Sometime early in his reign, the Vijayanagar emperor Devaraya II called his nobles together and asked a question no flatterer wanted to answer. His empire was larger than the Bahmani Sultanate to the north — bigger in land, bigger in people, bigger in revenue. So why, he wanted to know, had his predecessors been "reduced to pay them tribute"?1 A few of the advisers gave him the comfortable answer: Muslims, they said, held a divine mandate "for thirty thousand years" to rule the world, and there was nothing to be done about it.2 But others in the room had cooler heads. The Bahmani edge, they told him, was not destiny. It was horses. It was mounted archers — men who could ride and shoot at the same time, the most lethal battlefield technology of the age.3
What Devaraya II did next is the reason this page exists. He did not declare a holy war to reclaim Hindu dignity. He went shopping for the enemy's weapon. He enlisted Muslim cavalrymen into his own army, gave them estates to live on, built a mosque inside his capital so they could pray, and — by the account that comes down to us — set a Quran on a rich desk before his own throne, so that a Muslim soldier could make his obeisance to the Hindu emperor without sinning against his own faith.4 A Hindu god-king arranged his court so that his Muslim troops would not have to choose between loyalty and religion. That is the whole story in one image, and it detonates the tidy picture of Vijayanagar as a fortress of Hindu resistance against an Islamic north.
This is the documented moment when a Hindu emperor diagnosed a military problem correctly and solved it by adopting his rival's system whole — soldiers, tactics, and the religious accommodations needed to keep those soldiers. The chronicler Ferishta tells the tale, and an inscription from 1430 corroborates the core of it: Devaraya II absorbed a large body of "Turkish" mercenaries into his service.5 The numbers wobble between sources — the 1430 inscription speaks of 10,000 cavalry, while Ferishta himself reports a more modest initial induction of 2,000 men — but the direction is unmistakable.6 Vijayanagar imported the Bahmani military model rather than merely fighting it.
Three things travel together here, and they are easy to pull apart and weaken. First, the cavalry itself: trained mounted archers, the asset the Deccan ran on. Second, the religious infrastructure to retain them — the mosque "minus minarets and domes," which the city's later inhabitants knew not as a masjid but as a dharamsalai, just another rest-house, its true function half-forgotten even by the people living among its stones.7 Third, the trade policy underneath it all: Devaraya cultivated the west-coast ports and watched the import of warhorses from overseas, because cavalry is only as good as your ability to keep replacing the animals.8 The soldiers, the mosque, and the horse-trade are one system. Pull the religious tolerance out and the soldiers leave; cut the port economy and the horses die without replacement.
The logic running under Devaraya's decision is the logic of the better learner, not the truer believer. His advisers had handed him two explanations for Bahmani dominance — a metaphysical one (divine mandate) and a mechanical one (mounted archers). The metaphysical answer is a trap, because it tells you to either submit or wait for the cosmos to change. The mechanical answer is a gift, because a mechanism can be copied. Devaraya took the gift.
But copying a weapon is never just copying a weapon. A mounted archer is a man, and the best of them in the fifteenth-century Deccan happened to be Muslim — Turkish and Persian émigrés steeped in a riding-and-shooting tradition that South Indian warrior castes had not built. To recruit the technology you had to recruit the men, and to keep the men you had to make the court survivable for them. The Quran before the throne is not a gesture of personal piety; it is a retention policy. It removes the single friction that would make a devout Muslim soldier refuse to serve a Hindu king. Devaraya understood that the religious accommodation was not a cost of the military upgrade but a part of it — the part without which the rest does not function. None of this, Pillai is careful to note, was even unprecedented: three hundred years earlier the Hoysalas had employed Muslims in their armies, though those earlier recruits came from coastal settlements and were most likely of Arab descent, unlike the later Turkish and Persian waves.9
This page hands the vault a hard counter-example to the communal reading of Deccan history, anchored in a single concrete act. Whenever a sibling page argues that the Vijayanagar–Bahmani conflict was a war of religions, this one supplies the rebuttal in one sentence: a Hindu emperor built a mosque inside his Hindu capital to keep his army fed with mounted archers. It also emits a transferable mechanism — the borrowed edge, where a polity beats a rival by adopting the rival's own decisive technology plus the social arrangements that technology requires — which the behavioral-mechanics and Maratha pages can plug into directly. And it deposits the horse-trade thread (ports, overseas imports, the scramble for mounts) that runs all the way down to Krishnadeva Raya and the Portuguese, ready for the trade-economy page to pick up.
The single most revealing detail in the whole account is architectural. Among the ruins of Vijayanagar stood a pillared pavilion long taken for a rest-house. Only later did it prove to be a mosque — a mosque "minus minarets and domes," stripped of the features that would have announced it as Islamic, and known to the city's own inhabitants not as a masjid but simply as a dharamsalai.10 Sit with that. A house of Muslim prayer, raised by order of a Hindu emperor, designed quietly enough that the people who walked past it for generations filed it under the most ordinary category they had: a place to rest.
That building is the cosmopolitan thesis made of stone. It tells you the accommodation was real enough to be built and permanent enough to outlast the empire, yet undemonstrative enough that it dissolved into the everyday fabric of the city. There were no minarets crying the difference; there was a structure doing a job. The forgetting is itself the evidence. A society in a state of holy war does not lose track of which buildings belong to the enemy's god. Vijayanagar did, because the men who prayed there were not the enemy — they were the cavalry, granted estates, granted protection in "the exercise of their religion," folded into the working machinery of the state.11
Ferishta, who narrates the tale, is not a neutral witness, and the case study has to say so. He "generally tends to flatter Islamic kings," and there is a conceit in his telling — a hint that Muslim horsemen possessed some inherent superiority.12 A reader inclined to suspicion could dismiss the whole story as Muslim-chronicler self-congratulation. The inscription of 1430 is what saves it: independent of Ferishta's spin, it records the 10,000 cavalry the king absorbed.13 The spin and the fact can be separated. Strip away Ferishta's flattery and a stubborn core remains — Devaraya did enlist a large Muslim cavalry force, and Vijayanagar did make a home for them.
The episode also carries its own dark coda, which keeps it from becoming a fable of tolerance. Bhatkal, a west-coast port, was home to a large community of Muslim traders; decades after Devaraya, a "hot-headed" successor decided to massacre a good portion of them for selling to the wrong buyer, and the survivors fled to Goa — where Mahmud Gawan promptly scooped them up to secure his own Sultanate's horse supply.14 Tolerance here was a policy, not a principle. It held while it served the state and broke when a successor's temper or a rival's opportunism overrode it.
Stand in the great hall as the morning light comes in. The emperor is on his throne, and before him, set on a rich carved desk, lies a Quran. A Turkish cavalry officer crosses the stone floor — a man whose grandfather rode the steppe, whose bow can put an arrow through a man at a gallop. He has come to make his obeisance to a Hindu god-king whose priests anoint him as something close to divine. By every communal logic this should be impossible: the soldier should choke on the submission, or the Brahmins should choke on the soldier.
Neither happens. The officer bows toward the book that holds his own law, and in bowing toward it he renders his service to the emperor who placed it there. No line has been crossed in his own faith; the king has built the bridge for him. Off the hall, in a pavilion without minarets, his comrades will pray when the hour comes. Somewhere down on the coast, an agent of the crown is watching a ship from Hormuz unload horses, counting heads, because the man bowing in this hall is worth nothing without a fresh mount under him next season. The durbar, the mosque, and the harbour are one machine, and you are standing inside the gear that lets a Hindu empire run on Muslim cavalry without tearing itself apart.
The pattern goes wrong in two directions, and both are visible in the source. The first failure is misremembering it. The "enduring picture of Vijayanagar as a Hindu" bastion is the failure-mode of memory — the later, cleaner story in which a Hindu empire stood against Islam, with the mosque-in-the-capital and the Muslim cavalry quietly edited out.15 The diagnostic sign is any account of the Deccan that runs on religious sides and cannot explain why an emperor would build a mosque for his own troops. When a narrative has no room for the dharamsalai, the narrative has been laundered.
The second failure is in the practice itself: tolerance held as policy rather than principle is only as stable as the next ruler's judgment. Devaraya's accommodation worked; his "hot-headed" successor's massacre of the Bhatkal traders shows the same system thrown into reverse, and the merchants fleeing to enrich the rival Bahmanis shows the cost.16 The mobility that made the elite so adaptable — nobles "moving from patron to patron according to changes in political winds," not according to professed religion — cut both ways.17 A talented Muslim regent could rise to senior administrative rank in Vijayanagar; he could also, in a later case, permit his Muslim subordinates to eat beef over Brahmin protest, and it was during that very regent's rule that the empire collapsed.18 Pillai records the coincidence without claiming the cause — and that restraint is itself a diagnostic for how not to read the episode.
The evidentiary spine is two-stranded: Ferishta's narrative and a corroborating inscription of 1430. Pillai's method is to lean on neither alone. He flags Ferishta's pro-Islamic tilt openly — the chronicler flatters Muslim kings and hints at the innate superiority of Muslim horsemen — and then uses the independent inscription to rescue the factual core from the rhetorical frame.19 The numbers remain genuinely unresolved: 10,000 cavalry by the inscription, 2,000 by Ferishta's report of the initial induction. These are not necessarily contradictory (initial induction versus eventual total), but the page should not pretend the figure is settled.20
The deeper tension is interpretive. Pillai is arguing against a strong communal narrative of Deccan history, and the Devaraya episode is one of his sharpest weapons. The risk is that an author marshalling evidence against the communal story will overstate the cosmopolitanism — read every accommodation as principled tolerance rather than as the cold pragmatism it often was. Pillai mostly avoids this, and the Bhatkal massacre is his own evidence that he is not selling a fairy tale. The open question that survives: how much of this fifteenth-century pragmatism was genuinely shared cosmopolitan culture, and how much was simply that horses won battles and everyone needed the men who could ride them?
Pillai's whole frame in Rebel Sultans is set against what he calls the "enduring picture" — the inherited communal narrative in which Vijayanagar is the Hindu wall and the Bahmanis are the Islamic flood. The Devaraya episode is where he plants his flag, and his tension is less with another vault source than with that received story itself. He is using a hostile witness, Ferishta, against the witness's own grain: Ferishta wrote to flatter Muslim kings, yet Pillai turns Ferishta's tale of a Hindu emperor adopting Muslim cavalry into evidence for a shared military-political culture, not Muslim supremacy. The convergence that holds the argument together is the inscription of 1430, which lets Pillai keep the fact while discarding the chronicler's conceit — a clean instance of his cosmopolitan thesis disciplined by epigraphic corroboration rather than asserted on faith. Where the page meets the vault's Laine material on Shivaji, the same instinct surfaces from the other end: Laine studies how a Hindu warrior-king gets retrospectively recruited into a communal legend, and Pillai shows the fifteenth-century reality that such legends paper over — a Deccan elite that moved by political wind, not by professed creed.
Here is the plain version: Devaraya won by stealing the enemy's best weapon and the social setup that weapon needed — and that move shows up across the vault wherever a weaker or stuck player upgrades by borrowing rather than resisting.
The first handshake is to Army Composition: Heterogeneous Force as Adaptive Weapon. That page treats the deliberate mixing of fighting elements — different troop types, different traditions — as a designed adaptive advantage rather than an accident of recruitment. Devaraya is a textbook case with a twist the behavioral-mechanics frame sharpens: the adaptive force here required not just military integration but religious integration. The Quran before the throne is the retention mechanism that makes the heterogeneous force cohere. Set against the abstract principle, the Devaraya case adds a precise demand — a heterogeneous force composed of men whose loyalty conflicts with their faith will dissolve unless the patron engineers away the conflict. The insight that lands only when you hold both: the limit on borrowing your rival's elite troops is not military but social, and the polities that out-recruit their neighbours are the ones that solve the social problem first.
The second handshake is to Protector-Conversion Syncretism. That page tracks how a tradition absorbs an outside power by converting or domesticating it — folding a foreign or threatening force into the home framework until it serves rather than threatens. Devaraya's mosque-as-dharamsalai is the secular-political mirror: an alien religious presence is not converted but quietly domesticated, housed in a building stripped of its announcing features until the city files it under an ordinary local category. The tension between the two is the productive part — the spiritual page describes absorption that changes the absorbed thing's allegiance, while the Devaraya case describes accommodation that leaves the absorbed thing intact (the Muslim soldier stays Muslim, prays in his own way) and changes only its function within the host state. Reading them against each other surfaces a distinction neither makes alone: syncretism can run by conversion (change what they believe) or by accommodation (leave the belief, capture the use), and a state that needs the foreigner's skill more than his soul will reliably choose accommodation.
The third, lighter handshake is to Bhakti as Path — and here I will name the stretch. Bhakti's logic is that devotion, not caste or ritual purity, is the qualifying credential before the divine; the Devaraya court runs a worldly analogue where battlefield skill, not religious origin, is the qualifying credential before the emperor. The parallel is structural and loose: both dissolve an inherited gatekeeping criterion (purity / creed) in favour of a functional one (devotion / competence). It earns its place only as a faint echo, a reminder that the fifteenth-century Deccan's pragmatism about who could serve had a devotional-era cousin in who could approach God.
Sharpest implication. The single most durable picture in popular Deccan history — Hindu Vijayanagar versus Muslim Bahmanis — is contradicted by a building inside Vijayanagar's own capital. Devaraya II's mosque is not a footnote to the communal story; it is the disproof of it, and the fact that the city itself forgot the building was a mosque tells you how natural the accommodation had become.
Generative questions.