelligent man, recognised what had been done and realised the truth". Menander's sentence is flat: "This was the cause of the hostility between the Persians and the Turks."¹

Chavannes, who in 1903 set the Greek account beside the Chinese, summarised the sequence in a sentence that has not been improved on: "Dizaboul (Istâmi) ne se tint pas pour battu et envoya de nouveaux émissaires; ils périrent empoisonnés; trois ou quatre seulement d'entre eux en réchappèrent."¹⁰ Istämi did not accept defeat; the second embassy was poisoned; three or four survived. The costs of the transmission begin here, before any Roman is involved: a cargo of silk in the fire and most of a diplomatic party in the ground.

Maniakh's proposal and the road over the Caucasus

The proposal that turned a Sogdian trade mission into a Byzantine alliance was Maniakh's own. "Maniakh, the leader of the Sogdians, took this opportunity and advised Sizabul that it would be better for the Turks to cultivate the friendship of the Romans and send their raw silk for sale to them because they made more use of it than other people." He offered to go himself. Sizabul "consented to this proposal and sent Maniakh and some others as envoys to the Roman Emperor carrying greetings, a valuable gift of raw silk and a letter".¹

Menander's description of the journey is the earliest Roman description of the northern road: Maniakh "travelled very many roads and traversed very many lands, over huge mountains reaching near the clouds, through plains and woods, over marshes and rivers. Then he crossed the Caucasus and finally came to Byzantium."¹ The embassy arrived "at the beginning of the fourth year of Justin's reign", that is, late 568.¹ The letter was "written in Scythian", which Blockley takes to be Sogdian script, the chancery language of the qaghanate, read to the emperor "through an interpreter".¹

The audience is the one moment in the record where the receiving culture can be watched learning. Justin questioned the envoys "about the leadership of the Turks and their location". They answered that there were four principalities but that "power over the whole people was vested in Sizabul alone"; that the Türks "had conquered the Ephthalites and made them tributary"; that the Hephthalites "lives in cities", which the emperor took to mean that the Türks now held those cities. Then the question that mattered to Constantinople: how many Avars had escaped Türk rule? "Those who fled number, I think, around twenty thousand." The envoys "enumerated the tribes subject to the Turks and asked the Emperor for peace and an offensive and defensive alliance", adding that they were "very willing to crush those enemies of the Roman state who were pressing upon their territory". Then the oath: "Maniakh and those with him raised their hands on high and swore upon their greatest oath that they were saying these things with honest intent", calling down curses "upon themselves, even upon Sizabul and upon their whole race, if their claims were false".¹²

Theophanes of Byzantium adds a detail Menander omits, and it belongs to this embassy. "When king Justin afterwards showed the Turks how the worms were bred and how silk was made, he greatly surprised them, since at that time they had possession of the markets and harbours of the Seres, formerly held by the Persians."³ The Sogdians had come to sell raw silk to the one customer who, they believed, could not make it. The emperor's reply was to show them the worms. Li and Kordosis read the gesture as policy: "by demonstrating to the Türks that the Byzantines knew how to produce silk, Byzantium aimed to stall the Sogdians' plans of extending silk trade in Byzantium".¹⁹ The alliance was accepted; the commercial proposal that had carried it across a continent was, in the same audience, quietly declined.

Zemarchus goes east: August 569

Justin "decided to send an envoy to the Turks" and chose Zemarchus the Cilician, "who was at the time general in command of the eastern cities", the magister militum per Orientem.¹ The rank is the measure of the decision: the empire sent its eastern field commander, not a notary. Menander dates the departure with unusual care, "near to the end of the fourth year of Justin's reign, in the second year of the fifteen-year cycle and around the beginning of the Latin month of August", that is, August 569, and adds that "Zemarchus set out from Byzantium with Maniakh himself and his companions".¹ John of Ephesus, writing from Syriac memory, puts the mission in Justin's seventh year and says the party was away two years; the return in 571 is consistent with both.⁴

Mihály Dobrovits, whose 2011 study is the most detailed modern reconstruction of the journey, treats it as a problem in communication: how a Roman embassy with no precedent and no shared language managed a steppe court.¹⁸ Menander's answer begins with ritual. On entering Sogdian territory the party was met by Türks who "offered them iron for sale", a demonstration, Menander guessed, "to imply that they possessed land that contained iron"; and then by "exorcisers of ill-omened things", who heaped the baggage on the ground, set fire to frankincense boughs, chanted "in their Scythian tongue, making noise with bells and drums", and "led Zemarchus himself through the fire".¹ The Roman general was purified like any other guest of the qaghan.

They reached Sizabul "on a mountain called Ektag, or 'Golden Mountain' in Greek", in a valley where the ruler sat "in a tent, sitting upon a golden throne with two wheels, which could be drawn when necessary by one horse".¹ Where Ektag was is unresolved. Bury put it in the Altai, whose name means golden; Chavannes objected that Ektag is Ak-tagh, White Mountain, and placed the camp far to the south in the Tianshan, by the river Tekes; Blockley, surveying the dispute in 1985, declined to decide it.¹⁰¹ The record does the same. What is not in doubt is what Zemarchus saw: three days of feasts in pavilions hung "with multicoloured silken hangings", a couch "made completely of gold", "golden urns, water-sprinklers and also golden pitchers", a throne "supported by four golden peacocks", and wagons of silver "in no way inferior to those which we make; so wealthy is the ruler of the Turks".¹ The Sogdians had told Justin their master was rich. Zemarchus could now say so from sight.

Unknown Sogdian painters. Hall of the Ambassadors, Afrasiab (Samarkand), 7th century CE. Afrasiab Museum, Samarkand. Photograph: Northeast Asian History Foundation. Public domain (PD-Art) via Wikimedia Commons. · Public domain

Then the political theatre. Sizabul took Zemarchus and twenty attendants with him on a march against Persia, gave him "a female slave, a war-captive from the people called Kherkhir", and at a camp called Talas received a Persian envoy. He "invited the Roman and the Persian ambassadors to dine with him", seated the Roman in the place of honour, "heaped great reproaches on the Persians, telling the injuries he had received at their hands", and let the Persian lose his temper in front of the Roman.¹² John of Ephesus preserves a Syriac version of the same dinner in which the qaghan asks Zemarchus whether it is true, as the Persians say, "that the king of the Romans is their slave, and pays yearly tribute as a slave", and Zemarchus answers with Trajan's statue, before which, he says, no Persian dares pass on horseback; the Persians admit the statue exists, the qaghan dismisses them as liars, and Khosrow, hearing of it, "commanded that Trajan's statue should be overturned".⁴ Whatever the historical weight of the anecdote, John draws the right conclusion from it: the Persian king "was embittered in his enmity by this occurrence, for he imagined that the Romans were stirring up these tribes against him".⁴

The return by the northern road

Sizabul sent Zemarchus home with a new Türk envoy, Tagma Tarkhan, "since the former one, Maniakh, had died", and with Maniakh's young son, who "had been given his father's title" in recognition that Maniakh "had been very friendly and loyal to Sizabul".¹ The Sogdian who had proposed the alliance did not live to see its first exchange of embassies completed.

Menander's fragment 10.4 is the itinerary, and it is the record's second transmission. The road Maniakh had taken east was now taken west by a Roman general who wrote it down.

The itinerary has two features that mattered for everything afterwards. It worked: a Roman embassy and a Türk embassy had gone from the Türk court to the Black Sea without touching Persian soil. And it was contested: Persia had posted four thousand men on the Kuban and more in Suania to stop it, and had, John of Epiphania says, "planned to bribe the Alans through whose lands they were about to make their passage".⁶ Khosrow understood what the road meant before Constantinople did.

What changed and what was replaced

A second road to the East

The first thing Byzantium received was a route. Before 568 the empire's eastern trade had two doors, both Persian-keyed: the frontier towns of Mesopotamia and the Indian Ocean ports. After 571 it had a third, the steppe-Caucasus corridor that Zemarchus had walked: from the Volga across the Kuban and the Alan passes to Lazica and Trebizond. Nina Pigulevskaya, in the Soviet study that first set Byzantium's eastern trade in a single frame, treated the Türk embassies as the moment the "northern road" became a Byzantine policy rather than a merchants' rumour.¹² Mark Whittow, in the most recent survey, calls the result "Byzantium's Eurasian policy": a standing effort, from the 560s to the ninth century, to keep an ally on the steppe behind Persia.²²

The route was never a silk highway. Sogdian merchants continued to sell most of their raw silk through Persia, as they had always done, and Byzantium continued to buy it there.¹¹²⁰ But the road stayed open as a diplomatic and military fact. The Türks took Alania after 572, and the Alan king Sarosius fought for the Romans in the war that followed.¹³ In 584 Türk forces struck the rear of the Avars at Anchialos in Thrace, an action some scholars read as a renewed Byzantine-Türk understanding under Maurice.¹⁹ In 595 a Türk qaghan sent Maurice a letter styling himself "the great lord of seven races and master of seven zones of the world", preserved by Theophylact Simocatta.⁷ And in 627 Heraclius met the Türk yabghu outside Tiflis and marched into Persia with Türk horsemen at his side, the alliance the Sogdians had first proposed, delivered sixty years late.²⁷

A new category in Roman diplomacy: the steppe ally

The second thing received was a category. Roman diplomacy had long managed barbarians on its borders; it had never before courted a power on its enemy's far side. Between 569 and 576 Menander records a sequence of embassies that Ekaterina Nechaeva has analysed as a system, a "block organisation" of reciprocal missions, gifts and oaths.²⁶

The formula Sizabul and Justin swore in 569, that "the friend of the Romans was his friend and their enemy his enemy and that this should be unbreakable and inviolable", became the template Valentinus invoked in 576 and the Khazar alliance would inherit.¹ So did the manners. Maurice's Strategikon, written within a generation, contains the first Roman military ethnography of the Türks, and it is not admiring: "superstitious, treacherous, foul, faithless, possessed by an insatiate desire for riches. They scorn their oath, do not observe agreements, and are not satisfied by gifts."⁹ The empire had acquired an ally and, in the same decade, a prejudice.

The mechanics were new as well. Every embassy in the table travelled with the previous one's envoys: Maniakh went east with Zemarchus, Tagma Tarkhan and Maniakh's son came west with him, and by 576 there were, Menander says, 106 Türks in Constantinople waiting to be taken home with Valentinus.¹ Gifts moved both ways and were inspected on arrival, as the Türk gifts of 568 were, and the qaghan's letters came in Sogdian script with Sogdian readers attached. Nechaeva's point about this system is that it was built for repetition: oaths, hostages of courtesy and reciprocal escorts were the machinery by which two courts with no common language and no common god could keep a treaty alive for eight years across four thousand kilometres.²⁶ Byzantium would run the same machinery with the Khazars for two centuries.

What the Sogdians got: a seat, not a market

The senders' receipt was smaller than the receivers'. Maniakh had gone to Constantinople to sell silk. The emperor showed him silkworms. What the Sogdians kept was the role: from 568 onward they were the Türk qaghanate's diplomatic class toward the west as they already were toward China. Maniakh's son carried his father's title to Constantinople in 571; Tagma Tarkhan, the Türk envoy, "ranked" only "immediately" above him.¹ Li and Kordosis put the structural point in a sentence: "the Sogdians, being subjects of the Western Türks, always served as Türk envoys to Persia and Byzantium".¹⁹ The Afrasiab murals in Samarkand, painted around 650 in the hall of the Sogdian king Varkhuman, show the arrangement from the inside: Türk attendants with long plaits escort the ambassadors of China, Chaganian and Chach into a Sogdian audience hall.²⁸²⁹

Byzantine mint of Constantinople. Solidus of Justin II, 565–578 CE. The Metropolitan Museum of Art, New York, accession 04.35.3353. CC0 (Met Open Access) via Wikimedia Commons. · CC0

What was displaced

Three things were displaced, none of them completely.

• The premise of the Persian monopoly. Khosrow had burned the silk to show that Persia alone decided who sold silk to whom. After 571 that was demonstrably false: a Türk embassy had reached the Bosphorus around him, and a Roman one had come back. Persia's answer was to garrison the Caucasus passes, bribe the Alans, and finally go to war.⁶²⁰
• The Avar understanding. The Türks' price for friendship, stated in 563 and restated in 568, was that Byzantium abandon the Avars. Justin II refused the Avars tribute in 565 and the Türk alliance of 568 was "apparently directed against the Avars and Persians".²³ The understanding did not survive the Avar siege of Sirmium: by 574 Tiberius had bought peace from the Avar qaghan, and the Türks learned of it.²³¹⁹
• The idea of a single steppe. Before 563 Byzantium's north was Huns, Avars and Alans, a border zone. After 571 it was a map with an empire on it, eight provinces, a senior ruler named Arsilas, a western shad on the Volga, and a road to all of it.¹

What did not change

The record owes the reader the negative list, because the popular version of this story, the "Türk-Byzantine silk route", overstates it.

• Raw silk continued to reach Byzantium overwhelmingly through Persia; the northern road carried envoys, gifts and the occasional caravan, not the trade.¹¹²⁰
• Byzantine sericulture remained small; the empire kept importing eastern raw silk past the tenth century, and the kekolymena grades remained an imperial monopoly.
• No joint Türk-Roman campaign against Persia was ever fought. Whittow suggests one was planned for 573; it did not happen.²²
• The Türk alliance lasted eight years. Sogdian merchants outlasted it by two centuries.

What the cost was

The war the alliance helped start

The bill arrived in 572. Theophylact Simocatta lists the causes of the war that broke the Fifty-Year Peace, and the Türk embassies are on the list beside Suania, the Arab clients, the Armenian revolt and Justin's refusal to pay the annual 30,000 solidi.⁷²⁰ John of Ephesus, who lived through it in Constantinople, ranks them: "the first cause, he says, why the peace was broken was, that the Persarmenians surrendered themselves and their country to Rome: but there was a second cause, which greatly embittered the enmity, namely, that the king of the Romans had sent ambassadors to the barbarous tribes, who live in the interior of the Persian dominions, and whom they call Turchios".⁴ John of Epiphania, the lost historian on whom Theophylact drew, makes the Persian attempt to stop Zemarchus on the Alan road one of the opening acts of hostility.⁶ Michael Whitby, in the Cambridge Ancient History, treats the Türk entente as one of the several threads of Justin's "folly", none sufficient alone, all pulled at once.²¹

The record does not claim that the Sogdians caused the war. It claims what the sources claim: that the alliance Maniakh brokered was named by contemporaries, Roman and Persian, as one of the reasons the two empires went back to fighting, and that the fighting lasted from 572 to 591.

Apamea and Dara, 573

The first year's accounting is in John of Ephesus, book VI. While Khosrow sat before Dara, he sent the marzban Adarmahan with "a large body of troops" across the Euphrates into Syria. Adarmahan "stormed numerous castles, which fell in his way, and rased and burnt them", came to Apamea, and talked his way in: the bishop and the city's leaders, remembering that Khosrow had spared the buildings in 540, opened the gates, and the Persians "began to lay hands on and bind men and women, and spoil the city". They put the whole population and the bishop outside the walls, "set fire to it, and burnt the whole of it from one end to the other", and brought the captives back to the king. "And the captives were counted in the king's presence, and their number was two hundred and ninety-two thousand; and they were divided among the troops, and taken into the Persian territories."⁴ Evagrius, a native of the region, confirms the sack of Apamea and the capture of Dara in the same campaign.⁵

The figure of 292,000 is a chronicler's figure. No modern historian takes it literally; the Persians could not have fed such a column. What it records is a scale: the deportation of a provincial city and its countryside, counted as booty. Dara fell in November 573 after a siege of several months, and "all this and the people he took, and carried with him into his land".⁴²⁰ Justin II, "not bearing what had befallen him with resignation suited to a human being", in Evagrius's words, "falls into a state of frenzy".⁵ Sophia and Tiberius bought a year's truce in 574 for 45,000 solidi, then three years more at 30,000 a year, while the war went on in Armenia.²⁰

Two thousand women and a river

One passage in John of Ephesus connects the deportations directly to the alliance this record is about. Khosrow, "intoxicated with the glory he had gained in this expedition", ordered "that there should be selected from the captives two thousand virgins, full-grown, and of perfect beauty", dressed "in every thing like brides", and "sent as a present to the barbarians who dwell in the heart of his territories, and who are called Turks, in order to please and content them, and hire their services".⁴ The king who had poisoned the Türks' envoys six years earlier was now bidding for the Türks with Roman women.

John says the women were escorted slowly, "that they might not grow thin", and that within fifty leagues of their destination they reached "a very broad and rapid river". They asked leave to wash, asked the guards to stand off, "and all signed themselves in the Name of our Lord Jesus Christ, they threw themselves suddenly by one hasty rush into the river, and were drowned". The guards "saw not a single one of them alive".⁴

The record presents this as John presents it, a report that reached Constantinople through the Syrian captives who travelled with the escort, told in the register of martyrdom. Whether the number is exact, whether every detail is true, cannot be known. What can be said is that a contemporary believed the Türk alliance had a price, that the price was being paid by Syrian women, and that he wrote it down.

Bosporus, 576

The Türks collected from Byzantium too. In 576 Tiberius, now Caesar, sent Valentinus east with a party of 106 Türks who had accumulated in Constantinople from earlier embassies, to announce his elevation, renew the treaty "which Silzibul and the Emperor Justin had made when Zemarchus first came there", and ask the Türks to attack Persia.¹ He found Turxanthus, the western shad, in a war camp near the Volga, and Turxanthus had found out about the Avars.

"Are you not those very Romans who use ten tongues and lie with all of them?" Menander has him say, stuffing ten fingers into his mouth: "As now there are ten fingers in my mouth, so you Romans have used many tongues. Sometimes you deceive me, sometimes my slaves, the Uarkhonitai."¹ He accused the Romans of keeping the road to Byzantium secret from the Türks and of making treaties with their runaway slaves. Then, since "my father, Silzibul, is recently dead", he ordered the envoys to "follow the custom which prevails amongst us for the dead and slash your faces with daggers". "Immediately Valentinus and his companions slashed their own cheeks with their own daggers."¹ Turxanthus sacrificed four bound Hunnic captives at the funeral, sent Valentinus on to his brother Tardu at the Golden Mountain, and "swore that he would presently lay siege to Bosporus". He sent Bokhan "with a very large force", Anagai of the Utigurs joined him, and "the city of Bosporus was taken". The envoys "were detained by Turxanthus, insulted, mocked and otherwise ill-treated, and then sent away".¹

Bosporus is Panticapaeum, modern Kerch, the empire's oldest possession on the Crimean strait. By 579 Türk forces were raiding toward Chersonesus, and only the Türk civil war of 581 to 584 lifted the pressure on the peninsula.¹¹⁵ The alliance that had been sworn "unbreakable and inviolable" in 569 ended seven years later with a Roman city in Türk hands and Roman envoys bleeding from the face. Li and Kordosis draw the structural lesson: the nomads, they write, "were only treated as pieces on a chessboard for keeping the balance with the rest of the superpowers", and "the Western Türks realised the role of 'pawn' they had in the Byzantine 'nomadic' policy".¹⁹

The Sogdians' bill

The senders paid first and are counted last. One embassy's silk was burned at Ctesiphon. A second embassy's members, "all but three or four", were poisoned at the Persian court. Maniakh himself died somewhere between the Golden Mountain and the Bosphorus, on the road he had opened, and his title passed to a boy.¹ Sogdiana, which had asked only to sell silk in Persia, became a frontier between two empires at war, and in the following century the Türk-Sasanian-Chinese contest for it, then the Arab conquest of Samarkand in 712, would do to the Sogdian cities what the Sogdian merchants had spent four centuries avoiding.¹¹¹³

There is a last, smaller item. The Sogdians had asked for a market and been given a role, and the role carried the risk of the market without its profit. Sören Stark's survey of the archaeology of the Türk period shows Sogdian colonists, scribes and traders spread across the qaghanate from the Zarafshan to Mongolia in these decades, a diaspora whose prosperity depended on the peace of the steppe and the goodwill of whichever power held it.¹³ When the Türks and the Persians went to war, and when the Türks and the Romans fell out, it was Sogdian envoys who were poisoned, Sogdian envoys who died on the road, and Sogdian caravans that paid the Caucasus garrisons. The record cannot count them. It can say that the people who proposed the alliance were the people with the least protection from its failure.

The bill

• A cargo of raw silk, bought and burned at Ctesiphon, before 568.
• A Türk embassy to Persia, poisoned "all but three or four", before 568.
• Maniakh, dead on the road between 569 and 571.
• A twenty-year war, 572 to 591, of which the Türk alliance was a contemporary-named cause.
• Apamea burned "from one end to the other" and its people counted at 292,000 before Khosrow, 573.
• Dara taken and emptied, November 573; Justin II's collapse; truces bought for 45,000 and then 30,000 solidi a year.
• Two thousand captive women sent as a gift to the Türks who drowned themselves rather than arrive.
• Bosporus taken by the Türks, 576; Roman envoys made to slash their faces; the Crimea raided to 581.

The transmission was an alliance, a route and a way of thinking about the steppe, and all three outlasted the men who carried them. The bill was paid in Syria and the Crimea by people who had never heard of Maniakh. That is the usual arrangement, and the atlas records it.