Day 47 · A History of Education

The Decisive Changes in Education Were Never About What Gets Taught

Thursday, 13 August 2026 · BigCat's Time Machine
An exam paper with the name pasted over, a safe-conduct for travelling students, a village school rule fining absence by the day, a ledger recording household by household who could read — none of the four touches teaching content, yet all four decided who could learn and whether the learning counted.
EVENT · 01

The turning point of the imperial exams was not opening them, but covering the nameAnonymous Marking in the Song Examinations · 992–1015, Kaifeng

992–1015Northern Song · KaifengPing-ti Ho 1962

The Sui and Tang already held examinations, but Tang papers carried the candidate's name, and the pre-exam custom of xingjuan flourished: candidates sent their poems to powerful patrons to be recommended to the chief examiner, who also consulted circulating reputation lists when fixing the ranking. Formally open to all; in practice a negotiation among insiders, with a typical session admitting only twenty or thirty men.

In 992 the court official Chen Jing petitioned that palace examination papers have the head sealed, hiding name and native place; under Emperor Zhenzong the practice spread to the prefectural and departmental exams. But sealing could not defeat handwriting recognition, so in 1015 a Copying Bureau was established: clerks transcribed every original in vermilion ink, and examiners saw only the copy, cutting the last identifying thread. Locked compounds and separate sittings for examiners' relatives completed the apparatus. The scale changed with it: the two Song dynasties admitted over a hundred thousand jinshi, dozens of times the Tang total. Ping-ti Ho's The Ladder of Success in Imperial China (1962) found that roughly a third of Ming and Qing degree-holders came from families with no degree in three generations.

Without sealing and recopying, the examinations would likely have stayed in their Tang form: a supplementary channel open to the well-born rather than the trunk of civil recruitment. The dispute is over how much mobility followed. Ho read genuine upward movement; Robert Hymes, in Statesmen and Gentlemen (1986), used Fuzhou local records to object that elites held position through marriage alliances and charitable estates, so "no degree in three generations" means only that nobody passed, not that the family was poor. Benjamin Elman (2000) added that the classical training a candidate needed was itself a hidden gate shutting out illiterate households. Qian Mu offers a further reminder: the system was also a device that left educated men exactly one career.

Anonymised CV screening and double-blind review: equally effective at removing the reviewer's partiality, equally powerless against the disparity in resources that existed before anything was submitted.

Procedural fairness can block a judge's bias. It cannot move the starting line.
In your own review process, where does the largest remaining bias enter once names are stripped — before the submission, or after?
EVENT · 02

The university began as a guild, and its weapon was leavingUniversitas as a Guild · 1155–1231, Bologna and Paris

1155–1231Bologna · Paris · OxfordRashdall 1895

The twelfth-century student was a foreigner: he had travelled hundreds of miles to hear a particular jurist or theologian, held no civic status where he lodged, and could be gouged by landlords and handled at will by the town court. Universitas then did not mean "universal knowledge"; it was the ordinary word for a corporate body, a trade guild — at Bologna a guild of students, at Paris a guild of masters.

In 1155 Frederick Barbarossa issued the Authentica Habita, allowing a travelling scholar under suit to be tried by his master or the bishop rather than the town. In 1200, after a tavern brawl in Paris left students dead at the hands of townsmen, Philip II granted a charter placing students under ecclesiastical jurisdiction. In 1229 the Paris masters and students left the city en masse after a violent crackdown, and in 1231 Gregory IX issued Parens Scientiarum, confirming their right to set their own statutes and confer their own degrees. The real lever was portability: in 1209 Oxford students walked out after two of their number were hanged, and one contingent settled at Cambridge. The Bologna student guild went further and hired its professors — fining them for arriving late or failing to finish the syllabus.

Without these exemptions scholars would have remained appendages of cathedral schools, knowledge production would never have acquired independent legal personality, and the degree would not have become a credential valid across borders. Hastings Rashdall (1895) wrote the university as spontaneous guild order from below; A History of the University in Europe (ed. Rüegg, 1992) revises this, arguing the space was pried open by church and crown checking each other. Both rest on the same premise: autonomy came from the fact that masters and students could leave at any time, and the town lived off them.

Open-source foundations give a project a legal shell and trademark custody so no single company can absorb it; the leverage of remote teams and scarce engineers rests on the same fact — that the cost of moving is low.

Autonomy is not a principle granted from above. It is bargaining power bought with portability.
If your team relocated as a whole, who would bear the cost? That answer is the true boundary of your autonomy.
EVENT · 03

Compulsory schooling was a state project long before it was a child's rightPrussia's Compulsory Schooling · 1763–1810

1763–1810PrussiaAndy Green 1990

Frederick William I had ordered village children into school in 1717, and the order remained a dead letter. When the Seven Years' War ended in 1763 Prussia was depopulated and financially strained, and Frederick II needed tax officials who could keep accounts, sergeants who could read orders, and peasants who would not revolt.

On 12 August 1763 he issued the General-Landschul-Reglement, drafted by the Berlin pastor Johann Julius Hecker: children aged five to thirteen had to attend, and absence was fined by the day, collected from parents by the village headman. Enforcement was dismal — teachers were largely discharged soldiers, wages were scraped together by villagers, and classes stopped in winter for lack of firewood. The General State Laws of 1794 made schools state institutions with state-examined teachers. After the catastrophe at Jena in 1806, Wilhelm von Humboldt ran the education section of the interior ministry in 1809–1810: he founded the University of Berlin and turned the school-leaving examination, the Abitur, into the single credential for university entry and public office. By the 1830s more than eighty per cent of Prussian children of school age were enrolled, far ahead of Britain or France.

Had education not been taken over by the state, the German industrial and military catch-up of the mid-nineteenth century would have been close to impossible — yet liberation was never the design intent. Andy Green's Education and State Formation (1990) makes state-building the primary driver; James C. Scott's Seeing Like a State (1998) treats it as one more move to render a population legible, countable, deployable; Peter Lindert's Growing Public (2004) shows with data that mass schooling correlates strongly with later growth. The largest surprise came at the end: the literate population the state had produced became the readership of the 1848 revolutions. Once a capability is installed in a human head, it stops taking orders from whoever installed it.

Standardised corporate onboarding and compliance training aim at capability and conformity at once; and the best-trained people are usually the first to see which procedures are absurd — and the ones able to leave.

Universal provision is usually motivated by making a system governable, but the capability it spills over reshapes the governors.
What capability is your training meant to install? When it gets used in a direction you did not intend, will you count that as failure or success?
EVENT · 04

Swedes learned to read without schools, and never learned to writeReading Without Schooling · 1686–1800, Sweden

from 1686Sweden · FinlandEgil Johansson 1977

Seventeenth-century Sweden was a poor agrarian country of scattered villages with almost no schools. Lutheran doctrine required believers to read scripture for themselves, so the church took on a project the state could not.

The Church Law of 1686 required every parishioner to be able to read the catechism and set passages of scripture. The enforcement device was the husförhör, the household examination: each year the pastor visited every farm and had each person read and recite in front of him, with results entered in the husförhörslängd, the household examination register. The sanction was not a fine but a disqualification — those who failed could not take communion, and therefore could not marry. By the mid-eighteenth century reading ability among Swedish adults of both sexes commonly reached eighty to ninety per cent, the highest in Europe. The cost lay in the other half: the church tested reading and not writing, so writing stayed very low until schools spread in the nineteenth century. Egil Johansson of Umeå University reconstructed these rates household by household from the parish registers in the 1970s and named the phenomenon "literacy without schooling".

Measured the English way — by whether one could sign a marriage register — Sweden would rank as unremarkable, because most people precisely could not sign. The choice of metric rewrote the historical ranking. The deeper argument concerns what literacy does. Harvey Graff's The Literacy Myth (1979) used nineteenth-century Canadian data to show that literate workers did not thereby rise; literacy functioned more as an employer's filter. Jack Goody once held that writing itself produces abstract thought; Brian Street's New Literacy Studies (1984) replied that a capability is always embedded in its use — reading scripture to pass an examination and reading to draw up a contract are two different literacies.

Many people can now competently direct an AI to produce output while lacking the ability to verify or correct it: the new version of reading without writing. The same holds for metrics — the measured half gets optimised while the unmeasured half quietly decays.

Reading is not writing. The measured half gets optimised; the other half gets abandoned.
Which capability on your team has never entered any assessment? How much of it is left?

Four devices: the levers of educational history sat outside the classroom

Not one concerned teaching content; all four changed who could learn and whether it counted.
Device
Trouble to be solved
What it actually did
Unintended consequence
992–1015
sealed & recopied papers
rank negotiated through patronage
hide the name, then have clerks retranscribe every paper
the gate moved back to preparation resources
1155–1231
scholars' privileges
foreign students had no standing in town
exemption from civic courts, recognition of the guild as a legal body
the degree became a cross-border credential
1763–1810
compulsory attendance
the state lacked usable clerks and sergeants
daily fines for absence plus one examined exit
a literate public became the revolutionary readership
from 1686
household examination
no schools, yet everyone had to read scripture
tested face to face at home; failure barred marriage
a nation that read well and long could not write

Going Deeper

Which is closer to the prototype of the modern education system, the Chinese examinations or the Prussian school?
It depends which half you mean. The selection-and-credential half descends from the examinations: anonymous marking, a single standard, positions allocated by test result — logic that returned to the West through the British civil service reforms of the nineteenth century. The training-and-discipline half descends from Prussia: age-graded classes, subject timetables, state-certified teachers. Today's school is a suture of the two, and their aims conflict: selection wants spread, training wants everyone at the bar. Most long-running quarrels about schooling are that seam tearing.
Why was none of the four driven by an educator?
Because the binding constraint in education is rarely pedagogy. The Song wanted a credible ranking, medieval students wanted physical safety, Prussia wanted usable soldiers and tax clerks, the Swedish church wanted believers who could read scripture themselves — under those constraints, content is a secondary variable. The corollary is uncomfortable: to change an education system, changing the curriculum is usually the least effective move. Change the credential, the admission rule, and the metric.
In what form could "literacy without schooling" recur today?
Sweden's three conditions were a very narrow target capability (read the catechism), very frequent assessment (annually, household by household), and a sanction tied to a major life event (marriage). None is hard to reproduce — certification, frequent short tests, a link to promotion — and a capability can be spread with no classroom at all. The risk is isomorphic: the ability grows into the exact shape of the assessment. That Swedes read superbly and could not write was the direct result of the design.
If AI makes acquiring knowledge nearly free, what part of education remains irreplaceable?
Exactly what these four devices were doing: selection, certification, community, and assessment. Transmitting content stopped being scarce with printing, and schools still lasted five hundred years, because a school is simultaneously a signalling machine and a social body. What AI disrupts most is not lecturing but assessment — once homework can no longer evidence ability, the credibility of the credential starts to collapse, and the credential is what the system actually produces.