Edo: 268 Years of Peace No One Quite Believed Would Last

Between the rifle smoke clearing at Sekigahara in October 1600 and the rifle smoke rising at Toba-Fushimi in January 1868, Japan went 268 years without a major civil war. That is the headline of the Edo era, and it is the only feudal peace of comparable length anywhere on the planet. Even contemporaries struggled to believe their good luck. Almost every shogun spent his reign tightening the bolts of a system designed by Tokugawa Ieyasu and built on the fear that the next defection, the next famine, the next cult, would bring it all down.

I have spent years walking the ground that this era left behind, from the smoke-blackened beams of Edo Castle’s surviving baileys to the mossed-over post stones along the Nakasendō. The strangest thing about visiting Edo-period Japan today is realising how recent it feels. The 15th and last shogun, Tokugawa Yoshinobu, lived until 1913, took up cycling and oil painting in retirement, and posed for photographs in a three-piece suit. The world that had ended forty-five years earlier was still close enough to touch.

This is the long version of how Japan got 265 calm years out of the cabinet of a single warlord clan, what the price of that peace turned out to be, and where you can still walk through the architecture and policy traces of the regime that produced it.

Sekigahara to Edo: how a battle on a fog-bound morning produced a country

The Edo era does not begin where you might expect. It begins on a foggy October morning in 1600 in a narrow valley in central Mino Province, where roughly 160,000 men spent six hours killing each other so that one of them could later be appointed shogun.

I have walked the Sekigahara battlefield in autumn rain, and the eerie thing about it is how small the killing ground actually is. You can stand on the spot where Kobayakawa Hideaki defected and see almost the entire field that decided the next 268 years. The detailed mechanics of that morning live in my full piece on Sekigahara, the battle decided before lunch.

What matters here is the politics of the morning after. Ieyasu did not become shogun in 1600. He waited three years, doing what he had spent his whole life doing, which was nothing visible.

The Imperial title of Sei-i Taishōgun finally arrived from Emperor Go-Yōzei on March 24, 1603, and within two years Ieyasu had resigned the post in favour of his son Hidetada. He then ruled from semi-retirement at Sunpu for another eleven years. The lesson, repeated through the entire era, was that real power and the title for it should never sit in the same chair. You can read the long version of his career in Tokugawa Ieyasu, the patient man who outlasted everyone.

The unfinished business was Osaka. Toyotomi Hideyori, son of the man who had unified Japan a generation earlier, still held the largest castle in the country and a residual claim to legitimacy. Ieyasu manufactured a pretext over an inscription on a temple bell, laid siege to Osaka through the winter of 1614, breached the outer moats during a peace truce, and finished the job in the summer of 1615 with about 165,000 troops.

By the time Ieyasu died in June 1616, no rival house remained that could plausibly contest the Tokugawa. The Edo era, as a stable thing, dates from that summer.

The bakuhan system: 270 daimyō pinned to a chessboard

The political machine the Tokugawa built is usually called the bakuhan system, from bakufu (the shogun’s tent government in Edo) and han (the roughly 270 domains run by the daimyō). The shogun directly controlled about 7 million koku of rice-productive land out of a national total of around 26 million. That alone made the Tokugawa the largest single landlord in Japan by a factor of three. Everyone else held land at the shogun’s pleasure, with a contractual catch.

The catch was that the daimyō were sorted into three rings of trust. Shinpan were Tokugawa cadet branches, holding domains nearest to Edo. Fudai were houses that had been Tokugawa retainers before 1600, eligible for shogunate office and stationed across the strategic interior.

Tozama were the “outside lords” who had submitted only after Sekigahara, held the largest peripheral domains, and were forever ineligible for the central councils. The houses that eventually overthrew the regime, Satsuma and Chōshū, were both tozama, and they were never going to be anything else.

On top of this map of trust the bakufu placed a series of laws called the buke shohatto, the warrior house regulations, first issued in 1615 and updated through the era. Daimyō could not repair their castles without permission, could not coin money, could not marry into other daimyō houses without bakufu approval, and after 1615 could keep only one fortified castle per province under the ikkoku ichijō rule. About 170 castles were demolished in the first wave alone. The full landscape of what survived is in my guide to Japan’s 12 original castles.

Sankin-kōtai: the most expensive commute in pre-industrial history

The mechanism that broke the daimyō financially before they could break the shogunate militarily was sankin-kōtai, the alternate-attendance policy. Tokugawa Iemitsu, the third shogun, made it compulsory for tozama in 1635 and for fudai by 1642.

Every daimyō spent alternating years in his domain and in Edo. His wife and heir lived permanently in Edo as hostages. He travelled with a retinue scaled to his rice income, which in some cases meant more than 4,000 retainers strung out for kilometres along the Tōkaidō.

The receipts were ruinous. Modern estimates put sankin-kōtai expenditure at 70 to 80 percent of an average daimyō’s annual budget, with the Maeda of Kaga reportedly hitting figures even higher in some years. The Edo residences alone soaked up land, retainers, and prestige spending.

Inns along the Five Highways got rich, the road network expanded into a world-class system, and the daimyō slowly went broke. The whole machine is dissected in my standalone piece on how sankin-kōtai ate 75% of a daimyō’s income.

I find the side-effects more interesting than the main effect. Sankin-kōtai accidentally built a national tourism industry, an integrated postal courier service, the woodblock-print travel guide as a literary genre, and Edo itself as a consumer city. By 1700 the capital held about a million people, more than London or Paris at the time. Nearly half of them were samurai or their dependents, lodged in domain compounds that ringed Edo Castle and survive today as the addresses of central Tokyo ministries.

The four classes that Confucianism never quite enforced

Edo society was officially organised as shi-nō-kō-shō, the four-class hierarchy of samurai, peasants, artisans, and merchants in descending order. Below the four classes sat the eta and hinin, outcaste groups who handled tanning, butchery, and execution duties and did not receive legal recognition until 1871. Above all of it sat the imperial court in Kyoto and the daimyō, who counted as a separate caste of higher samurai.

The official theory was Neo-Confucian, imported from Ming China by Hayashi Razan and adopted by Ieyasu as the bakufu’s intellectual armour. It said merchants were the lowest productive class because they produced nothing but profit. The actual economy disagreed.

By the Genroku period in the 1690s, Osaka rice merchants were running futures contracts on the Dōjima exchange that resemble modern derivatives, samurai were borrowing from townsmen at compound interest, and the largest single creditor of the daimyō class was the Kōnoike merchant house. The pyramid was upside down and everyone knew it.

Samurai numbered about 400,000 households, roughly six percent of the population. They were a hereditary caste paid in rice stipends from the shogun or their daimyō. By the 18th century these stipends had not kept up with inflation, and many low-rank samurai were living in genteel poverty, taking second jobs as paper-umbrella makers or sword-tutoring instructors. The cultural products you most associate with the era, the swordsmanship treatises and the poetry, were often written on the spare time of an underemployed retainer who could not afford to drink.

The pyramid had a sex too. Women of the samurai class were locked into a system that combined arranged marriage, primogeniture, and the household codes of behaviour written down in books like Kaibara Ekken’s Onna Daigaku. Women of the merchant class, by contrast, often ran the actual shops, kept the books, and inherited businesses if no son survived.

There are surviving Edo-period account books in the hand of women who outlived two husbands and ran the family kimono shop better than either of them. The official theory and the lived reality of who was in charge of what diverged sharply, especially in cities, and the bakufu mostly looked the other way as long as the rice tax came in on schedule.

Outside the four classes sat the imperial court at Kyoto, which the Tokugawa kept on a tight financial leash and a generous ceremonial one. The emperor and his nobles received fixed stipends, were forbidden from political activity, and were maintained as a sacral institution which, in theory, legitimised the shogun. The Kinchū narabini Kuge Shohatto laws of 1615 reduced the court to ritual functions, with the imperial residence at Gosho confined behind walls and a clear bakufu commissioner permanently attached.

The court survived two and a half centuries on a budget that some senior daimyō households exceeded with their household entertainment expenses alone. That this fragile, ceremonial institution would eventually be the rallying point that overthrew the regime is one of the era’s bitter ironies.

1635: the year Japan locked the door

The seclusion edicts the Tokugawa issued between 1633 and 1639 are usually grouped under the term sakoku, “closed country,” though the term itself was a 19th-century coinage. The 1635 edict was the strictest. No Japanese could leave the country, no Japanese already abroad could return, and any ship attempting either was to be destroyed and its crew executed.

The Portuguese were expelled in 1639. Only the Dutch and the Chinese were permitted to keep trading, and only at Nagasaki, and the Dutch were confined to Dejima, an artificial island in Nagasaki Bay barely bigger than a city block.

The motive was Christianity, or more precisely the political risk that came with it. By 1614, when Ieyasu issued the first proscription, Japan had perhaps 300,000 baptised Christians, mostly in Kyushu, many of them tied to daimyō who had converted under Jesuit influence. The bakufu read the Spanish conquest of the Philippines as the playbook and decided to close the door before the same thing happened in Kyushu.

The Shimabara Rebellion of 1637-38, in which 37,000 mostly Christian peasants and rōnin held Hara Castle for nearly four months against a 125,000-strong shogunate army, removed any remaining doubt. Christianity went underground for 230 years and emerged only after Perry.

Sakoku was never as absolute as the name suggests. Korean diplomatic missions came through the Sō clan of Tsushima, the Shimazu of Satsuma ran the Ryūkyū kingdom as a tributary trade pipeline, the Matsumae clan handled Ainu trade in southern Hokkaido, and Dutch merchants delivered annual ship-loads of European books and instruments to Dejima.

From those books an entire intellectual subculture called rangaku, “Dutch learning,” developed inside Japan. By the 1770s, Sugita Genpaku and Maeno Ryōtaku were translating European anatomy texts into Japanese on the floor of an Edo townhouse without a dictionary, working from a single Dutch copy of a German anatomical atlas. The country was sealed, but the leaks were enormous.

Genroku: when peace started paying for itself

By the time the Genroku era arrived in 1688, the country was three generations into peace and the surplus had nowhere to go but into culture. Edo, Osaka, and Kyoto each topped 400,000 to a million inhabitants. Literacy among samurai approached 100 percent and among urban commoners reached an estimated 40 percent for men by the late period, the highest figures in the world before the industrial revolution. There was an audience for everything, and an artisan class hungry to supply it.

Ukiyo-e, the woodblock prints of the “floating world,” exploded under Hishikawa Moronobu in the 1670s and matured in Genroku. By the late 18th century the Tōshūsai Sharaku portraits of kabuki actors and the Kitagawa Utamaro studies of pleasure-quarter beauties were genre-defining. In the early 19th century Hokusai produced the Thirty-six Views of Mount Fuji and Hiroshige produced the Fifty-three Stations of the Tōkaidō, and you have probably seen both even if you have never been to a museum. Walk down certain streets in Asakusa today and you can still see the woodblock-print proportions of the eaves on the older shopfronts.

Theatre kept pace. Kabuki, born in Kyoto around 1603 in the dry riverbed performances of a shrine maiden named Izumo no Okuni, mutated through a women-only phase that the bakufu banned in 1629, a young-men phase banned in 1652, and finally settled into the all-male yarō kabuki form that survives today. Bunraku, the puppet theatre, peaked under playwright Chikamatsu Monzaemon, whose 1703 work The Love Suicides at Sonezaki sent so many young couples to commit shinjū that the bakufu repeatedly tried to ban the genre.

None of it worked, because the audience kept paying. To trace where puppet theatre eventually intersects with Japan’s older performing tradition, you can read my piece on the way of tea that killed its master as a parallel case of Edo aesthetic refinement.

Haiku found its definitive voice in Matsuo Bashō, a samurai’s son turned wandering poet who walked the Tōhoku roads in 1689 and produced Oku no Hosomichi, “The Narrow Road to the Deep North,” a travel diary so condensed that schoolchildren still memorise it. He died in Osaka in 1694 with the haiku “On a journey, ill, my dream goes wandering over withered fields,” which is the shortest summary of the Edo travelling-poet ethos you will ever find.

Prose came along too. Ihara Saikaku’s The Life of an Amorous Man appeared in 1682, followed by his Eternal Storehouse of Japan in 1688, both of which read like sociological field notes from a culture suddenly able to print 5,000-copy novel runs at affordable prices.

By the 1750s the kibyōshi, “yellow-cover books,” were essentially graphic novels carrying topical satire, and the bakufu’s Kansei reformers tried unsuccessfully to ban them. Print runs of 10,000 became routine for popular works in the early 19th century. The Edo population was reading.

If you want one walkable trace of what Genroku culture meant, take the train to Asakusa and find Sengaku-ji temple, where the forty-seven rōnin are buried beside their lord. The graves form a small open square, the incense never quite stops smoking, and the December anniversary still draws thousands. A short walk from there sits the Edo-Tokyo Museum complex in Ryōgoku, currently undergoing major renovation through the late 2020s, but the surrounding district still preserves the workshop streets and the ryōgoku riverside where Hokusai and Hiroshige sketched and where the great fireworks displays have run since 1733. You can also still see kabuki at Kabukiza in Ginza on the same site that has hosted theatres in some form since 1889, with the present building reconstructed in 2013.

The Akō Incident: when 47 men solved the wrong problem with the right answer

Of every story Edo Japan generated, the one that has refused to die is the Akō Incident of 1701-1703. In April 1701 the young daimyō of Akō, Asano Naganori, drew his short sword inside Edo Castle and slashed at Kira Yoshinaka, a senior bakufu protocol officer who had reportedly humiliated him over court etiquette.

Drawing a blade inside the shogun’s residence carried automatic death. Asano was ordered to commit seppuku that same afternoon. His domain was confiscated and his samurai dismissed.

Forty-seven of those samurai, led by Ōishi Kuranosuke, spent the next twenty-two months drinking publicly, neglecting their families, and otherwise persuading the bakufu spies that they had given up. On the snowy night of January 30, 1703, they assaulted Kira’s Edo residence in two coordinated columns, killed sixteen of his retainers, and beheaded Kira at the bottom of a charcoal storehouse. They carried the head across Edo to Sengaku-ji temple, washed it at Asano’s grave, and surrendered to the authorities.

The bakufu deliberated for forty-six days. The legal problem was that the men had carried out a vendetta the bakufu would normally have approved, but they had done it without permission, in defiance of the law that was supposed to be the entire basis of Tokugawa peace. Eventually the shogun’s councillors found the only answer that worked, which was to allow the rōnin the honour of seppuku rather than execution as common criminals.

Forty-six died on March 20, 1703. Terasaka Kichiemon, dispatched as a courier the night of the raid, was pardoned and lived to be 87. The story passed almost immediately into bunraku and kabuki as Chūshingura, “The Treasury of Loyal Retainers,” and is still being retold every December.

The Akō story matters as much for its reception as its events. Within months of the executions, popular printers were issuing illustrated chapbooks. Within fifteen years bunraku puppet productions were touring under the title Chūshingura.

By the 1820s every major kabuki actor had played Ōishi, every senior ukiyo-e artist had produced a complete forty-seven portrait set, and the cult of the rōnin had become a kind of unofficial civic religion. The bakufu disliked it but never managed to suppress it, partly because the official verdict had honoured the men with seppuku, and partly because the audience was simply too large. The Akō story is the moment Edo popular culture realised it could move a moral verdict against bakufu policy and live to tell the tale.

The economy slowly bends, then breaks

Beneath the cultural surface the rice economy was struggling. The fundamental problem was that the bakufu and the daimyō taxed in rice but increasingly spent in cash, and the merchants who converted rice to cash kept the spread. The fifth shogun Tokugawa Tsunayoshi, who reigned from 1680 to 1709, debased the gold coinage to plug deficits and is best remembered for his Laws of Compassion which made killing dogs a capital crime. There is more on him in my piece on the shogun who tried to legislate kindness.

The eighth shogun, Tokugawa Yoshimune, took office in 1716 and ran the Kyōhō Reforms through the 1720s and 1730s, returning the coinage to higher purity, encouraging new rice varieties and crop rotation, and personally taking up falconry as a public symbol of austerity. His reforms bought thirty years of stability, but the Kyōhō Famine of 1732, which killed an estimated 12,000 people in western Japan after locusts destroyed the rice crop, exposed the system’s brittleness. The peasantry began to riot. Between 1700 and 1850 there were more than 3,000 recorded peasant uprisings, mostly small and localised, but the curve trended upwards through the late period.

Matsudaira Sadanobu’s Kansei Reforms of 1787-93 doubled down on the Confucian script, restricting heterodox philosophy and pushing samurai back to traditional martial training. The Tenpō Reforms of 1841-43 under Mizuno Tadakuni went further, attempting price controls and forced returns of urban migrants to their villages. None of it worked. What I take from the late-Edo reform record is that everyone in power could see the economic problem clearly, but the structural solution would have required dismantling the four-class system and the rice-tax base, which is to say dismantling the regime itself.

The Tenmei Famine of 1782-88 and the Tenpō Famine of 1833-37 were the two worst rural disasters of the period, killing perhaps 920,000 and 1 million people respectively across northeastern Japan. The Tenmei catastrophe coincided with the eruption of Mount Asama in 1783, which dimmed harvests across the entire Kantō plain. Peasant uprisings and urban rice riots multiplied. In Osaka in 1837, a former bakufu official named Ōshio Heihachirō led a Confucian-inspired armed rebellion against rice merchants and the bakufu commissioner, which was crushed in a day but which sent a powerful message: the educated samurai class itself was beginning to look at the regime as illegitimate.

Black ships at Uraga, July 1853

What broke the regime in the end was not a peasant uprising or a daimyō revolt. It was four steam-powered American warships under Commodore Matthew Calbraith Perry, dropping anchor off Uraga at the mouth of Edo Bay on July 8, 1853. He carried a letter from President Millard Fillmore demanding a treaty.

Perry let the bakufu observe his pivot guns and his coal smoke, said he would return for an answer in a year, and left. The bakufu had no answer that did not amount to surrender. The next 15 years are called the Bakumatsu, the “end of the bakufu.”

The bakufu’s response to Perry was, in effect, a national psychological breakdown. Senior councillors did the unthinkable and circulated the American demands to all daimyō for opinion, which was the first time in 250 years that the bakufu had publicly admitted it could not decide a policy alone. Some daimyō wrote in favour of fighting, some in favour of trading, some in favour of waiting. The bakufu’s authority bled out from that moment, because the institutional fiction that the shogun’s word was final had been broken on paper, in writing, on the official record.

Perry returned in February 1854 with seven ships and stayed until the bakufu signed the Convention of Kanagawa on March 31, opening the ports of Shimoda and Hakodate, granting most-favoured-nation status, and accepting an American consul. Townsend Harris arrived as that consul in 1856 and by 1858 had negotiated the Harris Treaty, which opened Yokohama and four other ports, granted extraterritoriality, and fixed Japanese tariffs at five percent. The British, French, Dutch, and Russians signed parallel treaties within months. The closed country was over, on terms the bakufu had not chosen.

The economic shock was immediate. Japan’s domestic gold-to-silver ratio was about 1:5, against a world market ratio of 1:15. Foreign traders bought Japanese gold cheaply with imported silver and exported it. Roughly 70 tons of Japanese gold left the country between 1859 and 1860.

The bakufu re-coined to plug the leak, which produced inflation, which crushed samurai stipends, which radicalised the lower ranks. The slogan that began to spread was sonnō jōi, “revere the emperor, expel the barbarians,” and increasingly it pointed at the shogunate as the obstacle, not the solution.

Bakumatsu: assassination, alliance, and the slow surrender of authority

On a snowy March morning in 1860, eighteen samurai from Mito and Satsuma ambushed Tairō Ii Naosuke’s palanquin outside the Sakuradamon gate of Edo Castle and assassinated the regime’s most powerful councillor. Ii had pushed through the Harris Treaty without imperial sanction and had purged dissident nobles in the Ansei Purge. His killing announced the new political reality: the shogunate could no longer protect its own ministers in its own capital. The Mito ronin handed themselves in afterwards, and most were executed, but the message was already delivered.

The next eight years were a continuous low-grade civil war, fought through assassinations, court politics, and three full bakufu expeditions against Chōshū. The Hamaguri Gate Incident of August 1864 saw Chōshū troops burn Kyoto in a failed attempt to seize Emperor Kōmei.

Foreign navies bombarded Kagoshima in August 1863 and Shimonoseki in September 1864, demonstrating to Satsuma and Chōshū respectively that “expel the barbarians” was a slogan the West could refute with broadsides. By 1866, secretly brokered by Sakamoto Ryōma, Satsuma and Chōshū had formed an alliance against the bakufu. The figure most associated with the next phase, Saigō Takamori, runs through my profile of the Meiji hero who became Meiji’s enemy.

Tokugawa Yoshinobu, who had become the 15th shogun in August 1866 on the death of Iemochi, did the only thing left to do. On November 9, 1867, he announced the taisei hōkan, the formal return of governing authority to Emperor Meiji, in the Ōhiroma of Nijō Castle. He hoped to preserve a Tokugawa-led council under the imperial banner.

Satsuma and Chōshū had no intention of allowing this. On January 3, 1868, in the Imperial Palace, they engineered a court coup that abolished the shogunate outright and stripped Yoshinobu of his lands.

Yoshinobu was an unusual choice for last shogun. He had been raised at Mito, the most ideologically nationalist of the Tokugawa cadet domains, where his father Tokugawa Nariaki ran an academy that produced many of the early sonnō jōi activists. He spoke French, sat for portrait photographs, and reportedly read European military manuals.

Among his earliest acts as shogun was to commission a French military mission to retrain the Tokugawa army and a French engineer to begin building a modern arsenal at Yokosuka. Had he been crowned in 1855 instead of 1866, the Edo regime might have modernised in time. Crowned in 1866, with Satsuma already in coordinated revolt, he had no time and no allies left.

Boshin War and the bloodless surrender of Edo

The Boshin War opened on January 27, 1868, at Toba and Fushimi south of Kyoto, where roughly 5,000 imperial troops, mostly Satsuma and Chōshū with new Minié rifles, broke a Tokugawa force more than three times their size. Yoshinobu, who was not on the field, fled to Edo by ship. Saigō Takamori marched east as commander of the imperial Eastern Army, and by April his forces stood at the gates of Edo. The 30-square-kilometre city held perhaps 1.3 million people, mostly civilians.

What happened next is one of the strangest negotiated surrenders in modern history. Yoshinobu’s emissary Katsu Kaishū, the bakufu’s navy minister, met Saigō at Satsuma’s Edo residence on April 5 and 6, 1868. Yoshinobu would surrender Edo Castle, accept house arrest in Mito, and the imperial side would not loot, burn, or massacre.

Edo Castle was handed over peacefully on May 3. The bakufu fleet under Enomoto Takeaki sailed north to keep fighting, but the city of Edo passed to the new government without a shot fired. The whole episode is at the centre of my standalone piece on the castle that became the Tokyo Imperial Palace.

The last Tokugawa loyalists held out at Hakodate’s star-shaped Goryōkaku fort under Enomoto, declaring an Ezo Republic in December 1868. They surrendered on June 27, 1869, after an imperial naval assault. The Edo era, on any reasonable accounting, ended that morning. Yoshinobu, the man whose hereditary title had once meant supreme command of the Japanese military, lived another 44 years, was elevated to kōshaku (prince) in 1902, and died in 1913 quietly in Tokyo.

The fifteen Tokugawa shoguns: a quick map

Of the fifteen shoguns, only three or four did anything that the rest were not maintaining. Ieyasu (1603-1605) wrote the rulebook. Iemitsu (1623-1651) closed the country and made sankin-kōtai compulsory.

Tsunayoshi (1680-1709) presided over Genroku culture and broke the coinage. Yoshimune (1716-1745) ran the most successful reforms.

The shoguns of the long middle (Ietsugu through Iesada) were largely figureheads whose effective business was conducted by senior councillors. Ienari, who reigned 50 years from 1787 to 1837, fathered 53 children and managed to push the regime closer to bankruptcy than any of his predecessors. Yoshinobu (1866-1867) inherited an unsalvageable situation and spent fourteen months losing it gracefully.

The pattern that the names hide is institutional. The shogunate worked because the senior councillors, the rōjū, ran the government in monthly rotation regardless of whose seal sat on the documents. Roughly 17,000 bakufu officials staffed the various commissions that managed cities, finance, temples, and inspections.

When the institution broke under foreign pressure, the personalities of the last shoguns mattered very little. The system itself had run out of moves.

The Tokugawa intelligence apparatus: 17,000 officials and a culture of denunciation

Behind the visible courts and councils, the bakufu ran a domestic intelligence network of remarkable density. The metsuke were inspectors at the central level, watching daimyō for signs of rebellion. Below them, the ōmetsuke covered the senior daimyō themselves, while local yokome watched samurai households inside each domain.

Edo itself was policed by a small corps of senior machibugyō commissioners and a system of jishinban neighbourhood watchposts staffed by townspeople. The bakufu encouraged anonymous denunciation through public boxes called meyasubako first installed by Yoshimune in 1721, where any subject could drop in a complaint or accusation. Some boxes accumulated tens of thousands of submissions over the decades. The result was a city where everyone knew that everyone was being watched.

The intelligence machine had a religious arm too. After 1640, every Japanese household was required to register with a Buddhist temple in the terauke system, ostensibly to prove the household was not Christian, in practice to track births, deaths, and movements through a temple-based census. The Meiji government inherited this household registration system intact, renamed it koseki, and used it to draft soldiers, collect taxes, and track citizens for the next 80 years. It is still the legal backbone of Japanese family registration today.

Counting the cost of 268 years of peace

One number that is hard to absorb: in 1600 Japan had perhaps 12 million people, and by 1721 the population had stabilised at roughly 31 million, where it stayed until 1868. The first century of the Edo era was a sustained demographic boom, the next 150 years were essentially demographic stasis. The reasons are debated.

Some scholars point to the climate cooling of the Little Ice Age, some to deliberate fertility control through abortion and infanticide as response to the rice-tax system, some to the structural ceiling of an agriculture system that could not productively absorb more workers. Whatever the cause, by 1850 the country was running a balanced demographic ledger that masked deep regional collapses. The Tōhoku population fell during the late 18th-century famines and never fully recovered.

What you cannot see in those numbers is the extraordinary consolidation of culture inside them. The Edo era is when the country became Japan in a way that earlier centuries had not been. A national language, with Edo dialect at its core, slowly became standard through the woodblock-printed novel and the bakufu administrative document.

National highways tied the islands into a single travel network, while national currencies, after Yoshimune’s reforms, traded against each other at fixed rates. National educational forms, with terakoya in every village and hankō academies in every domain capital, ensured that a literate samurai or merchant moving between Kagoshima and Akita would find broadly the same intellectual furniture. The political fragmentation of Sengoku Japan, which had been the norm for two centuries, was replaced by a cultural integration that the Meiji state could plausibly call national.

Where you can still walk through the Edo era

The single most rewarding Edo-period site in the country is Edo Castle itself, now the eastern gardens of the Imperial Palace. The Honmaru and Ninomaru baileys are open daily except Mondays and Fridays, the Ōte-mon gate still carries the original Tokugawa hardware, and the foundation stones of the lost main keep, destroyed in the 1657 Meireki fire and never rebuilt, are accessible on the central plateau. The full visit guide sits in my dedicated piece on the castle that became the Tokyo Imperial Palace.

For Ieyasu the man, the destination is Nikkō Tōshō-gū, his deification shrine in the Kantō mountains. Construction began the year after his death in 1616 and reached its current form under Iemitsu in 1636. The shrine holds 8 national treasures and 34 important cultural properties, and the polychrome carving on the Yōmeimon gate alone is the densest surviving example of early Edo decorative woodwork in Japan. The site is easily reached on the Tobu Nikkō line from Asakusa.

For sankin-kōtai infrastructure, the best surviving fragment is the Hakone Sekisho checkpoint on the Tōkaidō, faithfully reconstructed in 2007 from Edo-period plans on the original site. The neighbouring Hakone-juku post town still preserves several inn buildings. The Kiso Valley posts of Tsumago and Magome on the parallel Nakasendō highway are the longer experience, an eight-kilometre walk between two preserved villages with cars banned during daytime hours and original cobblestones for much of the climb. I recommend doing the walk south to north, Magome up to Tsumago, because the gradient is gentler in that direction and the magome-jaya teahouse near the summit serves a properly seasoned oyaki to recover.

For the political theatre of the era’s end, Nijō Castle in Kyoto is non-negotiable. Built by Ieyasu in 1603 as the shogun’s Kyoto residence, it was the stage on which Yoshinobu announced the Taisei Hōkan in 1867. The Ninomaru Palace is a designated National Treasure and one of the most complete surviving Edo-period interior environments in the country, with original Kanō school sliding doors, “nightingale” floors that squeak deliberately to detect intruders, and the audience hall where the regime ended. Allow two hours minimum, and visit on a weekday morning.

For the Restoration side of the story, Hagi in Yamaguchi prefecture is the small castle town that produced an absurd density of Meiji-era figures: Yoshida Shōin, Itō Hirobumi, Yamagata Aritomo, Inoue Kaoru, Katsura Tarō. The Shōin Shrine and the surviving samurai districts can be walked in a half-day. Kagoshima, the old Satsuma capital, holds the Saigō Takamori statue at Shiroyama and the cave where he died in September 1877 during the Satsuma Rebellion. Hakodate, on Hokkaido’s southern tip, has the Goryōkaku star fort itself, walkable on raised paths over the original moats, with a small museum at the central administration building.

What the era left behind

What I think the Edo era ultimately bequeathed to the country is not the obvious thing. The obvious answer is the visual culture, the ukiyo-e and the kabuki and the haiku, and you can find variants of all three in the way contemporary Japan still markets itself. The deeper inheritance is bureaucratic.

The bakufu invented or refined the prefectural unit, the household registration system, the urban land tax, the highway post network, the licensed merchant guild, the population census, and a domestic intelligence service. The Meiji state inherited every one of these and ran with them.

The aesthetic that you now identify as traditional Japanese, the tatami room, the kimono, the cherry-blossom viewing, the tea ceremony, the formal garden, the matsuri parade, was largely codified or popularised under the Tokugawa. The textile rules of the kimono are unpacked in my piece on how to tell a kimono rank at a glance. The hanami tradition lives in my piece on the 1,300-year-old hanami tradition, and many of the festivals that visitors today queue for, including Asakusa’s three-day riot of mikoshi and the festival Tokugawa Ieyasu started in 1600, took their modern shape in this period.

What I find quietly remarkable is how many of the Edo era’s anxieties are still recognisable. The fear of foreign trade swamping a fragile domestic economy, the difficulty of taxing wealth that has migrated from land to commerce, the gap between an inherited ruling class and the people who actually run the markets, the slow erosion of legitimacy when the official ideology no longer matches what everyone can see, all of it survived the regime that produced it. Ieyasu would have recognised the structural shape of every Japanese fiscal crisis since 1990. The Tokugawa peace lasted because it was a working answer to those questions, not because it solved them, and the answer broke when the questions changed faster than the answer could.

If you only have one trip

If you can only fit one Edo-era trip into a Japan visit, I would build it around three nights in central Tokyo for Edo Castle and the Imperial Palace East Garden, a long day trip from there to Nikkō for the Tōshō-gū complex, and an overnight in the Kiso Valley to walk Magome to Tsumago at dawn before the day-trippers arrive. If you can stretch it to ten days, add three nights in Kyoto for Nijō Castle, then fly to Kagoshima for the Satsuma side, take the Sakurajima ferry, and return via Hagi for an afternoon. That itinerary covers Ieyasu’s tomb, his Edo capital, his Kyoto stage, his post-road infrastructure, and the two domains that finally undid his regime, in roughly the order the era itself unfolded.

For deeper context once you are there, I would also build in a slow afternoon at Sengaku-ji to visit the forty-seven graves and watch the local Tokyoites still leaving incense in front of Ōishi’s stone. Spend an evening watching kabuki at Kabukiza in Ginza, where the lobby still preserves the actor banners and the snack-vendor system that Edo audiences would recognise immediately. Stop at Tokyo National Museum’s Hōnkan in Ueno, where the standing collection of Edo-period swords, armour, kimono, lacquer, and ukiyo-e is one of the densest survival of period material culture in the world. The Hōnkan rotates exhibits regularly, but you can usually expect to see Hokusai or Hiroshige originals on display somewhere on the second floor, and the lacquerware galleries on the first floor are an underrated highlight.

Whatever you do, do not try to take this era in chronologically by reading dates in a row. It does not work. It works when you stand in one of these places and realise that the snowstorm in Hiroshige’s print, the carved peony on a Nikkō rafter, the rule about your wife being held in Edo while you commute from Kanazawa, and the four black ships that finally made the system stop pretending all happened to people who would have considered themselves contemporaries. The Edo era is one human lifetime, repeated for nine generations, in a country that mostly held its breath.