The time layer · 9 states · 32 places · 57 dated events
Contested Ground
Pick a year and the Deccan rearranges itself around it, and the map says why — what happened that year is written where it happened, and every mark on the ground will tell you what it is. Everything drawn is something the archive can cite — which is why there are no borders on it, and why, where the authorities disagree about a date, the disagreement is what you see rather than something smoothed away. Nine states carry 16 rival years between them; 20 of the dated events below were printed more than one way.
Colour is confidence, not territory: how sure the archive is that a place answered to a capital. Bare ground is no claim, not unclaimed land. Coast and rivers are Natural Earth, public domain — physical geography, not political.Where this map fails: pushed far enough, it says only that some capitals are further apart than others.
Nine states, one frame
The same frame nine times, so the eye compares instead of remembering. Each panel holds one state’s whole life — every capital it held, and the dashed line between them when the capital moved. Read the leans against each other and the point arrives without a sentence: Golconda’s capital moved eight kilometres in a hundred and sixty-nine years; Bengal’s moved two hundred and ninety; the Nizams’ moved four hundred and thirty-one, and did it in forty.
Where the dates fray
Solid bar: the years this archive uses. Frayed end: every other year a published authority gives for the same beginning or ending. Bidar carries five start dates and never held a coronation to date from; the Bahmani sultanate has four endings. Click a lane for the argument.
Two gravitational years, and a hole
Scrub across the corpus and it is not evenly weighted. Nine of the archive’s twenty-seven essays turn on 1565, the year of the battle fought at Rakkasagi–Tangadigi and remembered under the name of a town eighty-five kilometres away. Twelve run through 1686 and 1687, when Bijapur and then Golconda fell to Aurangzeb. Fifty-four per cent of every year this corpus mentions falls between 1500 and 1699.
Then set the year to 1700 and the map empties. Between 1688 and 1723 there is no state here at all — Golconda has fallen and the Nizams have not yet arrived. That gap is not a rendering failure; it is the shape of what has been written so far, and it is the most useful thing this page can tell you about its own limits. The density ribbon under the map shows the same fact along the whole timeline.
Why there are no borders
Every crisp historical border you have seen is a cartographer’s reconstruction. The best openly-licensed dataset of historical polygons does draw Bijapur, Bidar and Ahmadnagar, and draws them well. It also has a median snapshot span of eight years, and gives the Bengal Sultanate one polygon for eighty-seven years. A year-by-year map over that would invent eighty-six years of false stability. Its own authors say the apparent stability of ancient borders “may reflect limited historical records rather than the actual stability of the state itself.”
More simply: none of this archive’s 934 citations supports a single border line, so none is drawn. Tim Wallace, then of the New York Times graphics desk, put the editorial version of it well — “You don’t want to put a hard line around that. It’s not like you cross a river and all of a sudden it’s sparsely populated.” Minard labelled his own maps cartes figuratives et approximatives, declaring the uncertainty in the title, in 1869.
What you see instead is a confidence field: colour thickens near a capital the archive can actually place, thins with distance, and speckles where two capitals have an equal claim. Ground with no colour is no claim — not unclaimed land, not wilderness, simply somewhere this archive declines to speak for.
The graininess is doing work rather than being an artefact. Where one capital plainly holds the ground the field is drawn finely; where the call is close, or the nearest capital is far off, it coarsens — a bigger block is the map saying it knows this ground less well. And each state is named across its own field rather than boxed by a line, which is how cartographers labelled territory long before anyone could draw a border with confidence.
Coast and rivers are drawn, from Natural Earth, which is public domain. A coastline is not a sovereignty claim — and here the rivers carry argument, since the Krishna–Godavari doab is the ground Bijapur and Golconda fought over and the Kollur mine, which supplied the stones the world called Golconda diamonds, sits on the Krishna.
The method has a citation
This is not a new idea, only an unbuilt one. Alan MacEachren proposed it in 1992: an area may be shown “not bounded at all, but as fading in a continuous fashion from core to periphery.” Mark Ravina built the nearest-capital version for Tokugawa Japan in 2023, with the caveat worth repeating — the cells are“like isobars on a topographic map,” an abstraction, never a reconstruction. Luca Scholz’s Deceptive Contiguity is the case against the polygon; David Ludden made it for South India, where historical maps showing states“spread evenly across the landscape are quite deceptive” and the reality was “an archipelago of sacred sites and centers of agrarian power strung along routes of mobility.”
The smaller choices have sources behind them too — why doubt is drawn as blur rather than as a change of colour, why the turning points lead instead of the slider, and why the nine panels above hold one state each rather than one decade each. None of it is invented for the occasion.
Where this comes from
9 states, 32 places and 57 dated events, drawn from 46 essays and 934 citations. Every span is the value adopted in the archive’s register of contested dates, and every coordinate records the page it was read from — the four that no essay cites yet are drawn as hollow rings rather than solid marks, so you can see which ones are standing on thinner ground. Nothing here is asserted for the first time: everything on this page is argued somewhere else, and linked.