You will have three cities. You will have Bombay, arriving — Rushdie's midnight, Mistry's tenements, Chandra's police procedural the size of a phone directory, Mehta's non-fiction doorstop that half the country treats as a novel anyway. You will have Calcutta, ending — the same city in permanent elegiac twilight across a century of Bengali writing that begins mourning it before it has anything to mourn. And you will have Delhi, burning, in 1857 and 1947 and 1984, three dates that between them organise almost everything the capital has been allowed to be in fiction.

Three cities. In a country with more than fifty urban agglomerations over a million people.

Madras is not on that list, though a Tamil novel about the 1969 water crisis is one of the finest urban books written in this country in the last sixty years. Hyderabad is not on it, though the definitive novel of the Nizam's dissolution exists and has existed since 1977. Lucknow slipped off it sometime after Sharar. Kochi, Patna, Guwahati, Bhopal, Nagpur, Kanpur, Coimbatore — nothing, or nothing that travelled. Gurgaon has almost two million people and a skyline visible from an aeroplane and no novel of consequence at all.

This feature started as a straightforward comparison — how Indian fiction uses Bombay, Calcutta and Delhi differently — and fell apart under its own premise about a week in. Because the interesting question is not how the three cities are written differently. It is why there are three. The answer is not that nothing was written elsewhere. Ashokamitran wrote Madras. Sundara Ramaswamy wrote the small towns of the south. M. Mukundan wrote a Delhi that no English novelist has come close to — in Malayalam, over four decades of living there, and it took nine years to reach English and eleven to win the country's largest literary prize. The work exists. It simply arrived late, in a language the canon was not reading, at a point when the canon had already been assembled without it.

So the three-city map is not a map of Indian urban writing. It is a map of the translation economy — of what got carried across, when, by whom, and into which language. Bombay is the most written Indian city in the world because Bombay is the city that produced the most writers in English. That is a fact about publishing, not about Bombay.

A city does not become literature by being lived in, or even by being written. It becomes literature by being read somewhere else.

When he finally did publish — Kala Ghoda Poems and Sarpa Satra, both in 2004, both pushed out by friends because he had been told he was dying of cancer — he had the spines printed blank. No title. No author's name. No publisher's logo. A man who had spent half a century looking at one street in Bombay produced the definitive book about it and then designed it so that it could not be identified on a shelf.

There is your Bombay problem in a single object.

Because the other Bombay book of that era is Midnight's Children, and everything that followed it, and the shape of that literature is not the shape of Kolatkar's. Rushdie's Bombay is a city being arrived at — Saleem's pickle factory, the Methwold estate, a boy inheriting a metropolis and a country in the same breath. Rohinton Mistry's Bombay is a city being endured from inside a Parsi apartment block, written from Toronto. Vikram Chandra's Bombay is nine hundred pages of a policeman and a gangster circling each other. Suketu Mehta returned after twenty-one years in New York and produced Maximum City, which is reportage and reads like an epic and has done more to fix the international image of the city than any novel.

Every one of those books is superb. Every one of them is also, structurally, a book about arriving somewhere — or about having left and come back, which is arriving with better sentences.

That is the English-language Bombay novel's native form. The migrant-arrival narrative. The city as the place you get to, are overwhelmed by, and then attempt to render for somebody who has not got there. It is a genuinely great tradition. It also produced, almost as a side effect, the impression that Bombay is the most literary city in India — when what Bombay actually was, for reasons of commerce and Anglophone education and a film industry that paid writers, the city that produced the largest concentration of Indians writing English prose.

There is a second thing going on underneath that, and it is about who was in the room. Rushdie had Rugby and Cambridge. Mistry emigrated to Toronto in 1975 and has written Bombay from Ontario for fifty years. Mehta left at fourteen for Queens, New York, and came back at thirty-five with a reporter's notebook and an outsider's licence to ask rude questions, which is exactly what makes Maximum City work. Chandra teaches at Berkeley. This is not a complaint — the distance is productive, and half the great city books in any literature were written from somewhere else. But it does mean that the most visible body of Bombay writing was produced, in large part, by people looking back at it across an ocean.

Now go two streets over.

Namdeo Dhasal published Golpitha in 1972 — the year he founded the Dalit Panthers, modelled on the Black Panthers, at twenty-three. He had been a taxi driver. He had grown up in Kamathipura among pimps, sex workers, dealers and the rest of the trade that the city's respectable literature had spent a century declining to look at. The poems came out in a Marathi nobody had put on a page before: street register, caste slur, dialect, obscenity, and then a sudden swerve into European reference in high literary Marathi, all inside a single line.

It detonated. Marathi literature has not entirely recovered, and did not want to.

Dilip Chitre — poet, translator, the man who had already carried Tukaram into English — spent years working on Dhasal and finally published a selected volume with Navayana in 2007. Thirty-five years after Golpitha. Chitre described the translation problem plainly: the shifts of register are the poetry, and English has nowhere to put them.

His wife Viju put the reception problem more usefully than any critic has. Everybody walks past Golpitha every day, she said, and manages not to see what is there, and manages further not to feel it.

Note what she is describing. Golpitha is a book. Golpitha is also a place — a lane off Kamathipura, roughly eight hundred metres from the art galleries and the Wayside Inn where Kolatkar sat watching pi-dogs. The most anthologised Bombay poet in English and the most incendiary Bombay poet in Marathi were working within a twenty-minute walk of each other, in the same decade, on the same city. One of them is on university syllabi in three continents. The other required a thirty-five-year wait and a small activist press.

Dhasal got the Sahitya Akademi's only Lifetime Achievement Award of its golden jubilee year, in 2004. He is still, in English, essentially one book.

And Kolatkar is the hinge, because Kolatkar did both. He wrote in Marathi and in English and moved between them so completely that some of his English poems are labelled as the poet's own versions of Marathi originals that were never actually written down. There is no Marathi manuscript. The "original" existed only in his head, and what reached paper was the translation of a thing that had never been a thing.

He was not alone in that. Vilas Sarang worked the same way. A generation of Bombay writers moved between Marathi and English so freely that the concept of an original text starts to wobble in your hands.

And the machinery they came out of was not the publishing industry at all. It was the little magazine — the Marathi laghu-niyatakalike movement of the fifties and sixties, cyclostyled, unfunded, quarrelsome, printed in editions of a few hundred and read by everyone who mattered. Kolatkar ran one called damn you: a magazine of the arts, and it is where 'the boatride' first appeared in 1968. Dilip Chitre's Anthology of Marathi Poetry 1945–1965, published in 1967, is where a good deal of this work first surfaced in English at all. That is the actual infrastructure of Bombay's other literature: two poets, a duplicating machine, and an anthology.

Consider what that does to the premise of this whole feature. We have been talking about Indian writers as though each one has a language and the question is whether the work crosses. Kolatkar's practice says the border is not where we have drawn it. He was not a Marathi poet who also wrote English, or the reverse. He was a Bombay poet, and Bombay is a place where a sentence can start in one language and finish in another without anybody in the room finding it remarkable.

The city's actual condition is bilingual to the point of being uncountable, and its two literatures ended up in almost completely separate reading economies.

Which is the thing the arrival novel cannot see, and it is not a failing of the novelists. If your subject is the experience of coming into a city, the city necessarily appears as a surface — dense, dazzling, resistant, to be decoded. That is what arrival feels like and the books are honest about it. But a surface has an outside and an inside, and the writer who was already inside does not experience the city as a surface at all. Dhasal is not decoding Kamathipura. He lives there. There is no wonder in it. There is no first sighting of the skyline from a taxi.

Golpitha has almost no descriptions of Bombay in it, in the sense that a reader of Maximum City would recognise as description. Why would it. You do not describe your own street to yourself.

Kolatkar spent fifty years describing his, which is what happens when a man trained as a painter at the J.J. School sits in the same window long enough to see the place become strange again. The pi-dog who opens the Kala Ghoda sequence traces his ancestry to an imported English foxhound on his mother's side and to the dog that followed Yudhishthira on his father's. It is the joke of the entire city, told by a dog, in six lines.

None of this is a story about English versus the vernacular, incidentally, because the mill district settles that. Girangaon — the village of mills — held something like a quarter of a million textile workers and the densest concentration of chawls in the city, and it produced its own literature: Marathi, working-class, tied to the unions and the Shahiri song tradition and later to Dalit writing. Then the strike of 1982 broke the industry, the mills closed, the land became malls and towers, and the neighbourhood that had generated all of it was dispersed across the far suburbs within a generation.

There is no great English novel of the Bombay mill strike. There is a great deal of Marathi writing about it. You can see the problem.

Two poets. One decade. Eight hundred metres.

He had been unemployed for years at a stretch. He had lost a college lectureship in the thirties for being the most junior man on a payroll that needed trimming, and never really recovered from it. Most of what is now considered the greatest Bengali poetry after Tagore was published posthumously or in editions nobody bought. Rupasi Bangla — the book that gave him the permanent epithet, the Poet of Beautiful Bengal — was written in 1934 and did not appear until 1957, three years after a tramcar killed him.

The tram. Hold on to the tram, because it is going to do a great deal of work here.

Calcutta got horse-drawn trams in February 1873, the first in Asia, and electrified them in 1902. At the network's height there were more than three hundred and forty cars covering the whole city. By the 2020s the number in service had fallen to something under ten, running on two or three routes, out of a single surviving depot at Gariahat where a conductor of thirty-six years' standing still checked a duty roster pinned to a board. In 2024 the West Bengal transport minister announced that the service would be shut down except for a short heritage stretch, on the grounds that slow-moving vehicles were causing traffic snarls. Protests followed. So did a court case, and a stay, and an affidavit, and a change of government, and — as of this year — a set of ministerial noises about restoration that the Calcutta Tram Users' Association is treating with the caution of people who have heard noises before.

A hundred and fifty-three years, and the argument is still live.

Now here is the thing about Bengali writing on Calcutta that took me an embarrassingly long time to see. The elegy came first.

Jibanananda's Calcutta poems — 'Phutphate', On the Pavement, published in 1938 and later collected in Mahaprithibi — are already a literature of exhaustion, of a city that has gone grey and stopped meaning anything, of a man walking streets that no longer answer. Buddhadeva Bose was doing something adjacent. Samar Sen, Sudhindranath Datta, Bishnu Dey, the whole Kallol generation announcing its liberation from Tagore into a poetry of urban solitude and dust.

That is 1938.

In 1938 Calcutta is the second city of the British Empire. It has the jute, the port, the university, the presses, the Bengal Renaissance still within living memory, and roughly three hundred and forty trams running on a network that covers the entire municipality. Independence is nine years away. Partition, which will cut the city off from the eastern half of its own hinterland and pour refugees into it for two decades, is nine years away. The capital had been moved to Delhi in 1911, which stung — but the decline that every Calcutta elegy since has taken as its subject has not happened yet.

The writers were mourning a city that was still functioning.

I do not think this is a case of prophecy, and I would be suspicious of anyone who claimed it was. What I think happened is a good deal more interesting and slightly less flattering to Bengal, which is that the elegiac mode was available, it was prestigious, it suited a generation trying to get out from under Tagore's enormous shadow — and once a literature adopts a mode, it starts finding evidence for it. Then history obliged. Partition, refugees, the food movement, the Naxal years, the flight of capital, deindustrialisation, the long slow slide. By the time the decline was real, the vocabulary for describing it had been in place for thirty years and was fully furnished.

Which raises a question the city does not much enjoy: how much of Calcutta's identity as a place in permanent decline is a fact about the city, and how much is a fact about its literature?

And then the fiction picked it up and ran with it for fifty years. Sunil Gangopadhyay's young men drifting through a city that has nothing to offer them. Shirshendu Mukhopadhyay's interiors. The Hungryalists in the sixties, going at Bengali propriety with a hammer and getting prosecuted for obscenity for their trouble — an avant-garde that could only have formed in a literary culture confident enough to be attacked. Satyajit Ray's Calcutta trilogy in the seventies, in which the city stops being a backdrop and becomes an antagonist with a job market. Mrinal Sen doing the same thing angrier. The mode became so total that a Bengali writer who wanted to write Calcutta as a functioning place had to do it against the grain of everything around him, which is roughly the position Amit Chaudhuri has occupied for thirty years.

Chaudhuri, who has spent a career on exactly this, comes at it from the other side and in the other language. A Strange and Sublime Address is a small boy's summer holiday in a Calcutta house, and almost nothing happens in it, deliberately — the whole book is an argument that the ordinary texture of a place is more interesting than the catastrophe narrative attached to it. Years later, in Calcutta: Two Years in the City, he came back to write the non-fiction version and found himself confronting the tourism of ruin: the visitors who come for the crumbling facades, the pleasure that outsiders and insiders alike take in the idea of a grand thing going down.

He is the same writer who told the English-reading world that Kolatkar deserved to be as well known as Rushdie. Which is the connection between this piece and the last one, and it is not incidental. Chaudhuri is one of a very small number of people who have worked both sides — a Bengali novelist writing in English who spends his critical energy dragging bilingual and vernacular writers toward a readership that had no idea they existed. The traffic between Indian literatures does not happen automatically. It happens because specific people carry specific books across, usually at a personal cost, usually decades late.

There is also a quieter fight over the poet himself, which nobody in either country has any interest in settling. Jibanananda was born in Barisal, in what is now Bangladesh, taught there for thirteen years, and came to Calcutta for good only after 1947. Bangladesh claims him as the great poet of the Bengal countryside. West Bengal claims him as the great modernist of the city that killed him. Both are right, which is inconvenient for everybody.

And the trams are the whole argument in a physical object, which is why Calcutta cannot let go of them and why the fight has gone on for years in a High Court.

A tram is not efficient. It is slow, it blocks a lane, it costs money to maintain, and the transport minister was not lying about the traffic. A tram is also the last working piece of nineteenth-century municipal infrastructure in India, it is nearly free to run once built, it emits nothing, its cars last fifty to eighty years against five to ten for a bus, and it is threaded through the streets of north Calcutta in a way that makes the neighbourhoods legible — Lalbazar, the Nakhoda Masjid, Tagore's house at Jorasanko, Bagbazar, in a single line. The Calcutta Tram Users' Association has pointed out that the network would be a serious candidate for a UNESCO listing if only somebody would guarantee to maintain it for twenty years.

Nobody will guarantee twenty years. That is the entire problem, and it is not a Calcutta problem, it is an Indian one. We covered a version of it in these pages last month, in the feature on Indian audio drama — a medium that produced four thousand scripts a year for decades and kept almost none of them, because keeping requires a decision and nobody was ever asked to make it. An archive is not a recording. It is the decision to keep one.

A tram is not transport. It is the decision to keep one.

Sunil Gangopadhyay ran the moods of Neera, his lifelong fictional muse, against the reptilian crawl of a tram. Satyajit Ray put one in Apur Sansar. Tagore has them. The city's literature is so thoroughly wired into the sound of that bell that the removal of the network is being experienced, correctly, as an act of literary vandalism rather than a transport decision.

And Jibanananda, the poet of beautiful Bengal, the man who wrote the rural landscape so hard that Bangladesh and West Bengal have been quietly fighting over him ever since — Jibanananda was killed by the machine, in the city, in the dark, at fifty-five, eight days after the impact, having published almost nothing that sold.

The elegy did not need to wait for the decline.

That is William Dalrymple's account of it, and it is the least improbable thing that happened to Twilight in Delhi. Ahmed Ali finished it in 1939 and carried the manuscript to England that August, three weeks before the war. The Hogarth Press wanted it. The printers refused to set it — subversive of law and order, they said, meaning that a novel about what the British had done to Delhi after 1857 was not a thing to be putting into English shops while English soldiers were being recruited in Punjab. Ali was asked to cut the offending sections. He declined. E. M. Forster, Virginia Woolf and Desmond MacCarthy pushed it through and it came out in 1940 to genuine acclaim, and then the German air force settled the matter, and there was no reprint, and the book effectively vanished for decades.

Read the sentence again, because the shape of it recurs. The definitive novel of Old Delhi was written in English, published in London, physically destroyed by a European war, and disappeared from India for a generation.

It gets worse, and then it gets extraordinary.

In 1947 Ali was in China — British Council Visiting Professor at Nanking, appointed by the Government of British India, teaching English, learning Chinese, collecting the material that would become his book on Muslim China. He tried to come home in 1948. He was not permitted to. K. P. S. Menon, India's ambassador in China, would not let him return.

Why, exactly, is contested and the disagreement is worth keeping rather than resolving. The official version is procedural: Ali was a government employee who had failed to file his option — the form on which such men were required to state whether they were choosing India or Pakistan — and so the machinery could not process him. Ali's own account, given in the introduction to a later Indian edition of his novel, is not procedural at all. He wrote that he was prohibited from returning for no reason other than that he was a Muslim, and quoted Menon telling him it was a question of Hindu and Muslim, and that a letter to New Delhi would achieve nothing. Ali added that he found this incredible, and then made the connection that the whole business had been waiting for: this was his family's banishment from the city happening a second time, ninety years after the British had emptied Muslim Delhi in 1857.

He went to Karachi in 1948 and stayed. He joined the Pakistan Foreign Service at Liaquat Ali Khan's insistence and became Pakistan's first envoy to the People's Republic. He died in Karachi in 1994, at eighty-three.

The author of the greatest novel about Delhi was locked out of Delhi.

And then his wife translated him back in. Bilqees Jahan — poet, painter, whom he married in 1950 — put Twilight in Delhi into Urdu as Dilli ki Sham, published in 1963. A novel about the death of Urdu Delhi, written in English by a man who wrote Urdu fiction, had to wait twenty-three years and a marriage to reach the language it was actually about.

Somebody should write about that. Nobody has, much.

Now the dates, because Delhi in fiction runs on three of them and almost nothing else.

1857. Ali's subject, though he sets the book in 1911 to 1919 and lets the catastrophe sit underneath every room like damp. The Coronation Durbar passes through the novel the way the King's car passes through Mrs Dalloway, and the family in the courtyard watches an empire celebrate itself on the site of the one it destroyed.

1947. Anita Desai's Clear Light of Day, in which a Delhi family in Old Delhi comes apart in the same year the country does, and the house — dust, the koel, a brother lost to Hyderabad, a sister who stays — carries the whole thing. Khushwant Singh's Train to Pakistan, which is not set in Delhi at all but which is what most readers have instead of a Delhi Partition novel. Krishna Sobti, writing in Hindi, moving between the two Punjabs and the capital they emptied into.

1984. And here the pattern breaks in a way that tells you something.

Amitav Ghosh was a young lecturer at Delhi University living in Defence Colony on the thirty-first of October, 1984, when Indira Gandhi was shot a few miles from his flat. Over the following three days some three thousand Sikhs were murdered in the capital. Ghosh joined the citizens' relief effort, walked into neighbourhoods that had been taken over, and watched a city with the largest security apparatus in India decline, hour after hour, to do the thing it knows perfectly well how to do — declare the curfew, deploy the paramilitaries, send in the Army. Calcutta's authorities, he noted, moved promptly. Delhi did not.

He wrote about it eventually, in an essay, and opened by saying that the experiences had been decisive for him as a writer and that this was precisely why he had never attempted to write about them until now.

Sit with that admission for a moment, because it is the most honest sentence any Indian writer has produced about the capital. The most formidable novelist of his generation, present at the event, morally engaged in it, materially involved in the relief — and the thing he could produce was reportage. Not a novel. An essay, years late, that begins by explaining the delay.

Fire was everywhere, he wrote. It was the day's motif.

There is a mirror to all of this on the other side of the border, and it is worth a paragraph because it makes the point from outside. Intizar Husain, who left Meerut for Lahore in 1947, spent a career writing the towns of western Uttar Pradesh — the qasba, the small north Indian settlement with its trees and its water and its unbearable ordinariness — from a country he had moved to and a city that was not it. Basti is a novel about a man in Lahore who cannot stop remembering a place in India. Which means one of the finest novels of north Indian small-town life was written in Urdu, in Pakistan, by a man who could not go back, and reached English properly only through Frances Pritchett.

It is the Ahmed Ali shape again, one generation later. The novelist of a place, writing it from the country he was sorted into.

Delhi's problem is not that it lacks writers. Delhi has all the writers — the publishing houses, the universities, the magazines, the fellowships, the entire apparatus of Indian letters is headquartered within about eight kilometres of Connaught Place. Delhi's problem is that the city becomes narratable when it is burning and goes strangely opaque when it is not.

Bombay produced the arrival novel because the experience it had to offer was arrival. Calcutta produced the elegy because the mode was there before the decline and the decline caught up. Delhi produced the catastrophe novel because a city that has been destroyed and rebuilt seven or eight times, depending on who is counting, and emptied of its population twice within ninety years, offers catastrophe as its most legible narrative material — and because the ordinary Delhi, the Delhi of jobs and rented rooms and buses and the daily grind of being a person in a very large administrative town, is the least glamorous subject in Indian writing.

It is also, and this is the part that gets left out, a city that lost its own literary language. Delhi was an Urdu city — Ghalib's, Mir's, Zauq's, Dagh's, the whole apparatus of the mushaira and the ustad and the printing presses of Ballimaran. Twilight in Delhi is an elegy for exactly that world, written in English, by a man who wrote his own short fiction in Urdu, because English was where the readership had gone. Krishna Sobti, who came out of Gujrat in what became Pakistan and wrote in Hindi for sixty years, spent much of her career on the languages and registers of a Punjab and a Delhi that the map had cut up.

A city that changes its literary language twice in a century does not accumulate a continuous literature. It accumulates fragments in three scripts, and then somebody has to translate between them.

Somebody did write that Delhi. The migrant's Delhi, the clerk's Delhi, the Delhi of the badly-lit room in Govindpuri and the friend who lends you money and the decades of not quite belonging, running from the China war through the Emergency to 1984 and out the other side.

He wrote it in Malayalam.

Twenty-eight years. And in those twenty-eight years the canon of the Indian city novel was assembled, argued over, taught, anthologised and exported, and Thanneer was not in the room, because Thanneer was in Tamil and the room was reading English.

Consider what was actually in that hundred pages. Madras, late sixties, the first serious water crisis the city had known. A street. Three women — Jamuna, whose film career did not happen and whose lover is working out how to exploit her; her sister Chaya, married to an army man and surviving in a women's hostel; and Teacheramma, holding a household together around a sick husband. Around them, the tanker. The procession of pots. The queue. Ashokamitran does not waste a page on backstory — he will hand you an entire life in a paragraph and move on. By the end the crisis has not lifted. The street lights are out. Drainage has got into the drinking water. Nothing is resolved and the book is somehow not despairing.

That is a pitiless urban chronicle of the first order and there is nothing quite like it in Indian writing in any language. It is also, and this is the part that ought to bother us, a novel about municipal infrastructure — about what happens to ordinary people when a city's plumbing fails — which is a subject the English-language Indian novel has essentially never touched, in a country where it is the central daily fact of urban life for a few hundred million people.

The man wrote eight novels, fifteen novellas and more than two hundred short stories. His actual name was Jagadisa Thyagarajan. He grew up in Secunderabad, and in 1977 published Pathinettavathu Atchakkodu — The Eighteenth Parallel — about a boy in the Nizam's dominions during the months when the princely state was being folded into the Indian Union in 1947 and 1948. Which means the definitive fictional account of the end of Hyderabad State is a Tamil novel, written by a man who had moved to Madras, published thirty years after the events, and it is not on any list of Indian Partition novels that you or I were handed at university.

He died in Chennai in 2017. He was ninety.

And then there is Mukundan, who is the whole argument compressed into one biography and one very specific set of dates.

M. Mukundan was born in Mahe, on the Malabar coast, in the tiny French enclave that stayed French until 1954. He went to Delhi in 1962 and worked at the French Embassy as a cultural attaché for something like thirty-six years, retiring in 2004, and by his own count spent about four more years simply wandering the city — Old Delhi, Govindpuri, the localities where, as he put it, we cannot even imagine the conditions people were living in. Forty years in the capital.

He wrote Delhi Gathakal. It came out from DC Books in November 2011 — four hundred and ninety-six pages, in Malayalam, following a young leftist from Kerala named Sahadevan who arrives in Delhi and is met, more or less on the doorstep, by the news of the Chinese attack. From there the novel goes through everything: the war, the Emergency, the riots, the decades, a city seen entirely from the position of somebody who came to it from somewhere else and never became one of its subjects.

The English translation, by Fathima E. V. and Nandakumar K., appeared in 2020. It won the JCB Prize in November 2021.

Nine years from Malayalam to English. Ten to the prize. And on the night he took it, Mukundan described it as a novel of the poor and the wretched, and said the translators could not have done it better, and it had taken the three of them a year of work.

That is the Delhi novel of the second half of the twentieth century. The one the previous piece went looking for and could not find in English. It existed the entire time.

Note also the small horror in the timing. One of the JCB judges mentioned reading the book in February 2020, from a flat in East of Kailash, while Delhi burned again — Mukundan writing 1984, the reader looking up at the news, the recognition that you cannot get out of the way of this city's history by moving forward in time.

Three of the JCB Prize's winners in its first four years were translations from Malayalam. Three, out of four. That is not a coincidence and it is not evidence that Malayalam is the best literature in India. It is evidence that Malayalam has the strongest translation pipeline in India — publishers, translators, a reading public that takes it seriously, prizes that reward it, and DC Books, which has been carrying Malayalam literature for decades in a way most Indian language publishing has not managed.

Which brings me to what this piece is actually about, and it is the same argument this magazine ran last month in a different medium.

In August, in these pages, we tracked how a handful of Indian radio dramas escaped a medium that kept nothing. Andha Yug survived because in 1962 Satyadev Dubey put it on a stage, and theatre had directors with reputations, printed texts, festivals, critics, syllabi — an apparatus of memory that radio simply did not possess. The play did not survive by being good. It survived by walking out of a building with no archive and into a building with one.

A novel in Tamil or Malayalam or Odia is in exactly that position. It can be superb. It can be the best book published in India that year. What it cannot do, from inside its own language, is enter the thing we have agreed to call Indian literature — because that category, whatever anyone says at a festival, means the books that are legible in English.

So translation is not a service industry attached to Indian writing. It is the border control. It decides which cities exist.

And the delays are not neutral, because a book arriving twenty-eight years late arrives into a conversation that has already been had. Thanneer in 2001 is a rediscovery, a classic, an object of respect — the kind of book that gets an admiring review and a place on a syllabus about Tamil modernism. Thanneer in 1975, in English, alongside the early Rushdie generation, would have been a rival. It would have made the Indian urban novel argue with itself. It would have made Madras a literary city, which it is, and which the map does not show.

Sundara Ramaswamy in Tamil. Vinod Kumar Shukla in Hindi, writing small-town Madhya Pradesh with a strangeness nobody else attempted, and who reached English properly only through translators like Sara Rai — who then chaired the jury that gave Mukundan the prize, which is the kind of small closed circuit that tells you how few people are doing this work. Qurratulain Hyder in Urdu, who was so unhappy with the situation that she translated her own novels rather than wait — Aag ka Darya appeared in 1959 and she produced her own English version, River of Fire, herself, four decades later. She called it a transcreation rather than a translation, which is a novelist's way of saying that she did not trust anyone else to do it and was not going to pretend the two books were the same object.

There is a second-order problem underneath all this and it deserves naming, because it is the one nobody at a literature festival will say out loud. A translation pipeline does not carry everything equally. It selects. And what the English-language market has historically wanted from Indian-language fiction is the rural, the caste-marked, the folkloric, the anthropologically legible — the book that reads as authentically Indian to somebody in London, which in practice means the book that is furthest from a metropolitan English novel. Perumal Murugan reached the world through Tiruchengode and its fields, not through a city. A Tamil novel about a water tanker in a middle-class Madras street offers a foreign editor almost nothing to sell.

So the urban vernacular novel is doubly disadvantaged. It waits behind everything else in the queue, and when it gets to the front it does not look exotic enough to be worth the trouble.

The provincial silence is not a silence.

Rana Dasgupta had written two works of fiction — Tokyo Cancelled, Solo, the latter winning the Commonwealth Prize in 2010 — and when he came to write the city he had been living in since 2001, he abandoned the form he was good at. What he produced instead was a book of encounters: billionaires, metal traders, drug dealers, slum residents, the new black-money elite, with a century of the city's history woven through. It was shortlisted for the Orwell Prize, won the Kapuściński Award, and is now the standard reference for anyone trying to understand what liberalisation did to the capital.

A novelist looked at contemporary Delhi and concluded that the honest response was reportage.

He is not alone, and the pattern is worth naming because it is not an accident of temperament.

Suketu Mehta, on Bombay: Maximum City, non-fiction. Katherine Boo, on Annawadi — the settlement beside the Sahar airport, wedged between the international terminal and the luxury hotels, where the residents' economy runs on the airport's garbage: Behind the Beautiful Forevers, non-fiction, three years of reporting, and written by an American. William Dalrymple on Delhi: City of Djinns, non-fiction, and a Scot. Naresh Fernandes on Bombay's jazz age and its urban planning: non-fiction.

The moment the subject becomes the actual contemporary Indian city — its infrastructure, its settlements, its municipal reality — the form flips to reportage, and with some regularity the writer flips to a foreigner.

The Boo case is the sharpest of these and I want to be fair about it, because Behind the Beautiful Forevers is a magnificent book and the objection to it is not that an American wrote it. The objection is what its existence implies. Annawadi sat for years within sight of the international terminal that every Indian novelist in this piece has flown out of. Roughly three thousand people, living on the airport's waste, half an hour from the flats where the country's English-language literary establishment does its work. It took a reporter from Washington, three years of fieldwork, and a translator, to produce the book about it.

Dharavi is the largest informal settlement in Asia by most reckonings and has a manufacturing economy running into hundreds of millions of dollars a year. It has been filmed, photographed, toured, studied, redeveloped on paper about six times, and used as a set piece in a British film that won eight Academy Awards. It does not have its novel.

Meanwhile, look at where the novels are set.

The Bombay novel in English is overwhelmingly a South Bombay novel. Colaba, Kala Ghoda, Malabar Hill, Marine Drive, the Fort, the chawls of the mill district at a push. Which is to say: the old island city, roughly the bottom fifteen per cent of the metropolitan area, containing a small fraction of its population. Borivali, Mulund, Bhandup, Kandivali, Thane, Navi Mumbai — millions of people, decades of lives, almost nothing.

The exceptions prove the shape of it. Kiran Nagarkar put Ravan and Eddie in a chawl and made it a comic masterpiece. Aravind Adiga set Last Man in Tower in Vakola, a suburb by the domestic airport, in a decaying co-operative housing society where a builder wants everyone out — and it is one of the few English novels that treats an Indian housing society as a serious political arena, which, as anyone who has attended one of those meetings can tell you, it is. Jerry Pinto wrote a one-bedroom-hall-kitchen in Mahim with such precision in Em and the Big Hoom that the flat becomes the fourth character.

Adiga is worth staying with for a second, because Last Man in Tower does the thing almost nobody else attempts. Its subject is a redevelopment offer — a builder, a decrepit society block, a generous cash figure, and one retired schoolteacher who will not sign. That is not a metaphor for anything. That is the single most consequential social process in urban India over the last thirty years, conducted in stairwells and general body meetings, and it has produced roughly one novel.

The Delhi novel is Old Delhi or Lutyens or the south colonies. Gurgaon has close to two million people, a skyline you can see from a landing aircraft, an economy the size of a small country's, and no fiction of consequence at all. Noida the same. Salt Lake the same. Navi Mumbai the same. These are among the largest planned urban settlements built anywhere on earth in the last fifty years and Indian literature has, so far, produced essentially nothing about them.

Which is odd, until you notice what they have in common. They were all built by state agencies and private developers within living memory, they have no ruins, no Mughal layer, no colonial layer, no Renaissance, no elegy available. There is nothing to mourn yet. And an Indian city, in Indian literature, becomes narratable mainly by having something to mourn.

That is the through-line of this whole feature and I did not expect to arrive at it here.

Bombay is written as arrival, and arrival requires a place that was already there to arrive at. Calcutta is written as elegy, and the elegy was in place before the decline. Delhi is written as catastrophe, and its three usable dates are 1857, 1947 and 1984. Every one of those modes is retrospective. Every one requires the city to have a past that the writer can stand at the end of.

Gurgaon has no past. It has a land record.

And there is a version of this argument that says the new cities will get their novels in thirty years, once enough has happened in them to be remembered — that Gurgaon simply has to wait its turn, the way Bombay waited, and that the first generation raised entirely in a satellite town will produce the writer who does it. That is probably true and I hope it is.

But it also concedes the point. It says that an Indian city becomes writable once it has acquired a past, which means Indian literature can only ever describe the city as it was, from a position after it.

Ashokamitran did not do that. Fifty-seven years ago he wrote a hundred pages about a water tanker.

That is the counter-example and it is why he closes this feature rather than the fourth piece. Thanneer has no ruins in it, no lost aristocracy, no elegy, nothing to mourn. It is about a municipal failure in a working city and the specific women whose weeks are reorganised around a queue. It is the model for the novel of the Indian city that we do not have, and it was published in 1973, and the fact that it is not the founding text of Indian urban fiction is a fact about translation and nothing else.

There are signs of movement, and it would be sour not to say so. The JCB Prize has done more for the visibility of Indian-language fiction in a decade than the previous fifty years managed, whatever one thinks about prize culture. Publishers who did not exist twenty years ago — Navayana, Seagull, Blaft, Westland's translation lists, Eka — are commissioning translators rather than waiting for finished manuscripts to turn up. Translators are getting cover credit and prize money. Malayalam has shown what a functioning pipeline looks like and the other languages are watching.

But the map in your head has already been drawn, and it has three cities on it, and it was drawn by which books happened to be available in English at the moment you were being taught what Indian literature was.

Bombay is not the most written Indian city. It is the most translated one, and it did not need translating.

Four thousand kilometres of coastline, fifty-odd cities over a million, twenty-two scheduled languages.

Name the great Indian city novels.