Long before anyone called it the City of Victory, Devagiri was a fortress the poets sang about as the abode of the gods. The poets were being generous. What sat on that rock in the Marathi country was less a heaven than a vault — the seat of the Yadava kings, whose treasure rooms were so famous that talk of them carried hundreds of miles north, to the ear of a Muslim prince who liked the sound of it.1 When the poet-chronicler Abdul Malik Isami later described the city, he did not reach for the divine. He reached for an inventory: "Every house possessed diamonds and was full of silver and gold... In every street gold could be found in unlimited quantities... and where else can be had such variety of cloth except at Devagiri?"2
That was the Deccan in 1296: three great Hindu kingdoms, fat with generations of accumulated wealth, ruling three different tongues, and busy feuding with each other over backwaters.3 The Yadavas held the Marathi lands from Devagiri. The Kakatiyas held the Telugu east from Warangal. The Hoysalas held the Kannada south from Dwarasamudra.4 Each was sophisticated; each was rich; none of them was watching the horizon. Within fifty years all three would be gone — one king starved out by a ghost army, one dead by his own hand on a riverbank, one flayed and hung from a fort wall like a warning poster. This page is about why three kingdoms that could each have fought the invader chose instead to keep feuding with each other until there was nothing left to feud over.
This is the portrait of the pre-conquest Deccan — the world the Khiljis broke — and the first appearance of the thesis that runs through the whole of Pillai's Rebel Sultans like a fault line: that the people of this land lost not because they were weak, but because they would not unite.
Three dynasties anchor the picture. The Yadavas of Devagiri, lords of the Marathi-speaking country, ruled the kingdom Ferishta's chronicle calls the land of "the Marhat."5 The Kakatiyas of Warangal ruled the Telugu east — "the Tiling" — under a king, Prataparudra, who ran what was for its time a startlingly open society.6 The Hoysalas of Dwarasamudra held the Kannada south — "the Kanhar" — from a seat that would later be sacked into rubble.7 Ferishta, writing centuries later and reaching for a tidy origin, claimed all three peoples descended from a grandson of the biblical Noah conveniently named Dekhun.8 [FOLK ETYMOLOGY]
What matters is not the genealogy but the geometry. Three rich houses, three languages, one shared blind spot. They had spent so long warring over "some backwater or the other for as long as could be remembered" that the idea of a coordinated defence against a common enemy never occurred to them — and by the time it might have, the enemy was inside the gates.9 These three linguistic zones — Marathi, Telugu, Kannada — would survive the dynasties that ruled them and become the Deccan's enduring cultural map, the lines along which the region still organises itself.
The mechanism that doomed the three houses was not military inferiority. It was the refusal to add up.
Watch the fall of Devagiri, the first to go. Alauddin Khilji rode south in 1296 with 8,000 cavalry — a modest force — and met "practically no resistance" the whole way.10 At a single point, a lone chieftain backed by two unnamed Maratha women "who fought like lionesses" tried to stand in the road, and failed.11 One chieftain. Two women. That was the entire field resistance of the richest kingdom in the south, because the Yadava king Ramachandra had sent the bulk of his army off with his son toward the Hoysala frontier — to feud, of course, with one of the two houses that should have been his allies.12 When Alauddin arrived, the king shut his gates and waited out a siege he had never imagined could come, in a fortress so strong that "the only way to it being so narrow that but one horse, or one camel, can pass at a time."13 The fort was impregnable. The problem was that nobody had stocked it, because an invasion from the north "was the last thing Ramachandra expected." A week of hunger was enough to make him think about peace.14
His heir nearly saved it — sped back, nearly drove the invaders off — until Alauddin did something cheaper than fighting. He had 1,000 horsemen raise a "whirl of dust in the distance" and let the rumour spread that 20,000 fresh cavalry were arriving.15 The Yadavas believed the canard. A ghost army behind a blur of earth and wind brought Devagiri to its knees, and the treasure heaped before Alauddin was so vast that, sixty years later, portions of it were still said to sit untouched in the imperial vault.16 The internal logic is brutal and simple: each house guarded its own gold and its own grudges, and so each could be picked off alone, by a smaller force, by a trick.
This page seeds the foundation stone the rest of the Deccan material is built on. It establishes the before — the rich, fractured, pre-conquest order — so that the Khilji Conquest of the Deccan reads not as a clash of religions but as the demolition of houses that would not stand together. It supplies the deep-history floor for the Maratha story in The Four-Kingdom Deccan and the Rise of the Marathas: the Yadava Marathi country is the same soil Shivaji's people would later hold, and the disunity that doomed the Yadavas is the same disunity Pillai watches doom the Rebel Sultans three centuries on.
It hands the vault its earliest documented case of the Deccan Disunity Doom Thesis — the through-line, here in embryo — and it sets up Prataparudra and Kakatiya Meritocracy by introducing the king whose open society makes his fall the page's sharpest tragedy. The dust-army trick also emits straight into the vault's deception-and-asymmetry corpus: a smaller force beating a richer one by manufacturing a phantom threat is the Khilji ancestor of the Guerrilla Paradox: Weak Beats Strong.
No figure in this stretch carries more weight than Prataparudra, the last Kakatiya king of Warangal, because his story is the one where the loss most clearly was not deserved.
He was, by the record, a genuinely good ruler over a genuinely unusual kingdom. Kakatiya Warangal shows up in the inscriptions as one of medieval India's most mobile societies: nearly a third of the inscriptional evidence has fathers and sons holding different occupations, "free from rigid compulsions of caste and birth."17 The dynasty itself, almost alone among Indian ruling houses, refused to fabricate a glorious genealogy from the sun or moon or some legendary hero — "with only one exception, they embraced shudra status," claiming no exalted origin at all.18 They built thousands of irrigation tanks for their people. And Prataparudra pushed the openness further: by Richard Eaton's count, 45 per cent of his noblemen came from humble origins and only 12 per cent had inherited their place at court by blood.19 This was, for its century, something close to a meritocracy.
It did not save him. The fall came in stages, each more humiliating than the last. In 1310 Malik Kafur breached Warangal's walls; Prataparudra sued for peace and was spared on terms — submit, hand over wealth and elephants, and receive in return the Islamic khilat, the jewelled robe of honour, and the parasol of state sent from Delhi.20 With that robe, court culture from the north entered the southern world for the first time, and a king of the soil placed himself, dutifully, under another man's umbrella.
In 1318 he tried to wriggle "out into the sun." The punishment was ceremonial and cruel. He paid again in mountains of gold, but this time he had to climb the "eighteen steps leading up to the parapets of the citadel's stone wall," and there, in full view of his own Telugu warriors and the invading northerners, turn his face toward Delhi, bow slowly, and kiss the rampart in a gesture of submission.21 That image — a beloved king kissing his own wall while his people watch — is the whole tragedy compressed into one scene. He tried once more in 1320. The Tughluqs came back in 1323 with 63,000 mounted archers, took the fort, razed his family temple (a mosque rose in its place), and dragged him toward Delhi in chains.22 Fearful of what waited for him there, the last good king of the Kakatiyas killed himself on the banks of the Narmada, and the flame of his dynasty went out.23
You are standing on the parapet of Warangal's citadel, and the wind off the Telugu plain carries the smell of cooking fires from the camp of the northern army below.
Behind you wait your own warriors — men whose fathers were farmers and potters, men your house lifted on merit, men who love you. In front of you, down the slope, wait the men of Delhi who have already taken your gold once and have come back for more. You have climbed eighteen steps to get here, and every step has been counted by both armies watching. You feel the rough cut stone of the rampart under your hands. You know what comes next, because the terms left no room for invention. You turn your face north — toward a capital you have never seen, toward a Sultan who calls himself the shadow of God — and you bow. Slowly. You bend until your lips touch the cold stone of your own wall, the wall your ancestors raised, and you kiss it.
Nobody fires an arrow. Nobody from the other two kingdoms rides over the hill to break the siege, because the other two kingdoms are not coming. The Yadavas are already broken. The Hoysalas are minding their own ruin. You are the open society, the meritocracy, the good king — and you are entirely, structurally alone, kissing a wall while two armies watch you do it. That is what disunity feels like from the inside: not a battle lost, but a humiliation performed in solitude, with allies who never existed.
The failure-mode this episode embodies has a clear set of warning signs, and the three doomed dynasties show all of them.
The first sign is misplaced confidence in walls. Devagiri was one of the best fortresses in India, so narrow a single camel could pass at a time — and it fell in a week because the larder was empty. Strong defences breed the assumption that defence is enough, and the assumption rots the will to prepare.24 The second sign is the grudge that outranks the threat. Ramachandra Yadava had his main army on the Hoysala frontier when Alauddin arrived, busy with a feud, because old enemies feel realer than new ones.25 The third sign is premature boasting. As early as 1278 Ramachandra had compared himself to the Great Boar rescuing the earth "from the oppressions of the Turukas" — decades before he had encountered a single Turk.26 [SINGLE SOURCE] The rhetoric of strength ran years ahead of the substance, and the gap between them was exactly the gap Alauddin's dust-cloud exploited.
The deepest sign is picked off one at a time. The Kakatiyas fell after the Yadavas; the Hoysalas after the Kakatiyas. At no point did the three coordinate. When Muhammad bin Tughluq's cousin rebelled, the Hoysala king Ballala III sheltered him — and so invited Delhi's wrath onto his own realm, a kindness that doomed him the way the Yadavas' isolation doomed them.27 The failure is also a failure of memory: read as a "clash of civilizations" or "a feud between two antagonistic religions," the fall becomes a story about Hindu versus Muslim. Pillai's whole point is that the truth "is a little more complicated" — Islam had arrived peacefully on the coast by the tenth century, Muslim traders were "honoured and protected" in the Rashtrakuta kingdom, and even the eighteenth-century Brahmin Peshwas would call a Turkic sultan a just protector of the earth.28 The diagnostic failure is to mistake a story about disunity for a story about religion.
Pillai's evidence is a mosaic of partisan chroniclers, and he knows it. Isami's catalogue of Devagiri's diamonds and Amir Khusro's claim that "no measure or balance would suffice" to count the plunder are court poetry written to flatter conquerors — "a forgivable touch of poetic excess," Pillai concedes of Isami.29 [SINGLE SOURCE] The wealth was surely real; the figures are surely inflated. Ferishta's Noah-descended "Dekhun" is a genealogist's invention, useful for showing how later chroniclers tidied the Deccan into a single ancestry it never had.30 The dust-army rumour and the "20,000 fresh cavalry" survive in chronicle form and may themselves be a literary device dressing up a more ordinary collapse.31
The sharpest tension is interpretive, and it is the one Pillai is arguing against the whole time. The dominant popular reading casts the Khilji descent as a totalizing clash of civilizations. Pillai answers with the cosmopolitan counter-evidence — peaceful early Islam, protected traders, Sanskrit inscriptions casting the foreign sultan as Vishnu's representative "sleeping in peace on the ocean of milk."32 But he cannot fully escape the violence in his own sources: temples razed and mosques raised in their place, Prataparudra dead on the Narmada, a king flayed alive. The honest position the page holds open is that both are true at once — the encounter was neither a painless communion nor a simple religious war, and "no battle was ever a painless communion."33 An open question worth keeping: did the three houses ever attempt coordination, or is their isolation a fact, an artefact of hostile sources, or both?
Pillai writes this opening against a ghost — the communal narrative that turns the Deccan's medieval history into a two-religion grudge match. His method is to crowd that narrative out with detail it cannot absorb: the Rashtrakuta Muslims who grew their beards like their Hindu neighbours, the Sanskrit inscription that makes Sultan Balban into Vishnu, the Peshwas calling a Yavana king a just protector. The tension inside his own framing is visible and he does not hide it. He wants the period to be cosmopolitan, a "patchwork of principalities" absorbing newcomers as it always had — but the same chapter has him reporting a king kissing his own ramparts in forced submission and another flayed and stuffed with straw. Pillai's resolution is not to deny the violence but to refuse its religious framing: the brutality is statecraft, the language of conquest in any age, not the special signature of one faith against another.
Set against the vault's Laine material, the convergence is methodological. In Shivaji: Legend as Interested Knowledge, Laine shows how later partisans built a Shivaji to serve their politics; Pillai is doing the upstream version of the same move, dismantling the clash-of-civilizations legend that later partisans built over the Deccan's actual, messier past. Both treat the inherited story as a thing made by interested hands rather than a record of what happened. Where Laine excavates one constructed hero, Pillai excavates a constructed villain-narrative — and both insist the construction is itself the data worth studying.
Here is the plain version: a smaller, poorer force beat three rich kingdoms by attacking them one at a time and by making them believe in an army that wasn't there — and that is a pattern that shows up far outside thirteenth-century India.
The first handshake is to Guerrilla Paradox: Weak Beats Strong. Alauddin's 8,000 cavalry should not have toppled the wealthiest kingdoms of the south; they did it because the south was divided and could be fought in sequence rather than in sum. The structural parallel is exact — strength is not a property of a side but of a coalition, and a coalition that never forms is weaker than the smallest force that does. The dust-cloud at Devagiri is the precise mechanism the guerrilla paradox depends on: the weaker party wins by controlling what the stronger party believes, not by matching its numbers. The insight this produces, that neither page reaches alone, is that the famous Maratha asymmetry three centuries later was first practised on the Deccan by its conquerors before it was practised by the Deccan against later conquerors — the trick was learned where it was first suffered.
The second handshake reaches to Assassins (Nizari Ismaili). The Nizaris, a small sect surrounded by vastly larger powers, survived not by armies but by reputation — by making far stronger enemies believe a single dagger could find them anywhere, so that the belief in the threat did the defensive work that numbers could not. Alauddin runs the same play in reverse: where the Nizaris manufactured a phantom threat to deter the strong, Alauddin manufactured a phantom reinforcement to demoralise the strong. Both are victories of staged perception over material force. The tension that sharpens the parallel is the direction of the lie — the weak Nizaris fabricated danger to survive, the ambitious Khilji fabricated overwhelming strength to conquer — and seeing them together shows that the manipulation of an enemy's belief is a single tool that cuts equally for defence and for predation, which neither the military nor the spiritual page would conclude on its own.
A third, shorter handshake runs to Maratha Guerrilla Warfare Doctrine. The Yadava Marathi country that fell to a dust-cloud in 1296 is the same hill country whose people would later turn terrain, mobility, and the enemy's own assumptions into a doctrine of asymmetric war. The doom of the Yadavas and the rise of the Marathas are the two ends of one geography: the same land taught two opposite lessons to two different generations, and the second generation learned, precisely, not to sit behind a wall and wait.
Sharpest implication. The three doomed dynasties are the founding case of Pillai's thesis, and the thesis is colder than it first looks: the Deccan was not conquered because the north was stronger but because the south would not add itself together. Wealth, sophistication, even an open and meritocratic society were no defence against the simple arithmetic of being fought one at a time. The lesson the land taught — that disunity is the true enemy — would have to be taught again, to the Rebel Sultans and to everyone after them, because it is the one lesson the Deccan never seems to keep.
Generative questions.