Near the Great Zab in northern Iraq, the Abbasid army commanded by Abd Allah ibn Ali confronted the Umayyad army of Caliph Marwan II. Al-Tabari dates Marwan's rout to 25 January 750.
The Umayyad cavalry commander al-Walid ibn Mu'awiya attacked despite Marwan's order not to begin the battle, and drove back the Abbasid right wing under Abu Awn.
Musa ibn Ka'b advised Abd Allah to dismount his troops. They crouched with extended lances, checked the Syrian attack, and then drove Marwan's forces into flight.
The bridge
Marwan fled and the pontoon bridge was cut, trapping part of his army. Al-Tabari says more men drowned than were killed in combat.
The Abbasids retied the bridge and recovered bodies — about 300 by that account — and took Marwan's camp, weapons and wealth.
News of the victory went to Caliph Abu al-Abbas, who ordered 500 dirhams for every participant and set their monthly salary at 80 dirhams.
Marwan got away
Abd Allah stayed at the battlefield for seven days. Marwan escaped toward Harran and went on resisting.
He was killed near Busir in Egypt in August 750, not at the Zab, and Wasit and other centres still held out after the battle.
These narratives reach us through al-Tabari, compiled generations later from multiple transmitted reports, so the speeches and individual tactical anecdotes belong to their transmitters rather than to eyewitnesses.
Researched 23 Aug 2026·4 sources·not yet auditedTraditional date
How this was checked
Researched from the web into a fact sheet and rewritten from that sheet. This article has not yet had an independent model audit. How the pipeline works →
Traditional date. This date is the conventional one, not one fixed by a contemporary record.
What the sources leave uncertain
The 25 January date is the standard conversion of 11 Jumada II 132 AH. Al-Tabari's translated text calls it Saturday, while Moshe Sharon's modern reconstruction calls it Sunday; both give 25 January 750.
Battle narratives preserved by al-Tabari were compiled generations afterward and present multiple transmitted reports, so speeches and individual tactical anecdotes should be attributed rather than treated as eyewitness certainty.
Ziyad ibn Salih advanced beyond Chach and occupied Talas in 751, which prompted Gao Xianzhi to march from Kucha with Tang, Ferganan and Karluk forces to confront him.
The quarrel behind it was recent. Gao had deceptively captured and sacked Chach in 750 and Emperor Xuanzong had executed its king, driving an escaped Chach prince and other Sogdians to appeal to Ziyad for vengeance.
The year is secure but no exact day survives; July is commonly given from later Islamic chronology.
Five days, then a defection
Chinese sources describe a five-day confrontation near Talas, after which the Karluks changed sides and joined Ziyad's forces in routing Gao's army.
Gao's total force including allied contingents is put at about 20,000 to 30,000. He escaped with a few thousand men, who reportedly had to fight through panicked Ferganan allies obstructing the retreat.
No extant source gives a reliable size for Ziyad's army, and the battlefield has not been identified beyond the Talas region.
What it did not settle
Gao was soon removed from the Anxi command, and Abbasid influence expanded into Chach and possibly Talas.
Tang power in Central Asia did not end there: the Anxi forces were rebuilt and campaigned successfully in 753. No primary source supports the frequent claim that a Tibetan army fought for the Abbasids at Talas.
Among the captured Tang soldiers was Du Huan, who remained in the Islamic world and returned to China by sea in 762.
Researched 23 Aug 2026·4 sources·not yet audited
How this was checked
Researched from the web into a fact sheet and rewritten from that sheet. This article has not yet had an independent model audit. How the pipeline works →
What the sources leave uncertain
The year 751 is secure, but no exact day survives. July is commonly given from later Islamic chronology.
The precise battlefield has not been identified beyond the Talas/Taraz region.
The surviving accounts are brief. Chinese sources describe Karluk defection, but their motive is a modern inference; other traditions may treat the Karluks as Abbasid allies from the outset.
No extant source gives a reliable size for Ziyad's Abbasid army.
No contemporary source records the founding of a Samarkand paper mill or a transfer of papermaking technology in 751. The year was attached retrospectively, because that was the year of the Battle of Talas.
The familiar story first appears in a work by al-Tha'alibi, a Persian author who died in 1038: he said Chinese prisoners captured at Talas introduced papermaking to Samarkand.
The phrase paper mill misleads too — the sources do not document a newly constructed mechanized mill at Samarkand that year.
What was already there
Paper was already used and probably manufactured in Central Asia, including Samarkand, before 751, which makes a sudden introduction by Talas prisoners unlikely.
The material points the same way. Early Islamic paper was characteristically made from rag fibres, unlike the bast-fibre paper then typical in China.
Jonathan Bloom argues from chronology, materials and source criticism that the transfer story is unlikely to be factual.
The one captive who wrote it down
Some Tang soldiers were certainly captured at Talas, but none can be securely identified as a papermaker.
Du Huan, the known captive, returned in 762 and described Chinese painters, silk weavers and metalworkers in Kufa. He did not identify papermakers among them.
Samarkand did become important in paper's westward diffusion; what fails verification is the dramatic one-year origin story.
Researched 23 Aug 2026·5 sources·not yet auditedDate disputed
How this was checked
Researched from the web into a fact sheet and rewritten from that sheet. This article has not yet had an independent model audit. How the pipeline works →
Date disputed. Sources disagree about when this happened.
What the sources leave uncertain
This should not be presented as a verified event. The year 751 was attached retrospectively because it was the year of the Battle of Talas.
The term 'paper mill' can also mislead: the sources do not document a newly constructed mechanized mill in Samarkand that year.
Some scholars allow that individual captives could have contributed techniques, but that weaker possibility does not establish the traditional founding story.
At Saint-Denis in 754, Pope Stephen II anointed Pepin and his sons Charles and Carloman as kings of the Franks. That the papal anointing took place that year is well supported.
What is securely attested is anointing with holy oil, not the placing of a physical crown.
Pepin had been king since 751. What the rite supplied was direct papal consecration, extended to the next Carolingian generation.
What the Clausula adds
The Clausula de unctione Pippini says Stephen also blessed Queen Bertrada and the Frankish nobles, and bound the Franks not to choose a king outside Pepin's descendants.
Pepin and his sons received the title Patricius Romanorum, although the title's precise legal force remained obscure.
The familiar date of 28 July comes from that Clausula tradition; a modern medieval-source repertory notes that the surviving text may be a ninth- or tenth-century composition or forgery.
Why the pope was in Francia
Stephen had crossed the Alps seeking a protector against the Lombard king Aistulf, after Byzantine help proved ineffective. Pepin had already agreed to aid the papacy, and the Saint-Denis rite sacralized that alliance and the Carolingian succession.
The territorial side of the bargain came later: the Donation of Pepin was chiefly fulfilled after campaigns in Italy in 756.
Modern scholarship also questions whether Pepin was anointed at all in 751, so calling 754 a re-consecration needs caution.
Researched 23 Aug 2026·5 sources·not yet auditedTraditional date
How this was checked
Researched from the web into a fact sheet and rewritten from that sheet. This article has not yet had an independent model audit. How the pipeline works →
Traditional date. This date is the conventional one, not one fixed by a contemporary record.
What the sources leave uncertain
The ceremony certainly occurred at Saint-Denis in 754, probably in July. The exact date 28 July comes from the Clausula tradition and cannot be held firmly by all chronologies.
The surviving Clausula's date and authenticity have been debated; a modern medieval-source repertory notes that the extant text may be a ninth- or tenth-century composition or forgery.
Modern scholarship questions whether Pepin was anointed at all in 751. If he was, Boniface was not demonstrably the officiant; therefore 're-consecration' should be used cautiously.
Stephen is numbered Stephen II in modern lists that omit an unconsecrated namesake, and Stephen III in lists that include him.
Abd al-Rahman ibn Mu'awiya crossed from the Maghreb to the southern coast of al-Andalus in 755, beginning his attempt to build a new political base after the Abbasid overthrow of his Umayyad family.
The date is not secure. One scholarly study places the event only within Rabi I or II of 138 AH — 14 August to 11 October 755 — while the TDV İslâm Ansiklopedisi gives the traditional conversion of 14 August.
Most medieval Arabic chronicles name al-Munakkab, modern Almuñécar. The Fath al-Andalus instead describes the ship being driven to Biṭruh Riyāna, identified by a modern study with Burriana beach farther west.
Met at the shore
He did not step onto an empty coast. The supporters Ubayd Allah ibn Uthman and Abd Allah ibn Khalid met him and took him to Ubayd Allah's stronghold at Torrox.
The freedman Bedr had prepared the ground, obtaining promises of support from Umayyad clients and Yemeni-aligned leaders before the crossing.
What arrived was one dynastic claimant with a small following, leaning on clients and factions already established in al-Andalus.
The governor tried talking first
When the governor Yusuf al-Fihri learned of the landing, he sent envoys with offers intended to make Abd al-Rahman abandon his political ambitions. Abd al-Rahman rejected them.
The landing did not by itself establish the Emirate of Córdoba; he still had to gather an army and defeat Yusuf in May 756.
What he eventually founded was an emirate. The caliphal title at Córdoba was adopted only by Abd al-Rahman III, in 929.
Researched 23 Aug 2026·3 sources·not yet auditedDate disputed
How this was checked
Researched from the web into a fact sheet and rewritten from that sheet. This article has not yet had an independent model audit. How the pipeline works →
Date disputed. Sources disagree; the research places it at ~0755-09.
What the sources leave uncertain
The traditional September 755 dating is not secure. One scholarly study places the event only within Rabi I or II 138 AH, equivalent to 14 August–11 October 755; the TDV İslâm Ansiklopedisi gives the more precise traditional conversion 14 August 755.
Almuñécar is named by most Arabic chronicles, but a modern analysis of the Fath al-Andalus argues for Biṭruh Riyāna/Burriana beach, west of Almuñécar. The exact landing site should therefore not be presented as certain.
Narrative details survive in chronicles written well after the event and may contain dynastic legend.
Negotiations had failed: Yusuf al-Fihri's proposed settlement, which included marriage and land, did not resolve the competing claims to rule. In May 756 Abd al-Rahman moved his forces across the Guadalquivir during the night and attacked Yusuf's army near al-Musara, outside Córdoba, on the morning of 15 May in the TDV chronology.
He defeated the forces of Yusuf and of al-Sumayl ibn Hatim al-Kilabi, who commanded the established Qays-aligned regime alongside him.
The exact day is disputed. Some modern accounts give 14–15 May, and the 1911 Encyclopaedia Britannica gives 16 May.
Into the city
The victory opened Córdoba. Abd al-Rahman entered, took the governor's palace and was proclaimed emir, replacing Yusuf's governorship with an Umayyad regime centred on the city.
The New Cambridge Medieval History summarizes the sequence the same way: Abd al-Rahman defeated Yusuf and al-Sumayl outside Córdoba in May, took the city and palace, and was proclaimed amir.
Both defeated commanders escaped.
What 15 May was not
Seville belongs to an earlier stage of the campaign, entered or taken probably in March 756. The claim that it was captured without violence appears in popular summaries but is not established by the stronger sources consulted.
Winning at al-Musara did not deliver uncontested control of the peninsula either; Yusuf, his family and other factions went on resisting.
The Great Mosque of Córdoba belongs to Abd al-Rahman's later reign, not to the day he entered the city.
Researched 23 Aug 2026·5 sources·not yet auditedDate disputed
How this was checked
Researched from the web into a fact sheet and rewritten from that sheet. This article has not yet had an independent model audit. How the pipeline works →
Date disputed. Sources disagree; the research places it at 0756-05-15.
What the sources leave uncertain
The month and sequence are secure, but sources disagree over the exact day. The TDV encyclopedia gives the morning of 15 May; some modern accounts use 14–15 May, while the 1911 Encyclopaedia Britannica gives 16 May.
Seville was entered or taken earlier, probably in March 756. It should not be merged into a single May 15 event.
The assertion that Seville was captured 'without violence' appears in popular summaries but is not established by the stronger sources consulted.
The Zizhi Tongjian dates An Lushan's capture of the Tang eastern capital to the dingyou day of the twelfth lunar month of Tianbao 14 — commonly converted as 18 January 756, though another modern military chronology gives 19 January.
Rebel troops entered through all four gates and began killing and plundering.
Feng Changqing's newly recruited loyalist army had already been beaten at Wulao, at Kuiyuan and at Luoyang's Upper East Gate.
Out through the garden wall
Feng kept fighting at the Duting relay station and the Xuanren Gate, then broke through the western wall of the imperial garden and fled.
The Tang officials Li Zheng, Lu Yi and Jiang Qing refused or failed to flee and were executed after the city fell.
Lu Yi had sent his family away, put on court dress, and waited at the Censorate.
The numbers, and the pause
The Zizhi Tongjian says An mobilized 150,000 troops at the start of the revolt while publicly claiming 200,000. Feng's emergency levy was 60,000 recruits raised in ten days.
Feng retreated west to join Gao Xianzhi and urged withdrawal to the defensible Tong Pass; that pause let the Tang reinforce the western approaches to Chang'an, which remained in Tang hands.
An stayed at Luoyang preparing to proclaim a new dynasty, and on 5 February declared himself emperor of Yan.
Researched 23 Aug 2026·4 sources·not yet auditedDate disputed
How this was checked
Researched from the web into a fact sheet and rewritten from that sheet. This article has not yet had an independent model audit. How the pipeline works →
Date disputed. Sources disagree; the research places it at 0756-01-18.
What the sources leave uncertain
The Zizhi Tongjian dates the capture to the dingyou day of the twelfth lunar month of Tianbao 14. The commonly published conversion is 18 January 756, but another modern military chronology gives 19 January, producing a one-day conversion disagreement.
The title’s '200,000-strong army' repeats An’s advertised total. The Zizhi Tongjian says he actually mobilized 150,000 while claiming 200,000.
The annal is a major historical source but was compiled centuries after the rebellion from earlier records.
At the beginning of 756 the Lombard king Aistulf broke his previous agreement and besieged Rome for three months, encircling the city and devastating property, churches, cemeteries and the territory outside the walls.
The papal biography says his troops dug into the suburban cemeteries and carried away saints' bodies.
Pope Stephen II sent urgent appeals to Pepin and the Franks by sea, because Aistulf had blocked the land routes; one of them was written in the voice of St Peter himself, addressing the Franks as the apostle's adopted sons.
Pavia again
Pepin led a second expedition into Italy in May, broke the Lombard defenses near the Susa passes and besieged Aistulf in Pavia once more, forcing his surrender under harsher conditions.
Those terms took one third of Aistulf's royal treasury.
A Byzantine envoy, George the protoasecretis, asked that the former imperial territories be restored to Constantinople. Pepin rejected the demand and declared that the recovered cities would be transferred to the Roman Church.
Keys laid on the shrine
Pepin commissioned Abbot Fulrad of Saint-Denis to receive the surrendered cities. Fulrad collected their keys, hostages and leading inhabitants, then deposited the keys and the grant at the shrine of St Peter in Rome.
The transfer covered Ravenna and cities in the exarchate, Emilia and the Pentapolis — 21 cities and castles, in addition to their territories.
Pepin's original grant document is lost. The papal biography says it was kept in the Roman Church's archive, and Treccani notes that it is unclear whether the transfer gave the papacy the exarchate and Pentapolis in their complete territorial extent.
Researched 23 Aug 2026·5 sources·not yet audited
How this was checked
Researched from the web into a fact sheet and rewritten from that sheet. This article has not yet had an independent model audit. How the pipeline works →
What the sources leave uncertain
Pepin’s original grant document is lost. The papal biography says it was preserved in the Roman Church’s archive but does not reproduce it.
Treccani notes that it is unclear whether the transfer actually gave the papacy the exarchate and Pentapolis in their complete territorial extent.
The sources are strongly shaped by the papal viewpoint; the Liber Pontificalis biography is particularly hostile to Aistulf.
Stephen is called Stephen II in modern lists that exclude a pope-elect who died before consecration, but Stephen III in some older scholarship.
The chronicle records the arrival of Shi Chaoyi's severed head at Chang'an on the jiachen day, the thirtieth of the first lunar month, conventionally converted to 17 February 763. It does not record his suicide on that date.
The textual-critical note in the Zizhi Tongjian compares contradictory accounts — suicide by hanging, capture and beheading, death in an earlier month, different locations. Its editor concludes only that Shi died during the first lunar month of 763.
The New Book of Tang assigns the suicide to the jia-shen tenth day, but the Zizhi Tongjian editor says the basis for that exact date is unknown.
Refused at every gate
After repeated defeats, Shi escaped the siege of Mo Prefecture and rode toward Fanyang seeking troops to continue the Yan resistance. Tian Chengsi, who had urged him to go for reinforcements, then surrendered the city and Shi's family.
Li Huaixian, Shi's own military governor at Fanyang, had already offered to submit to the Tang and barred him from entering. His subordinate Li Baozhong told Shi openly that Yan had lost Heaven's support and that the city now served the Tang.
Shi tried Guangyang, was refused there too and turned toward the Xi or Khitan frontier. Li Huaixian's men caught up with him at Wenquan Stockade, where he hanged himself in a forest rather than be captured.
How the war was closed out
Li Huaixian had Shi beheaded and sent the head to the Tang court. The pursuers were his own forces, not Tang troops.
With the last Yan emperor dead and the remaining northern commanders submitting, Shi's suicide conventionally marks the end of the An–Shi Rebellion, traditionally spanning 755 to 763.
Emperor Daizong accepted their submission and retained several of them as regional military governors, so the rebellion ended without an immediate restoration of direct Tang administration throughout Hebei.
Researched 23 Aug 2026·3 sources·not yet auditedDate disputed
How this was checked
Researched from the web into a fact sheet and rewritten from that sheet. This article has not yet had an independent model audit. How the pipeline works →
Date disputed. Sources disagree; the research places it at ~0763-02.
What the sources leave uncertain
17 February 763 is securely associated with the arrival of Shi Chaoyi's severed head at Chang'an, not securely with his death.
The textual-critical note in Zizhi Tongjian compares contradictory accounts: suicide by hanging, capture and beheading, death in an earlier month, and different locations. Its editor concludes only that Shi died during the first lunar month of 763 and records the head's arrival on that month's thirtieth day.
The New Book of Tang assigns the suicide to the jia-shen tenth day, but the Zizhi Tongjian editor explicitly says the basis for that exact death date is unknown.
According to Ibn al-Qifṭī's later account, an Indian mission including the scholar Kanka arrived at Baghdad in AH 156 — 772/773. Al-Bīrūnī instead dated the mission to AH 154, or 770/771.
The visitor brought a Sanskrit astronomical work related to, but apparently not identical with, Brahmagupta's Brāhmasphuṭasiddhānta.
Modern scholarly summaries prudently say the early 770s, or soon after 770.
What was actually translated
Caliph al-Manṣūr requested an Arabic translation. Al-Fazārī collaborated with the Indian astronomer and based his Great Sindhind astronomical tables on the imported work.
Indian parameters and computational methods entered the developing Baghdad astronomical tradition by that route. The episode is known through later writers, because al-Fazārī's works and the original Arabic translation do not survive intact.
Al-Fazārī's identity is itself uncertain: medieval biographers may have turned one astronomer into a father-and-son pair.
Zero is not what the record shows
The surviving evidence describes the arrival and translation of an astronomical handbook. It does not document a ceremony or a day on which the number 0 was introduced to Baghdad.
Zero has separate histories as a placeholder, a numeral glyph, a number, a name and a set of arithmetic rules, and no consulted source shows them arriving together in one datable event.
Al-Khwārizmī's systematic arithmetic using the nine Hindu digits and zero came later, around 825; the earliest universally accepted dated Indian inscription using a zero symbol is from 876.
Researched 23 Aug 2026·5 sources·not yet auditedDate disputed
How this was checked
Researched from the web into a fact sheet and rewritten from that sheet. This article has not yet had an independent model audit. How the pipeline works →
Date disputed. Sources disagree about when this happened.
What the sources leave uncertain
The year is not independently secure: Ibn al-Qifṭī gives 772/773, al-Bīrūnī gives 770/771, and modern scholarly summaries prudently say 'early 770s' or 'soon after 770.'
The episode is known through later writers because al-Fazārī's works and the original Arabic translation do not survive intact.
It is uncertain exactly which Sanskrit work arrived. Scholars describe it as related to Brahmagupta's tradition rather than confidently identifying it as the Brāhmasphuṭasiddhānta itself.
No consulted scholarly source demonstrates that a particular zero glyph, zero as a number, and the complete decimal positional system all arrived in Baghdad in one datable 773 event.
Charlemagne had besieged Pavia since September 773. He returned from an Easter visit to Rome and captured the Lombard capital in June 774.
The Royal Frankish Annals report that he seized King Desiderius, his wife, a daughter and the palace treasure. Lombards from Italy's cities then submitted to Charles and the Franks.
The exact day of the surrender is not consistently reported; the first surviving document calling Charlemagne king of the Franks and Lombards is dated 5 June 774.
The kingdom kept its name
Desiderius was removed from power and sent into Frankish captivity, while his son Adelchis escaped to Constantinople.
Charlemagne retained the Lombard kingdom as a recognizable political unit rather than erasing it into Francia, and much of the existing Lombard administration initially stayed in place.
Several Lombard officeholders kept their posts; one of them, Hrodgaud of Friuli, rebelled in 776.
The Easter document
During Easter at Rome, on 6 April 774, Charlemagne and Pope Adrian I discussed papal territorial claims and drew up a document restating promises attributed to Pepin's agreement at Quierzy.
That list survives through the papal Liber Pontificalis, and historians dispute whether it represented enforceable sovereignty, future promises, papal patrimonial claims, or a later expansion of the record. Maritime Venetia remained within the Byzantine sphere, and Frankish conquest of Istria is usually dated around 788.
Becoming king of the Lombards did not make Charlemagne a Roman emperor; that coronation came in 800.
Researched 23 Aug 2026·6 sources·not yet auditedDate corrected
How this was checked
Researched from the web into a fact sheet and rewritten from that sheet. This article has not yet had an independent model audit. How the pipeline works →
Date corrected. The research places this at 0774-06.
What the sources leave uncertain
The sources securely place Pavia's capitulation in June 774, but the exact day of surrender is not consistently reported. The first surviving use of Charlemagne's double title is dated 5 June.
The title-as-listed combines at least two phases: Charlemagne's Easter visit and papal agreement at Rome, followed by Pavia's fall in June.
The extensive territorial list attributed to the Easter agreement survives through the papal Liber Pontificalis. Historians dispute whether it represented enforceable sovereignty, future promises, papal patrimonial claims, or later expansion of the record.
The political consolidation of Carolingian Italy was a process, not a single-day transfer of every Lombard or Byzantine territory.
The council opened on 24 September 787 in the Church of Holy Wisdom at Nicaea. Patriarch Tarasios addressed the assembly, and three formerly iconoclast bishops were reinstated that first day after reading professions of faith.
Across seven formal sessions the assembly read Pope Adrian I's letters, reconciled further repentant bishops, gathered patristic arguments for images and refuted the iconoclast Council of Hieria of 754.
It met at Nicaea at all because an earlier attempt at Constantinople in 786 had been broken up by iconoclast soldiers; Irene, regent for her son Constantine VI, reconvened it once the hostile guard was replaced.
Signed without recorded debate
At the seventh session, on 13 October, the bishops read and signed the council's doctrinal Definition without any debate recorded in the Acts.
The Definition ordered sacred images of Christ, Mary, angels, saints and holy people to be displayed in churches, on vessels and vestments, in homes and beside roads.
It distinguished honorable reverence toward an image from the true worship reserved to the divine nature, explaining that honor shown to an image passes to the person represented.
Carried back, then reversed
Tarasios immediately dispatched bishops, monks and clergy to Constantinople to explain the proceedings and decrees to Constantine VI and Irene.
The council also issued 22 disciplinary canons, reaching well beyond art to relics in altars, clerical finances, monasteries and episcopal discipline.
The settlement held less than three decades: Byzantine iconoclasm was imposed a second time from 815 to 843.
Researched 23 Aug 2026·6 sources·not yet auditedDate corrected
How this was checked
Researched from the web into a fact sheet and rewritten from that sheet. This article has not yet had an independent model audit. How the pipeline works →
Date corrected. The research places this at 0787-10-13.
What the sources leave uncertain
The secure date for the doctrinal decision is 13 October 787; the council itself began on 24 September.
A ceremonial imperial reception at Constantinople on 23 October is often called an eighth session, but the modern critical edition presents seven formal sessions at Nicaea and treats the Constantinople event as reception of the decrees.
Attendance figures vary substantially in older accounts—roughly 330 to 367—so no single exact bishop count should be treated as secure.
Pope Adrian sent no surviving formal letter acknowledging the completed council, apparently because his territorial demands were not met. His representatives participated, however, and his defense of the council in 794 demonstrates acceptance of its doctrinal decision.
View full image ↗ 
