Two blocs, neither of them a race
The Bahmani sultanate, founded in the plateau country in 1347, governed with an immigrant service class. Encyclopaedia Iranica describes its elite as drawn 'mainly of Iranian, Turk, Dakani, or Muslim migrants from northern India, in addition to the local Hindu population', with Iranians, Turks and Afghans doing most to shape the state's destiny [1]. A rich revenue base and a chronic shortage of cavalry drew men from Gilan, Shiraz, Hamadan, the Hijaz and the Horn of Africa for a century and a half; what the state could not do was digest them, and those who had arrived early enough to become native hardened into one party against those still arriving. Iranica's account of the Adil Shahis of Bijapur, where the split outlived the Bahmanis, gives the sharpest definition: one class 'consisted of fresh immigrants, usually Iranian and Shi'ite'; the other, 'called Deccanis, consisted of the descendants of Turkish, Abyssinian or Arab immigrants. Most often they were Sunnite and native-born Indians, preferring to speak indigenous languages, especially their own dialect of proto-Urdu known as Dakhni' [17]. Read that carefully and the ethnic reading collapses: the Dakhanis are themselves descended from Turks, Arabs and Africans. What divided the blocs was length of settlement, language of daily use, and the direction in which patronage ran.
Modern scholarship presses the point. A. C. S. Peacock notes that 'Richard Eaton has argued these factions were at least as much political and constructs as ethnic realities' [14]; Meia Walravens warns that the two words 'are used in the sources as shorthands for complex identities', and that we should not take them for 'monolithic groups in a continuous state of confrontation' [3]. The leadership proves it. The head of the Dakhani party, Malik Hasan Bahri, titled Nizam-ul-Mulk, was on Ferishta's own account a Brahmin of Vijayanagara, born Timapa son of Bheiroo, taken captive and converted [13]; the head of the Afaqis, Yusuf Adil Khan, was in the same chronicler's phrase a Turkish slave [9].
The quarrel was built into the architecture
This was not informal rivalry. Peacock states that 'In the Bahmani state this distinction was institutionalised, so that gharīban and Dakani nobles were arranged opposite sides of the royal hall of audience' [14]. The capitals carried the same charge, Bidar associated with the newcomers, Gulbarga with the older establishment. Ferishta traces the poisoning of relations to one ceremonial decision: after an Afaqi commander's victory over Khandesh, Ala-ud-din Ahmad II ruled that the Persians and Turks should thereafter take precedence over the Dakhanis and Abyssinians, and from that day, he writes, 'the most rooted inveteracy has existed between the Deccanies and Moguls' [7]. Whether one precedence ruling can bear that weight is doubtful; what it shows is what a seventeenth-century Deccan historian took the quarrel to be about, which was rank and pay.
The African dimension is essential. Habshis, Abyssinians who mostly arrived as military slaves, were foreign-born and yet belonged to the Dakhani bloc without ambiguity. Briggs's Ferishta notes that the Deccanis and Abyssinians 'on this as well as all other occasions seemed to make common cause against the foreigners' [8], and Eaton supplies the mechanism: 'Habshi ex-slaves generally allied themselves both culturally and politically with the Deccani class', within a regional identity already visible in the mid-fourteenth century and gathering force in the sixteenth [20]. One building may record the alliance. Mohit Manohar reads the Chand Minar at Daulatabad, raised in 1446 by an African military slave, as the product of a moment when the tower 'impressed upon Ahmad II the importance of retaining in his court dark-skinned officers from India and Africa (dakkaniyan) at a time when their standing was threatened by the lighter-skinned gharīban' [16].
Chakan, 1446: the mechanics of an atrocity
The largest killing of the feud followed a military disaster. Malik-ut-Tujjar Khalaf Hasan Basri, an Afaqi who had insisted on taking an all-Foreign army into the Konkan, was destroyed near Sangameshwar with seven or eight thousand men. The survivors fell back on the fort of Chakan, north of present-day Pune, and the Dakhani officers, fearing they would report the debacle first, got their own version before the drunken sultan, who ordered them put to the sword. A two-month siege failed; the besiegers then produced a forged pardon and swore on the Quran, and about 2,500 Foreigners came out, 1,200 of them Sayyids. Three hundred of the principal men were invited to eat; armed men rushed in at a signal and 'put every soul to the sword', while four thousand Dakhani horse fell on the camp outside and 'put every male to death, even the very infants at the breast' [6]. Haig's accounting differs - 'put to the sword 1200 Sayyids, 1000 other foreigners, and five or six thousand children' [21]. These are chroniclers' numbers and they do not agree; what they agree on is a mass killing under a false oath, which Ferishta compares to Karbala.
Chakan settled nothing: a survivor reached Bidar, and the sultan executed the Dakhani leaders responsible. Walravens records that the mutual accusations 'led to siyāsa convictions of the Foreigners, and five years later, in 855/1451, of a group of Deccanis who had been detained on account of their role in the Foreigners’ deaths' [5] - the state punished both blocs in turn under its discretionary criminal jurisdiction. Forty-five years later the wheel turned: a Dakhani and Habshi conspiracy tried to murder Mahmud Shah Bahmani II in his own palace, was beaten off by Turkish guards, and at sunrise the sultan 'commanded the foreigners to enter the houses of the treacherous Deccanies and Abyssinians, and put them to death without distinction', orders executed on three consecutive days [10]. No side of this quarrel holds the moral ground.
Mahmud Gawan: the reform, the killing, and the contemporaries
Mahmud Gawan, a Gilani merchant who landed at Dabhol in 1453 and rose to be chief minister, was the Afaqi who tried to hold the ring. Eaton places him in the flow of migration: 'Gawan joined the stream of other Westerners who for decades had been migrating to the Deccan and had taken up service in Bidar, the Bahmani capital', adding that the enrichment such men brought 'came at a heavy cost' [19]. His trading network explains why the bloc kept renewing itself: the Deccan bought its warhorses overseas, and foreign status was less a sentiment than a portfolio of transregional assets - horses, credit, Persian chancery skill - the Dakhanis could not match and resented. Gawan's reform split four provinces into eight, resumed districts as crown land, took military appointments from governors and introduced muster inspection. Haig is unsentimental: 'These reforms were most unpopular. The older nobles disliked them because they curtailed the power and diminished the wealth of the provincial governors... They rendered their author more odious than ever to the Deccani faction, headed by Malik Hasan, who had been the first to suffer by them' [22]. Haig calls the new governments fairly divided; the eight he lists divide evenly only if the two African governors are counted apart from the Dakhanis, and counted with them - as the rest of the evidence here suggests they should be - the Dakhani bloc came out of the reform holding five of the eight [22]. What is not in dispute is that the reform cost Malik Hasan the undivided Telingana he had expected, and Malik Hasan is the man who then destroyed its author.
The chronicle version of his death is famous: conspirators got his seal-keeper drunk, stamped a blank sheet, wrote above the seal a letter inviting the Raja of Orissa to invade, and the intoxicated sultan, shown the seal, went no further before having him cut down. Iranica dates it: 'unfortunately his tenure was cut short when he was put to death by Mohammad III in Safar, 886/April, 1481, and buried at Bidar' [2]. But Walravens has read the documents written within months of the event - Bahmani chancery letters in Nimdihi's Kanz al-ma'ani, the chronicle of Tuni, al-Sakhawi's biographical dictionary - and the forged letter is in none of them. The charge the Bahmani state itself laid was corruption: 'The Bahmani letter to the Ottoman sultan explicitly mentions fasād as the reason for Maḥmūd Gāwān’s execution' [4], meaning the extortions of his revenue agents, for which a minister could be liable though personally innocent. The forged-letter plot appears only in the 1590s, in the Tabaqat-i Akbari, from which Ferishta inherits it: the factional conspiracy is a late-sixteenth-century settlement of an old score, not a contemporary report - which does not make it false, but does mean it must be argued rather than assumed [3].
Was it also a Shia-Sunni quarrel?
The standard account layers a sectarian divide over the factional one: Afaqis largely Twelver Shia, Dakhanis largely Sunni, the Friday sermon the flashpoint. There is evidence for it: the Chakan victims were disproportionately Sayyids, and Iranica ties the Bijapur cleavage to the newcomers' Shiism [17]. But Ferishta supplies the counter-evidence himself. When Yusuf Adil Shah proposed in 1502 to proclaim Twelver Shiism publicly at Bijapur, his advisers objected that 'most even among his own nobles were Toorks, Deccanies, and Abyssinians of the Soony persuasion', as were the neighbouring Deccan kings [11]. The leader of the Foreigners commanded a largely Sunni following.
Roy Fischel has argued the sceptical case at length; a review of his Local States in an Imperial World reports that on the widely held theory of a Shia-Sunni layer beneath this tension, 'Fischel disagrees with this' [24]. The same review records a further claim worth sitting with: that the Deccani-foreigner divide 'was the main source of tension among the elites of the time rather than the relations between Muslims and non-Muslims' [23]. The fault line that mattered ran inside the Muslim service class. A voice recovered by Peacock complicates matters from the other side: the Shirazi poet 'Iyani, himself a gharib, describes the revolt of the Abyssinian commander Dastur Dinar without the factional vocabulary at all - 'Although modern scholarship sees Deccani politics as riven between gharīban and Dakani factions, the later are nowhere mentioned. Rather 'Iyani describes the conflict in racially charged terms as a battle between the pale-skinned gharīban and the black Ethiopians' [15]. An eyewitness reaches for colour where later historians reach for party.
Partition: how the split made five sultanates
The Bahmani sultanate was not conquered; it was divided by its own provincial governors along the seam the faction had cut, and the identity of each founder is the argument. The sequence: 'The initial revolts of Yusuf Adil Shah, Malik Ahmad Nizam Shah I, and Fathullah Imad-ul-Mulk in 1490, and Qasim Barid I in 1492 saw the end of any real Bahmani power, and the last independent sultanate, Golkonda, in 1518' [26]. Bijapur went to Yusuf Adil Khan, leader of the Afaqis and in Ferishta's phrase a Turkish slave; Ahmadnagar to Ahmad, son of the Dakhani leader Malik Hasan; Berar to Fathullah Imad-ul-Mulk, the second Brahmin captive-convert of the Dakhani party; Bidar to Qasim Barid, a Turk; Golconda to Sultan Quli, a Turk of Hamadan [21]. They were converts and immigrant soldiers rather than a native aristocracy, and one acquired a grander ancestry afterwards: the Bijapuri histories make Yusuf a son of the Ottoman Murad II smuggled to Sava as an infant, a claim Iranica rejects because 'the non-Bijapuri sources fail to support this tradition, and epigraphic evidence suggests that Sāva was Yūsof’s ancestral home' [18]. The faction did not destroy the Deccan sultanate so much as reproduce itself five times.
Nor did the quarrel end with partition; at Bijapur it reversed. Yusuf's Shia proclamation was undone by his grandson Ibrahim Adil Shah I, who restored Sunni rites, moved the public accounts out of Persian into a local language under Brahmin clerks, and, in Ferishta's phrasing, 'Instead of foreign troops he enlisted Deccanies in his service, and permitted only four hundred foreigners to remain in his body-guard' [12]. The dismissed Persian cavalry were promptly hired by Vijayanagara: faction at this level was a labour market with an ideology attached. The physical archive is unusually good. Gawan's madrasa, founded at Bidar in 876/1471 and praised by Ferishta for its thousand rooms and its library of rare Persian and Arabic manuscripts, still stands in the old city [1], while the Chand Minar is on Manohar's reading the Dakhani-Habshi answer to it [16]. Chakan survives as a land fort in Pune district, where a reference entry notes that 'The two groups had an intense rivalry and a massacre of thousands of foreigners was carried out by the Deccanis at the fort in 1457' [25] - a date the chronicles do not support, and a reminder that even the calendar of this quarrel is unsettled.
Dates in this essay
Massacre of the Foreigners at Chakan: Ferishta in Briggs's translation and Wolseley Haig's Cambridge History narrative place the Sangameshwar disaster and the killing that followed in 1446, under Ala-ud-din Ahmad Shah Bahmani II; Walravens, from Firishta's Persian and Sherwani, dates the Konkan campaign to 850/1446-47 with the counter-convictions of Dakhani officers in 855/1451. The Wikipedia article on Chakan fort, citing M. S. Mate and M. S. Siddiqi, gives 1457 [wikipedia-chakan-fort]. The archive follows the chronicle chain at 1446 and flags 1457 as unverified against Mate's own wording.
Execution of Mahmud Gawan: Encyclopaedia Iranica gives Safar 886/April 1481 [iranica-bahmanid-gawan-death]; Walravens, from the contemporary documents, gives 6 Safar 886/6 April 1481; Wikipedia and Haig give 5 April 1481. Haig also makes Gawan seventy-eight at his death, which is irreconcilable with a birth around 813/1410-11 as Walravens and the reference works give it.
End of the Bahmani sultanate and start of the successor states: Iranica's BAHMANID DYNASTY entry runs the dynasty to 934/1528 under Kalim-Allah, its DECCAN entry only to 'the early 16th century'; Wikipedia dates formal dissolution to 1527 and the end of effective power to the revolts of 1490-92, with Golconda last in 1518 [wikipedia-bahmani-end]. For Bijapur, Iranica's ADELSAHIS entry dates Yusuf's khutba in his own name to 895/1490 and its BIJAPUR entry to 895/1489, while Yusuf and his son used the title Adil Khan rather than Shah, Ibrahim Adil Shah I formally claiming independence only in 1538.
Where authorities disagree about when a state began or ended, the archive follows one rule — a polity is dated from the assumption of sovereign power in its own capital, and to the settlement that extinguished it — and records the alternatives rather than hiding them.
Open questions
H. K. Sherwani's The Bahmanis of the Deccan (1985) is the standard authority on the Chakan massacre and on the factional carnage at Bidar in 892/1487, and both Walravens and Peacock cite it at length; no copy was opened here, so the 1487 Bidar killings are known to this entry only at second hand through Peacock's citation.
Roy Fischel's Local States in an Imperial World (Edinburgh, 2020), chapters 2-3, is the current standard study of the Dakhani and Foreigner categories, but only publisher matter and a review were read; what Fischel actually argues against the Shia-Sunni reading, in his own words, has not been verified.
Eaton's A Social History of the Deccan carries a table of prominent Westerners in Bahmani service between 1422 and 1458 which would give the prosopographical spine of the Afaqi bloc. Only the publisher's chapter abstracts on the book contents page were reachable, and the two Eaton quotations used here are taken verbatim from those abstracts rather than from the chapters themselves, whose full text is paywalled.
Mohit Manohar's Chand Minar article is paywalled beyond its abstract, so his detailed reading of Firishta's use of dakkaniyan and gharibān, and the minaret's inscriptions, remain unchecked here.
How many died at Chakan, and who they were, is reported differently in every rendering - Ferishta in Briggs gives about 2,500 Foreigners of whom 1,200 were Sayyids and 'every male' in the camp; Haig gives 1,200 Sayyids, 1,000 other foreigners and five or six thousand children. The Persian text has not been collated and these should be read as chroniclers' figures, not as counts.
Sayyid Ali Tabataba's Burhan-i Ma'athir (1004/1596), the Nizam Shahi counterpart to Ferishta and a witness with its own factional interest, is used here only at second hand through Peacock; no edition or translation was opened.
Yusuf Adil Shah's origin remains a transmission problem rather than a settled fact: the Bijapuri court histories make him a son of the Ottoman Murad II, while Iranica holds that 'the non-Bijapuri sources fail to support this tradition, and epigraphic evidence suggests that Sāva was Yūsof’s ancestral home' [iranica-adilshahis-yusuf]. Ferishta elsewhere calls him a Turkish slave, other accounts a Georgian slave or an Aq Qoyunlu Turkman. What the evidence bears is migration from Iran around 865/1460 and a dynasty with strong motive to acquire a royal pedigree.
Sadiq Naqvi, The Iranian Afaquies Contribution to the Qutb Shahi and Adil Shahi Kingdoms (Hyderabad, 2003), is the one monograph dedicated to the Afaqis; no copy or excerpt could be located online for this entry.
The later life of the faction was only glimpsed: Malik Ambar at Ahmadnagar as the last flowering of the Dakhani-Habshi bloc, and the Afghan-Abyssinian factionalism at Bijapur in the 1670s that Iranica names among the causes of the Adil Shahi collapse, were not pursued into a substantial source.
The obvious outside witness to Bidar in exactly the Gawan years is the Russian merchant Athanasius Nikitin, with Barbosa and Varthema a little later; none of these travel accounts was reached, and they would test the chronicles from outside the court.