Again, Today

A bridge that is load-tested about 137,400 times a day by people who are not testing it — they are just going to work

R831’s Wow! signal had the loneliest epistemology on record: one facing, ever, and no way to earn a second. Now take the exact opposite shape. What of a proof that is re-run constantly — daily, hourly, by everyone — and passes so routinely that nobody registers it as a proof at all? A test whose thousandth facing stopped counting as evidence precisely because it has never yet failed. There is one hanging over the East River, and its first facings were not routine at all. People were afraid of it, and the fear was front-page news.

At first, every crossing counted. The Brooklyn Bridge opened on 24 May 1883 as the longest suspension bridge in the world — a main span of 1,595.5 feet, the first suspension bridge hung from steel wire, the only land passage between the separate cities of New York and Brooklyn. On its first day, some 150,300 people and 1,800 vehicles crossed it. Six days later, on 30 May, a woman fell on the wooden stairs at the Manhattan approach, another screamed, a rumour ran through the holiday crowd that the bridge was falling, and the stampede on the narrow stairway killed 12 people and seriously injured dozens. The bridge had not moved. What failed was the public’s willingness to treat a crossing as a proof — and at that point a crossing still was treated as one, which is why a rumour about its result could kill.

The last crossing that was staged as a proof had elephants in it. P. T. Barnum had offered, before the opening, to march his elephants across as a demonstration, and was turned down. After the stampede, the offer was accepted on his terms rather than the city’s fee. On 17 May 1884 he paraded 21 elephants and 17 camels over the bridge, the great Jumbo bringing up the rear, before thousands of spectators; the New York Times reached for a Noah’s Ark comparison for the sight of the menagerie crossing under the electric lights. The animals crossed, the bridge did not budge, and the demonstration did the work it was hired to do. Notice what kind of proof it was: a single, deliberate, witnessed load, performed because individual crossings — already numbering in the hundreds of thousands — had somehow not added up, in the public mind, to the same conclusion. The elephants were not stronger evidence than the traffic. They were evidence of a kind people could watch.

Now the test runs itself, about 137,400 times a day. The city’s Department of Transportation gives the bridge’s daily averages, as of 2024, as 103,051 vehicles, 28,845 pedestrians and 5,504 cyclists — 137,400 crossings a day, each one a genuine load placed on the deck, the suspenders and the four main cables, each one answered, within seconds, by the bridge still standing. Nobody counts these as results. There is no printout, no red circle in the margin; the outcome of the test is that a commute happened. The thaler of R828 at least knew it was being examined — a holder bit it, weighed it, read its edge. The bridge’s examiners do not know they are examining. The proof has become so reliable that it has dropped out of the category of proof and into the category of ground: the thing you stand on while thinking about something else.

And here is the honest edge: a crossing proves a narrow thing. Each passage demonstrates that the structure carried that load, on that day, at that temperature. That is real evidence, and there is a mountain of it. But a crossing inspects nothing. It says nothing about the inside of a main cable — each of the four is a bundle of nineteen strands of roughly 280 wires apiece, wrapped tight, its interior visible to no one crossing over it. The Roeblings built in a wide margin — the cables are reckoned strong enough to carry more than five times the load they actually bear — and the places where time has actually bitten are precisely the places no traveller ever sees: city records describe corrosion where the strands splay inside the two anchorage chambers, inspectors reduced to fibre optics to see the wires where they pass around the eyebar pins, and 1990s rehabilitation contracts that rehabilitated the cables in both anchorages and dehumidified them. None of that was learned from traffic. All of it was learned by stopping, unwrapping, and looking — the deliberate act the daily proof never performs.

Another bridge shows what the routine can hide. The Silver Bridge at Point Pleasant, an eyebar-chain suspension bridge opened in 1928, was crossed routinely for thirty-nine years — tens of thousands of unremarkable, successful facings, each one silently filed by its users as evidence. On 15 December 1967, at rush hour, it fell into the Ohio River and 46 people died. Investigators traced the collapse to a cleavage fracture in the eye of a single eyebar in the suspension chain, grown by stress corrosion and corrosion fatigue — in a joint where, the safety board found, the flaw could not have been discovered by the inspection methods of the time without taking the joint apart. Thirty-nine years of passes had tested the bridge’s strength every day and its flaw not once, because the flaw was not in the part of the bridge a crossing tests. The national response is the tell: the Federal-Aid Highway Act of 1968 and the inspection standards that followed by 1971 did not ask Americans to cross more attentively. They required bridges to be inspected — at least every two years, by qualified inspectors, looking on purpose. The country formally decided that survival is not an inspection.

The edges, kept straight. This page is not an alarm about the Brooklyn Bridge, which is not the Silver Bridge: its load hangs from thousands of parallel wires, not a single non-redundant chain of eyebars; its design margin is large; and it lives under the very inspection regime the Silver Bridge’s fall created. Nor is the daily mountain of passes worthless — it rules out whole families of failure, the ones that would have shown up under load long ago. The claim is narrower and stranger. Evidence has a half-life in the mind: the first crossing was an event, the elephants were a spectacle, the millionth crossing is a commute. The test did not get weaker as it repeated. It got invisible — and the one thing it never tested, the hidden interior, went on changing at its own pace, in the dark, entirely indifferent to how many times the visible part had passed.

The test, schematically. The numbers in the toy are schematic, as stated: a day’s crossings are set to the real 2024 daily total (137,400), but the hidden-wear figure is an invented unit, included to show what crossings cannot touch. Run as many days as you like: the crossings counter and the loads-proved row grow without limit, and the interior row does not change — wear advances only with simulated years, and it is revealed only by inspecting, which no number of crossings substitutes for.

Crossings so far in this toy: 0. The bridge has been proved, today, by nobody yet — in the world, it already has been, about 137,400 times, before lunch.

The Wow! signal waits for a second facing that will not come. The Brooklyn Bridge gets a hundred thousand facings before breakfast and cannot spend any of them on the one question that matters most — what is happening inside the cables — because a facing that consists of driving over the answer cannot see it. Somewhere between one facing and a hundred million, a proof stops being information and becomes the floor. The elephants knew they were in a demonstration. Nobody since has been.