Day 52 · The Rise & Fall of Cities

Cities Are Not Built. They Are Kept Fed.

Tuesday, 18 August 2026 · BigCat's Time Machine
A siege that cut the aqueducts, a treaty that traded submission for self-rule, a calculation that doubled a pipe, an expressway never built — a city's fate usually turns on what one person decided under deadline about one of four things: supply, legal standing, capacity, veto.
EVENT · 01

The Moment the Aqueducts Were Cut, a City of a Million Began Reverting to VillagesThe Siege of Rome · Aqueducts Cut, March 537 CE

March 537 CEthe city of RomeWard-Perkins 2005 · Wickham 2005

Second-century Rome held roughly a million people; local farmland could not have fed a fraction of them. Two external pipelines kept it alive: grain shipped across the sea from Egypt and North Africa under the state dole (annona), and water delivered by eleven aqueducts. In 439 the Vandals took Carthage and the African grain route was severed first. After the Western empire ended in 476, the Ostrogothic king Theodoric still kept the grain administration running. In 535 Justinian sent an army to reconquer Italy.

In March 537 the Ostrogothic king Vitiges laid siege and made a precise decision: do not storm the walls — cut every aqueduct. The city fell back on Tiber water, the watermills stopped, and the defending general Belisarius mounted millstones on boats moored in the current as a stopgap. The siege lasted a year and nine days; over the following two decades Rome changed hands repeatedly, and by the end of the sixth century its population had fallen from about a million to a few tens of thousands. The city was not destroyed by any one massacre — the sea-borne system that fed it was snipped, segment by segment.

Without that twenty-year war of reconquest, the urban order under Ostrogothic rule might well have continued. The dispute is whether to call this collapse or transformation: Bryan Ward-Perkins, The Fall of Rome and the End of Civilization (2005), reads the archaeological distribution of pottery, roof tiles and coinage as evidence of genuine material regression; Peter Brown's “Late Antiquity” school stresses continuity and recomposition. Chris Wickham (2005) splits the difference: what broke was the state machine that converted taxation into long-distance supply.

Every city whose scale rests on external transfers sits on the same curve: mining towns, company towns, subsidised new districts. Scale never justifies itself — who is paying for it, and how many single points does the chain have?

A city's ceiling is set by its supply chain, not by its buildings.
In the systems you depend on, which chain is invisible until the day it snaps?
EVENT · 02

Cities Won a Battle and Bought Themselves a Legal IdentityLegnano & the Peace of Constance · Lombardy, 1176–1183

1176–1183LombardyPirenne 1925 · Wickham 2015

In eleventh- and twelfth-century northern Italy, Milan and its neighbours revived on trade and were already electing their own consuls, writing their own statutes and levying their own taxes — while remaining, in law, subjects of the Holy Roman Empire. At the Diet of Roncaglia in 1158, Frederick I (Barbarossa) announced the recovery of all royal prerogatives — minting, taxation, appointments — declaring decades of accomplished fact illegal at a stroke. The cities formed the Lombard League.

At Legnano on 29 May 1176, the League's civic infantry held against the emperor's heavy cavalry; Barbarossa was unhorsed and missing for days. The Peace of Constance followed on 25 June 1183: the emperor kept nominal overlordship and the rituals of homage, while conceding in substance that each city elected its consuls, made its own law and collected its own taxes. This was not a declaration of independence but a trade — the form of submission exchanged for the content of autonomy. The maxim “city air makes you free” now had something behind it.

Had Barbarossa prevailed, Italy might have been absorbed into a territorial-princely order as Germany was, and the city-state competition that later produced Florence and Venice would have lost its soil. Two debates: Henri Pirenne, Medieval Cities (1925), attributed the urban revival to the restart of long-distance trade, later revised by Michael McCormick (2001) into an earlier, more continuous process; and Weber's East–West contrast, cited more often still — contemporaneous Kaifeng under the Northern Song tore down its ward walls and abolished the curfew, with commerce dwarfing any European city, yet remained an administrative seat of the court. That contrast is easily read as a ranking, which is exactly where Qian Mu's caution applies: the Chinese “city” was a node of prefectural administration, not a political unit — the two kinds of city were solving different problems.

Autonomy is rarely granted; it is conceded once suppression costs more than it is worth. Platforms and their developers, local versus central fiscal power — all are arguing the same thing: who bows in form, who writes the rules in substance.

Institutional freedom is seldom won by declaration; it is won by making recognition cheaper than repression.
In your organisation, which real autonomy is paid for with formal deference?
EVENT · 03

He Doubled the Pipe Diameter He Had Calculated — and the Sewers Lasted 160 YearsThe Great Stink & Bazalgette's Sewers · London, 1858–1865

June–August 1858LondonHalliday 1999 · Szreter 1988

In the first half of the nineteenth century London grew from one million to over two and a half. Flush toilets pushed waste into the sewers and the sewers discharged straight into the Thames — from which the city also drew its water; the cholera outbreaks of 1848 and 1854 killed more than twenty thousand. Joseph Bazalgette became chief engineer of the Metropolitan Board of Works in 1856, and his river schemes were rejected by the Treasury one after another: spending a fortune on a river that was merely foul did not pay politically.

In June 1858 a heatwave turned the Thames into an open fermentation vat, and Parliament — sitting on the riverbank — could not conduct business. A bill that had gone nowhere for ten years passed within eighteen days of MPs being gassed by their own river: on 2 August the Board was authorised to borrow three million pounds without further Treasury approval. The decisive move, though, was technical. Having calculated the pipe diameter from the population and the densest usage of the day, Bazalgette doubled it outright — on the grounds that you build this sort of thing once in a lifetime. The Crossness pumping station opened in 1865, with some 82 miles of intercepting sewers and over a thousand miles of street sewers. London never suffered another major cholera epidemic, and the network remains the backbone of its drainage today.

Had that summer been cooler, or Parliament not sat on the water, the bill would likely have slipped another decade — priced in another cholera outbreak. The scholarly argument is not about the engineering but about attribution: Thomas McKeown (1976) held that the nineteenth-century mortality decline came mainly from better nutrition; Simon Szreter (1988) countered with evidence on municipal investment and local politics, arguing that precisely this kind of forced public-health spending turned urban mortality around. And an irony sits on top: the miasma theory Bazalgette worked from — bad smells cause disease — was wrong, while the prescription of moving sewage away from people was right.

The doubled diameter was waste at the time and is London's most valuable inheritance 160 years on. The other half is worth remembering too: what moved the project was not argument but decision-makers being gassed themselves — when the pain never reaches whoever signs the cheque, the right proposal simply queues. That is not stupidity but incentive structure: costs now, benefits to successors.

Infrastructure earns its value not on the day it opens, but in the extra margin you left for a future you could not predict.
Which of your systems is sized for today's peak? If tomorrow needs ten times that, who pays for the rebuild?
EVENT · 04

An Expressway That Was Never Built Saved Half of ManhattanJacobs vs. Moses & the Lower Manhattan Expressway · New York, 1961–1969

1961–1969New YorkCaro 1974 · Glaeser 2011

Robert Moses never won an election, yet by holding a dozen public-authority posts at once he controlled New York's bridge toll revenue and construction approvals, and for forty years shaped nearly every expressway in the city. The Federal-Aid Highway Act of 1956 supplied 90% matching funds, turning “demolish the city, build the road” into a near-free option for city hall. His Lower Manhattan Expressway would have run a ten-lane elevated road along Broome Street, slicing through SoHo and part of Chinatown and displacing roughly 2,000 households and 800 businesses.

In 1961 Jane Jacobs published The Death and Life of Great American Cities, aimed squarely at the planning doctrine of reading a city off a map: a street is kept safe by people continuously watching it, not by making it wider. The fight ran nearly a decade, and its method was to turn every public hearing into procedural terrain — on 10 April 1968 Jacobs and the crowd surged the stage and upended the stenographer's record, and she was arrested on charges including inciting a riot. On 16 July the following year, Mayor Lindsay formally killed the project.

Had the expressway been built, the SoHo and Chinatown blocks that are now the liveliest in the city would have been cut in half — Moses's Cross-Bronx Expressway displaced tens of thousands and left that district unrecovered for thirty years. The argument has reversed in recent years: Robert Caro's The Power Broker (1974) fixed the canonical story of Moses as unaccountable power; Ed Glaeser's Triumph of the City (2011) and others counter that the community veto Jacobs won was later used to block all new housing, and is the institutional source of runaway prices in New York and San Francisco. Both are right: a veto protects what is already good and locks out what might be.

Every design that answers “who gets to say no” carries this double edge — code review, architecture boards, community governance. A veto exists to prevent irreversible damage; once everyone holds one, the system can only preserve the status quo.

Whoever you hand the veto to, you have decided how much change the system can still absorb.
Who can say no in the system you run? What was that veto created to prevent — and is it still preventing the same thing?

Four Cities, One Ledger

Supply, legal standing, capacity, veto — nearly every rise and fall is booked in these four columns.
When / Where
The specific decision
What the city gained / lost
Unintended consequence
537
Rome
cutting 11 aqueducts in a siege
lost its sea-borne supply chain
a city of a million shrank to a few districts
1183
Constance
nominal submission traded for real autonomy
gained self-legislation and corporate standing
city-state rivalry bred the Renaissance
1858
London
passed in 18 days, pipe diameter doubled
gained 100% capacity headroom
a Victorian network still running today
1969
New York
the Lower Manhattan Expressway cancelled
kept SoHo and Chinatown
the community veto became a housing shortage

Going Deeper

Why is a city's rise always easier to narrate than its fall?
Rises have a founding date, a protagonist, a ribbon to cut. Decline is the slow failure of a supply chain link by link, with no attributable event. Rome did not die from one breached wall — the grain route, the aqueducts and the fiscal transfers went one at a time, and no single day of it was newsworthy. Hence a durable bias: we celebrate completion and argue about maintenance — and the first line cut from any budget is the one whose effect nobody can see.
European urban autonomy versus the administrative Chinese city — which was more “advanced”?
Wrong question. Weber's contrast rests on a real observation (a corporate community versus a prefectural seat), but reading it as a ranking is teleology. Northern Song Kaifeng had the densest commerce on earth in the eleventh century, at the price of no legal standing independent of the court; the Italian communes had that standing, at the price of permanent internal war and a ceiling on scale. Borrowing Ray Huang's angle: one lowered coordination cost through bureaucratic hierarchy, the other raised the rate of trial and error through polycentric competition.
Are redundancy and veto power two faces of the same thing?
The doubled pipe and the community veto are both damping reserved against an uncertain future; the difference is liquidity. Redundancy is a usable reserve that only pays off when spent; a veto is an unusable reserve that only accumulates. Distributed systems obey the same rule — replicas raise availability, while a consensus protocol in which everyone holds a blocking vote stops moving as soon as it grows. What you want is depth of the first and a strictly limited roster of holders of the second.