Sejong made an alphabet in 1443. Korea adopted it in 1894.

Before: a kingdom that wrote in a language nobody spoke

In the winter of 1443 the kingdom of Chosŏn had been in existence for fifty-one years, kept the fullest court records of any state on earth, and could not write down what its people said. Its historians were compiling the Veritable Records of each reign in literary Chinese. Its law code, the Kyŏngguk taejŏn, was being drafted in literary Chinese. Its civil examinations tested composition in literary Chinese, its diplomacy with the Ming was conducted in literary Chinese, and the land registers that determined what a farmer in Chŏlla owed were written in a language that farmer had never heard spoken.²² Korean itself — a language with a phonology and a grammar that Chinese shares almost nothing with — existed in the record only as a set of workarounds, and the workarounds had been in use, without fundamental improvement, for something close to eight hundred years.²¹

This is the calibration the rest of this record depends on, so it is worth being exact about it. Chosŏn Korea was not illiterate. It was among the most heavily documented societies in the world. It was a society in which writing and speech had been prised apart so completely that the two had become separate skills, and the skill that mattered was the one that took five years and a tutor to acquire.

Three systems, one prerequisite

Korean scribes had not been idle about the problem. By the Three Kingdoms period they had built three distinct methods of forcing Chinese characters to carry Korean, and all three survived into Sejong's lifetime:²¹¹

• Idu (吏讀, "clerk reading") — Chinese characters used partly for meaning and partly for sound, with a set of conventional characters appended to mark Korean case particles and verb endings. This was the working script of the clerical class: contracts, petitions, land deeds, local government paperwork. It was still being written in the nineteenth century.
• Hyangch'al (鄕札, "local letters") — a fuller system in which characters were pressed into representing all the sounds of a Korean text. It is preserved chiefly in the twenty-five surviving hyangga poems, and it was already archaic by 1400.
• Kugyŏl (口訣, "oral formulae") — annotation marks, often abbreviated characters, inserted into a Classical Chinese text to show a Korean reader how to parse it into Korean word order and where to supply grammatical endings. A reading aid, not a script.

Every one of these had the same prerequisite: you had to know Chinese characters first. Idu did not spare a clerk the memorisation of several thousand graphs; it only gave him a way to bend them. There was no route into Korean writing that did not run through the Chinese lexicon, which meant there was no route into writing at all for anyone without the years and the money to acquire it. Literacy in Chosŏn tracked class with almost no slippage, and the systems were not the cause of that — but they guaranteed that nothing would change it.¹⁰

What writing was for, and who it was for

It is also worth being exact about what the state wanted writing for, because the answer shapes the resistance that came later. Writing in Chosŏn was the medium of office. The examination system was the mechanism by which the yangban reproduced itself as a hereditary governing class, and mastery of literary Chinese was the barrier that made the mechanism work.²² A man who could compose an eight-legged essay in the Ming manner was a candidate for the state; a man who could not was, by definition, not. The barrier was expensive to clear and that expense was the point.

Against this sat a practical problem the court could not ignore. Sejong's government was an activist one — it published a farming manual in 1429, standardised rain gauges in 1442, recalculated the astronomical tables for the latitude of Hanyang — and its publications kept running into the same wall. In 1434 it issued the Samgang haengsil to, an illustrated compendium of exemplary filial and loyal conduct, with pictures precisely because the people it was aimed at could not read the text. That is a state discovering the limit of its own reach, and discovering it in the form of a picture book.

The transmission: twenty-eight letters, made in a palace, announced without warning

The notice of the twelfth month, 1443

The Sejong sillok, under the twelfth lunar month of the twenty-fifth year of the reign — a period falling between late December 1443 and late January 1444 in the Gregorian calendar — records that the king had personally made twenty-eight letters of a vernacular script, that the letters were called the Hunmin chŏngŭm, the Correct Sounds for the Instruction of the People, and that although simple they could be combined without limit.² The notice is brief and it is startling in what it does not contain. There is no commission, no committee, no prior debate in the record, no memorial proposing the work. The verb is 親制 — personally made. Whatever help Sejong had, the court's own official history says the alphabet arrived as the king's own object, presented to a bureaucracy that had not been consulted.

Historians have argued about that verb ever since, and the argument matters because the popular story of hangul is a story about the Hall of Worthies working collectively under a benevolent king. Lee Ki-Moon has made the sustained case for taking the sillok at its word: the design is too unified, too much the product of a single analytical intelligence, and the men who are supposed to have made it collectively are the same men who, a few months later, signed a memorial saying they would rather it had not been made.¹⁵ What the Hall of Worthies demonstrably produced was the explanation of the letters, three years later. The letters themselves appear in the record fully formed, out of a palace, in a month when nobody was expecting them.

How the letters were built

What Sejong's court produced was not a set of signs. It was a phonology, rendered as shapes.

The Hunmin chŏngŭm haerye of 1446 states the design principle without hedging: five basic consonant letters are diagrams of the organs that make the sound.³ ㄱ is the root of the tongue blocking the throat. ㄴ is the tongue touching the upper palate. ㅁ is the outline of the mouth. ㅅ is the incisor. ㅇ is the throat. From these five, the rest of the consonant inventory is generated by adding strokes, and the added stroke means something: it marks the sound as more strongly articulated. ㄱ becomes ㅋ; ㄴ becomes ㄷ and then ㅌ; ㅁ becomes ㅂ and then ㅍ. The visual relation between two letters encodes the phonetic relation between two sounds, which is a thing no other script in the world was doing in 1443, and which nobody in Europe would think to do before the nineteenth century.

Government of Chosŏn, Hunmin chŏngŭm haerye, 1446. Kyujanggak Institute for Korean Studies, Seoul National University. Public domain via Wikimedia Commons. · Public domain

The vowels work on a different and more openly cosmological principle. Three elements stand for heaven (a dot, ㆍ), earth (a horizontal, ㅡ) and the standing human being (a vertical, ㅣ), and the remaining vowels are built by combining them, with the position of the dot — above or below the horizontal, right or left of the vertical — determining whether the vowel counts as yang or yin.³⁷ That distinction was not decorative. Fifteenth-century Korean had vowel harmony, and the letter shapes encode which vowels may appear in a word together. The Neo-Confucian cosmology in the Haerye's exposition is not a varnish applied to a linguistic result; it is the conceptual apparatus the designers were using to think with, and Kang Sinhang's edition remains the standard guide to how much of the argument is genuinely phonological and how much is Song dynasty numerology.⁷

Finally, the letters are not run in a line. They are stacked into syllable blocks, so that a written Korean text occupies a grid of the same visual weight as a Chinese one. This is often read as deference to Chinese aesthetics, and it partly is. It is also a decision that keeps the syllable — the unit Korean phonology actually operates on — visible on the page, and it is the reason a Korean text can be scanned at the speed a Chinese text can.¹⁴

Geoffrey Sampson's 1985 survey of the world's writing systems gave hangul a category of its own on the strength of this: a featural script, in which the graphemes encode articulatory features rather than merely segments. His assessment has been quoted so often in Korea that it has become a kind of national epigraph, and it is worth quoting exactly, because the hedge in the first clause is part of what makes it credible: "Whether or not it is ultimately the best of all conceivable scripts for Korean, Han'gul must unquestionably rank as one of the greatest intellectual achievements of humankind."⁶

Where the shapes came from, and the argument that will not close

There is a complication in the record, and the honest version of this story does not paper over it. The same sillok notice of 1443 that announces the twenty-eight letters says that they "imitate the Old Seal" — 古篆.² The Haerye of 1446 says the consonants imitate the speech organs.³ These are not the same claim, and scholarship has spent sixty years on the gap.

Gari Ledyard's 1966 Berkeley dissertation proposed the most influential answer: that a handful of basic consonant shapes were adapted from the ʼPhags-pa script, the alphabet the Mongol Yuan had commissioned from a Tibetan lama in 1269 and which Korean officials of the late Koryŏ period would have known well. Earlier writers had tried to derive ten or more hangul letters from ʼPhags-pa; Ledyard judged most of those derivations contrived and argued for five — ꡂ to ㄱ, ꡊ to ㄷ, ꡙ to ㄹ, ꡎ to ㅂ, ꡛ to ㅅ and ㅈ — with the remaining consonants generated from those by the Haerye's own stroke-addition rule.⁴ By 1997 Ledyard had also concluded that the alternative reading of 古篆 as Chinese seal script was hopeless: the tortuous curves of seal forms have nothing in common with hangul's geometry.⁵ A third position, associated with Kim Wan-jin, derives the shapes instead from simplified kugyŏl characters, on the analogy of how Japanese katakana emerged from abbreviated Chinese graphs.

The position this record takes is that these are not competing accounts of the same thing. The articulatory analysis is the design principle — it is what makes the system work, it is what the designers wrote down, and no borrowing hypothesis explains it. The ʼPhags-pa resemblance is a plausible account of where some of the strokes came from, and of what gave a fifteenth-century Korean king the idea that a small alphabet was a thing a state could commission at all. A court that had spent a century as a Mongol son-in-law state knew that scripts were made objects. That knowledge is itself a transmission, and it is the quiet Mongol contribution to a Korean achievement.

Ch'oe Malli's memorial, 1444

In the second lunar month of 1444, within weeks of the announcement, Ch'oe Malli, first counsellor of the Hall of Worthies, submitted a memorial against the vernacular script on behalf of himself and six colleagues.⁹ It is one of the most valuable documents in this record, because it is the receiving culture's elite explaining, in its own words and without embarrassment, exactly why it did not want the gift.

The memorial's arguments, set against what each was actually defending:

Read the fifth row again. It is the only one of the six that is not a cultural argument at all, and it is the one that gives the game away. Ch'oe Malli was not confused about whether the alphabet worked. He had evidently grasped that it worked extremely well, and had followed the consequence through to its end: that if a man could become an official without the Chinese classics, the classics would stop being the thing that made officials. The memorial is not a failure of imagination. It is an accurate piece of forecasting by a man who did not like what he forecast.

Sejong dismissed Ch'oe from office and reinstated him the following day. Ch'oe resigned, went home to his village, and died on 23 October 1445 — a year before the script he opposed was formally promulgated. Chŏng Ch'angson, another of the signers, was dismissed outright after arguing that translating the Samgang haengsil into the vernacular would not make people virtuous.²⁹ It is worth holding onto the modesty of these punishments. Nobody was killed over the alphabet. The Chosŏn court argued about it, and the argument was won by the king and lost by the bureaucracy, and then the bureaucracy spent four hundred and fifty years winning it back by other means.

The Haerye, 1446

The formal promulgation came in the ninth lunar month of 1446, with a volume that has no parallel in the history of writing: the Hunmin chŏngŭm haerye, in which the makers of a script explain, letter by letter, why each one has the shape it has. Sejong wrote the preface and the presentation of the basic letters; the commentary was composed by Chŏng Inji together with Ch'oe Hang, Pak P'aengnyŏn, Sin Sukchu, Sŏng Sammun, Kang Hŭian and Yi Kae — Yi Sŏllo is counted as an eighth hand in some accounts — and runs in six parts: the design of the letters, the initials, the medials, the finals, the combination of letters, and examples of use.³⁸

Sejong's preface is the most quoted Korean text in existence and deserves to be given in full rather than paraphrased: "The sounds of our country's language are different from those of the Middle Kingdom and are not confluent with the sounds of characters. Therefore, among the ignorant people, there have been many who, having something they want to put into words, have in the end been unable to express their feelings. I have been distressed because of this, and have newly designed twenty-eight letters, which I wish to have everyone practice at their ease and make convenient for their daily use."³

Chŏng Inji's postface makes the practical claim that has been tested ever since and has never failed: that "a wise man can acquaint himself with them before the morning is over; even a stupid man can learn them in the space of ten days."³

There is a second motive in the record, and flattening it out does the king no favours. Within a year the same letters were serving the Tongguk chŏngun, a rime dictionary fixing the "correct" Sino-Korean readings of Chinese characters, compiled by Sin Sukchu, Ch'oe Hang and Pak P'aengnyŏn, completed in 1447 and published in 1448.¹⁷¹ Several letters in the original twenty-eight never wrote a native Korean word; they exist to transcribe Chinese initials. The Hunmin chŏngŭm was one instrument built for two jobs — to give Korean a script, and to give Korean Sinology a phonetic notation — and Ch'oe Malli's memorial was aimed at least as much at the second job as the first. The popular telling keeps the first purpose and drops the second, which makes Sejong more sentimental and less interesting than he was.

What changed, and what was replaced

The first books, 1447–1459

The court moved quickly, and the first decade of vernacular publishing tells you precisely what the court thought the script was for.

Two things stand out. The first is that the alphabet's earliest sustained use was dynastic propaganda and Buddhist devotion — the Yongbi ŏch'ŏn ga exists to prove that the Yi seizure of the throne had been mandated by heaven, and it is the new script's first major employment¹⁹ — one to legitimise the house, one to serve a religion the state's own Neo-Confucian ideology was busy suppressing. The second is visible only in the layout.

Cultural Heritage Administration of the Republic of Korea (문화재청). Wŏrin ch'ŏn'gang chi kok, vol. 1, 1447; National Treasure No. 320. CC BY-SA 4.0 via Wikimedia Commons. · CC BY-SA 4.0

In almost every other early vernacular book, hangul appears in small type beside dominant Chinese characters: a gloss, an aid, something subordinate. In the Wŏrin ch'ŏn'gang chi kok the relation is reversed, the Korean letters set large and the Chinese reduced to annotation.¹⁸ That is what the script looks like when someone has decided it is the text rather than the apparatus, and it would be roughly four hundred and fifty years before a Korean government printed anything that way again.

The script enters the clerical examination

Sejong did not stop at publishing. In the tenth month of 1446 he submitted a long report written in the new script to the State Tribunal concerning the conduct of the Office of the Inspector-General — the king using the vernacular in a document of real administrative weight, months after promulgation. On the twenty-sixth day of the twelfth month of 1446 he ordered the script included in the state examinations for clerical officials, and on the twentieth day of the fourth month of 1447 extended the requirement to the provincial examinations.²¹

This was the moment Ch'oe Malli had forecast, and it was the high-water mark. The requirement touched the clerical examinations — the route into the sub-bureaucracy, not into the yangban officialdom proper — and it did not spread upward. The great civil examination, the one that made careers and families, went on being conducted in literary Chinese for another four hundred and forty-seven years. Sejong had put his alphabet into the antechamber of the state and there it stayed.

He died in April 1450. In 1456 his grandson Tanjong's supporters attempted to restore the boy king Sejo had deposed; the plot was betrayed, and among the six men executed for it were Sŏng Sammun, Pak P'aengnyŏn and Yi Kae — three of the scholars who had written the explanation of the letters ten years earlier. Sejo abolished the Hall of Worthies the same year. Sin Sukchu, their colleague on the Haerye and on the rime dictionary, served the new king and rose to chief state councillor. The institution that explained the alphabet was destroyed within a decade of explaining it, and it was destroyed for reasons that had nothing to do with the alphabet.

1504: Yŏnsan'gun burns the books

On the nineteenth day of the seventh lunar month of 1504, in the middle of the second of the great literati purges, anonymous posters written in the vernacular script appeared in Hanyang mocking Yŏnsan'gun for his tyranny. His response was immediate and total. He banned the study, the teaching and the use of the script; ordered books written in it burned; and directed the five administrative districts of the capital to identify and report every person known to be literate in it, with punishment for neighbours who failed to denounce.¹¹⁶

This is the only state persecution in the record that is unambiguously a persecution of the script, and it deserves the specificity the atlas owes it. What is being described is a literacy register, compiled by a city's own residents against each other, on pain of punishment for silence. The ban ran about five months. Yŏnsan'gun himself ended it in the twelfth month by ordering calendar books translated into the script — an autocrat rediscovering that the tool is useful once it has stopped being pointed at him. Two years later, after Yŏnsan'gun's deposition, Chungjong abolished the Ŏnmunch'ŏng, the government office for the vernacular script, altogether.¹

After 1506 the Chosŏn state had no institution responsible for hangul, no requirement that anyone learn it above the clerical grade, and no programme of any kind for its use. It had, in other words, arrived at the settlement Ch'oe Malli had asked for sixty years earlier, and it kept that settlement until 1894.

The long margin

And the script did not die. This is the part of the story that makes it worth six thousand words rather than six hundred, and it is the part that the phrase "successfully popularized" both captures and obscures. Hangul survived four centuries with no institutional support whatever, carried by exactly the people whose literacy the state did not count.

Letter of Wŏn's mother, Andong, 1586, excavated from the tomb of Yi Ūng't'ae in 1998. Andong National University Museum. Public domain via Wikimedia Commons. · Public domain

The physical evidence is unambiguous and it comes from ordinary places. The oldest surviving hangul letter was written in the 1490s by a man named Na Sin'gŏl to his wife. In April 1998, an archaeological team from Andong National University, clearing tombs on the south bank of the Naktong ahead of an urban renewal project, opened the grave of Yi Ŭngt'ae, a provincial gentleman of the Kosŏng Yi lineage who had died in 1586 at thirty-one. On his chest lay a letter from his pregnant widow, written in the vernacular script, in colloquial Middle Korean, addressing him directly and asking how he could have gone ahead of her. The tomb also held sandals she had woven from hemp and her own hair. That is a household 140 years after promulgation, in a provincial county, with no state encouragement of any kind, writing its private grief in the king's alphabet — and doing it so naturally that the choice of script evidently required no thought at all.

The pattern that produced that letter has three main strands:

1. Women's correspondence. The great majority of surviving private letters from Chosŏn are in hangul, and women wrote and received most of them. Women were excluded from the examinations and so from the Chinese literacy the examinations enforced; the alphabet was the only writing available to them, and they used it comprehensively — letters, household accounts, ritual instructions, recipe books.
2. Buddhist and popular printing. A religion pushed out of the state's favour had every reason to reach readers the state was not reaching, and the vernacular devotional book was the result.
3. Fiction. From the seventeenth century the hangul novel became a major genre with a real commercial market, circulating through lending libraries and copyists, read aloud in households, and despised in print by the men who ran the country.

The one piece of official infrastructure the script did acquire in those centuries came from a translator rather than a ministry. In 1527 Ch'oe Sejin published the Hunmong chahoe, a Chinese-character primer for children, and prefaced it with an account of the vernacular letters that gave most of them the names Koreans still call them by.²⁰ A script with no government office was standardised, for four hundred years, by a textbook.

The contempt was part of the arrangement. Ŏnmun, "vernacular writing", was the polite name, and it carried the sense of something low. Amk'ŭl, "women's script", was the other one — a term attested in later usage rather than in fifteenth-century documents, which is worth saying plainly rather than quoting as though a Chosŏn official had said it — set against chinsŏ, "true writing", for Chinese.¹⁰ But the contempt had a consequence its holders did not intend. A script that nobody official was watching was a script that nobody official was controlling, and when the peninsula's politics finally needed to address a population rather than a bureaucracy — in the wartime proclamations of the 1590s, in the vernacular petitions of the eighteenth century, in the posters that so enraged Yŏnsan'gun — the instrument was already there, already known, already in every provincial house.¹³

1894: the edict that arrived with an army

On 21 November 1894 a royal edict declared that the laws and decrees of Korea would henceforth be written in the national script, with translation into Chinese or the use of mixed script permitted.¹⁶¹² Four hundred and fifty-one years after Sejong made the letters, and four hundred and forty-eight after he promulgated them, the Korean state adopted its own alphabet.

The circumstances deserve to be stated without softening. The edict was one of some two hundred issued by the Deliberative Council of the Kabo reform cabinet, a body formed after Japanese troops seized the royal palace in July 1894 and installed a government willing to legislate at speed.¹⁶ The single most consequential act in the history of Korean writing was performed by a reform administration operating under foreign military pressure, and the men who drafted it were Korean reformers who had wanted this for years and finally had the leverage to get it. Both halves of that sentence are true and neither cancels the other.

What followed came fast, and came mostly from outside the state. The Tongnip sinmun, founded in 1896, printed entirely in hangul with no Chinese characters at all, and demonstrated that a serious newspaper could be produced that way.¹² Protestant missionaries, who had been publishing scripture and tracts in the vernacular since the 1880s precisely because it reached people Chinese could not, had already built much of the printing and pedagogical infrastructure the new policy required, and had begun the revaluation of the script from low writing to national possession.¹¹ Chu Sigyŏng, the grammarian who did more than anyone to standardise the modern orthography, coined the name han'gŭl — "great script", or "Korean script" — in the early 1910s.¹ The name everyone now uses is barely a century old. For most of its life the thing did not have a dignified name to be called by.

What it cost, and who paid

The bill at the point of transfer

Nothing. This has to be said first and said plainly, because the atlas's habit is to look for the cost and there are records where the honest answer is that the transfer itself had none. No one was conquered so that Korea could have an alphabet. No one was enslaved, displaced, taxed or killed in its making. The source and the target were the same civilisation; the frontier crossed was a class line inside a single society. The nearest thing to a casualty at the point of transfer is Ch'oe Malli, who lost his post for a day, resigned it permanently, and died in his village at the age of about fifty-seven — a man ruined, mildly, by being right about the consequences and wrong about whether they were bad.

The five months of 1504

The first real item on the bill is Yŏnsan'gun's proscription, and it should be counted for what it was rather than inflated. For roughly five months in 1504, being able to read was a reportable condition in the Korean capital. Books were burned. Neighbours were required to denounce neighbours. The surviving record does not permit a body count, and this record will not invent one; what it permits is the observation that a state which has ordered a literacy register has crossed a line, and that the line was crossed over a script whose entire declared purpose was to let people say what they had been unable to say. The ban's proximate cause was that somebody had used it to say something about the king. The ban ended when the king found the script convenient again.

The structural cost is larger than the five months. The abolition of the Ŏnmunch'ŏng in 1506 removed the last institution with a brief for the alphabet, and nothing replaced it for three hundred and eighty-eight years.

What 451 years in the margin cost

The largest item on this record's bill is not a violence. It is an omission, and it is quantifiable in the sense that the thing omitted is nameable.

From 1446 to 1894 the Korean state possessed an alphabet that its own promulgating document claimed an unschooled person could learn in ten days, and conducted none of its business in it. Law, tax assessment, land registration, judicial procedure, appointment, official history and the examinations were transacted in a language that the people governed by them did not speak. The gap between what the state could have made legible and what it chose to make legible is the cost, and it was paid, every year for four and a half centuries, by everyone on the wrong side of it — in petitions they could not read before affixing a mark, in obligations they could not check, in judgments they could not follow.

It is not that nobody noticed. Sejong's preface says the thing outright: there are many among the people who have something to say and cannot in the end express it. The instrument for fixing that was built, published, explained, and then left on the shelf by the class whose position depended on the problem continuing.

The colonial chapter, and what may not be billed to it

In the last decades of the Chosŏn dynasty and the first of Japanese rule, hangul stopped being low writing and became national property — and it did so largely because the Japanese colonial state made it into a thing that could be taken away.

From the late 1930s, under wartime assimilation policy, Korean was progressively removed from schools and public life in favour of Japanese.¹⁵ In October 1942 the colonial police, acting on a schoolgirl's confiscated diary in which she recorded being punished for speaking Korean, arrested the members of the Korean Language Society in Keijō on charges under the Peace Preservation Law. The Society's offence was compiling a Korean dictionary and standardising Korean spelling — it had issued the Unified Hangul Orthography in 1933, which remains the basis of South Korean spelling.¹ Its members were tortured. Yi Yunjae, born 1888, died in prison in 1943. Han Chin, born 1886, died there in 1944.

Those two deaths are real, they are named, and they are dated, and they are not a cost of Sejong's transmission. They are a cost of colonial occupation. The atlas's cost framing is only worth anything if it refuses this kind of accounting shortcut: the fact that a thing was attacked later does not transfer the attacker's bill to the thing's maker. What does belong in this record is the consequence — that four and a half centuries of official contempt were reversed in about forty years by an occupation's attempt at erasure, and that hangul acquired the meaning it now carries, as the emblem of Korean nationhood, from being suppressed rather than from being adopted.

The same period supplied the one great gift to the record's own scholarship. In 1940, in an old house in Andong, the collector Chŏn Hyŏngp'il acquired from the Yi Han'gŏl family the only known copy of the Hunmin chŏngŭm haerye — the volume that explains the design of every letter, unseen by scholars for centuries, surviving in a provincial household through the entire period in which the script it explains was supposedly beneath notice. Until that moment nobody living knew why the letters had the shapes they had, and the ʼPhags-pa and pictographic theories had been competing in an evidential vacuum. The book was designated National Treasure No. 70 in 1962 and entered the UNESCO Memory of the World Register in 1997.²³

The severance

There is a final cost and it is the one a Korean reader recognises before it is explained. Korean is now written in hangul, essentially alone. Chinese characters have been progressively removed from South Korean newspapers, textbooks and public signage since the 1970s and were never much used in the North. This is the alphabet's complete victory, and it arrived within a century of the edict that made it official.

The consequence is that the documentary record of the peninsula from roughly the fourth century to the nineteenth — the Veritable Records, the law codes, the land registers, the vast majority of the literary and historical corpus, almost all of it in literary Chinese and the rest in idu — is unreadable to nearly everyone whose inheritance it is. Translation projects have been running for decades and are enormous, and they will not close the gap, because a translated archive is a curated archive: someone decided what was worth rendering. A society that could once read only its rulers' documents can now read everything except its own past.

The ledger

Cost severity 2. The transmission killed nobody and displaced nobody. Against that: a five-month state proscription in 1504 with book-burning and compulsory denunciation of the literate; the abolition of the script's only government office in 1506 and 388 years without a successor; four and a half centuries in which a state that possessed a solution to its own stated problem declined to use it, at a cost borne entirely by the governed; and a modern settlement that secured universal literacy at the price of severing that literacy from a millennium of records. The named casualties are small in number and ambiguous in sympathy — Ch'oe Malli, ruined for opposing it; three of the men who explained it, executed in 1456 for something else entirely. The deaths of Yi Yunjae and Han Chin in colonial prisons belong to the history of the script and to the account of Japanese rule; they do not belong on Sejong's bill, and this record declines to put them there.

What the ledger cannot balance is the other side. An alphabet designed in a palace in 1443, on a correct analysis of a language nobody had analysed, by a man who wrote down why he was doing it, is now the sole script of some eighty million people and among the most quickly learned writing systems in existence. Its designer's stated ambition — that everyone should practise it at their ease and find it convenient for daily use — was met, in full, four hundred and fifty years after he was dead.