Civilisation and education are connected by a remarkably practical relationship: what people learn changes what they can build, maintain and improve, and the world they build changes what the next generation can learn. An upward spiral begins when useful knowledge becomes dependable capability, its benefits widen opportunity, and enough of those benefits return to support further learning. A downward spiral develops when mistakes, exclusion, exhaustion or neglected foundations make each new round of work harder than the last.
This helps explain why a society can become richer without becoming more secure, why impressive buildings can coexist with weakening institutions, and why a quiet improvement in a classroom may matter more than it first appears. Education is not a magic explanation for every rise or fall. Resources, health, peace, institutions, power, technology and the environment matter too. But education is one of the principal ways a civilisation turns yesterday’s experience into tomorrow’s starting point.
Imagine arriving in a city before its shops open.
Someone is checking a machine. Someone is preparing a classroom. Someone is making sure that a delivery reaches the right place. A tired person is finishing a night shift, and another person is arriving to take over. The city has not remained alive merely because its buildings survived the night. It has remained usable because people know what to do next.
Now imagine a child walking through that city on the way to school.
The child did not design the water system, invent written language or negotiate the rules that allow strangers to work together. Neither did most of the adults passing on the pavement. We arrive inside an inheritance, not at the beginning of a blank world.
The wonderful part is that we can add to it.
The worrying part is that we can lose the ability to carry it.
Between those possibilities lies this essay’s central question: after each generation has lived, learned and worked, does the next generation begin with a stronger foundation—or with a larger bill for the foundations we have used up?
The spiral is a way of examining that question, not a law that assigns nations a predetermined destiny. The examples described as imagined are illustrations, not field observations. Where the essay discusses research or particular institutions, the relevant sources are linked beside the claim and collected at the end.
Choose a route through the essay
Understand the spiral · Chapters 1–5. Inheritance, retained learning and the meaning of progress.
Follow learning into society · Chapters 6–10. Classroom foundations, work, Singapore and reliable standards.
Examine tools and prosperity · Chapters 11–15. AI, human capability, shared abundance and growing complexity.
Recognise and repair decline · Chapters 16–20. Wrong lessons, missing successors, interrupted access and recovery.
Decide what counts as better · Chapters 21–25. Purpose, inclusion, exchange, truth and seven practical questions.
Turn understanding into action · Chapters 26–30. Two towns, families, institutions and the next generation.
Explore all 30 chapters
- The City Has to Learn How to Be a City Again
- Why a Circle Does Not Necessarily Become a Spiral
- The Two Returns: Feeding the Foundations or Feeding on Them
- Wealth, Stability and Direction Are Different Questions
- Education Lets a New Mind Begin Somewhere Other Than Zero
- A Small Gap Can Become a Large Detour
- Attendance, Certificates and Capability Are Not the Same Thing
- Learning Needs Somewhere Useful to Go
- Singapore Makes the Connections Easier to See
- A Gold Standard Is Something Another Person Can Reliably Do
- The Words We Can Use Change the Tools We Can Use
- AI Can Raise Performance Without Raising the Same Kind of Capability
- We Do Not Need to Know Everything the Machine Knows
- When Private Wealth Becomes a Public Starting Point
- Success Can Make the System Harder to Carry
- The Downward Spiral Often Begins With a Misread Signal
- The Future Can Develop a Shortage Before the Present Notices
- Sometimes the Door Is Closed From the Outside
- Recovery Begins by Finding the Actual Starting Point
- Learning Has a Physical and Human Floor
- Better at What—and Better for Whom?
- There Is More Than One Route Into a Capable Society
- Civilisations Learn Across Borders, Not Only Within Them
- Truth Needs Somewhere Safe to Arrive
- Seven Questions That Make the Direction Easier to See
- Two Towns, Two Very Different Returns
- What a Family Can Do Without Turning Home Into a Factory
- What Institutions Can Do Before the Crisis
- The Questions That Keep This Essay Honest
- The Next Generation Should Inherit More Than Our Buildings
1. The City Has to Learn How to Be a City Again
A city can look reassuringly permanent. Its station is where it was yesterday. Its familiar bridge crosses the same river. The school gate opens beneath a sign that several generations recognise. We find comfort in these continuities because they suggest that the world has not forgotten how to hold together.
Yet permanence in appearance can conceal extraordinary turnover underneath.
The person repairing the bridge today is not the person who built it. The teacher explaining fractions this morning was once a child encountering fractions for the first time. The manager responsible for the station had to learn procedures that existed before their appointment. The people who keep the place running are continually changing, even when the skyline seems still.
For the purposes of this essay, think of civilisation as a shared arrangement for carrying human life and capability through time. It includes much more than cities, governments and machines. It includes language, caregiving, practical knowledge, institutions, art, agreements, memory and the possibility of asking a useful question. Its achievement is not simply that something impressive exists. Its achievement is that people can continue using, understanding, questioning and renewing what exists.
This gives education a larger meaning than attendance in a classroom. School is a central part of it, but so are apprenticeship, conversation, professional training, libraries, supervised practice, community knowledge and the deliberate transmission of judgement. Someone learns how to recognise a problem. Someone learns whom to consult. Someone learns why a rule exists, and when its assumptions no longer hold.
Consider an imagined neighbourhood workshop. Its most experienced technician can identify a recurring fault quickly. When she hears a particular irregularity, she knows which inspection should come next. A younger colleague hears only noise. The difference is not a difference in human worth. It is a difference in experience, interpretation and access to learning.
If the expert fixes every fault alone, the workshop may have an excellent month. If she also teaches a colleague to investigate safely, record the evidence and recognise the limits of the diagnosis, the workshop may have a better future. The immediate repair and the future capability are different outputs. A sensible organisation needs both.
Now enlarge the scene. Substitute a clinic, a school, a transport operator, an archive or a water utility. The details change, but the handover question remains: who is becoming able to do the work that currently depends on someone else?
A civilisation therefore has to reproduce more than its population or its buildings. It has to reproduce access to useful competence. That does not mean every person must master every task. It means the shared system must keep producing enough people, connections and trustworthy records for important work to remain possible.
There is a hopeful implication. Every act of good teaching is an interruption of the need to start from nothing. A learner can inherit a method without repeating every mistake through which that method was discovered. A young musician can begin with an instrument, a tradition and a teacher. A new technician can begin with drawings, demonstrations and supervised practice. Their learning is still real work, but it begins from an offered platform.
The platform must not become a cage. Education also needs to give people the means to recognise when an inherited practice is inadequate, unjust or obsolete. A society that passes on instructions without permission to question them preserves obedience more reliably than understanding.
This is the delicate beauty of the task. We need enough continuity that knowledge does not disappear with its carriers, and enough freedom that its next carriers can improve it. We need a handover that says both “Here is what we have learned” and “Here is how you can discover where we were wrong.”
The earlier reflection on civilisation’s familiar landmarks begins with the comfort of something enduring. Education reveals the human activity inside that comfort. The building may look the same. The competence that protects its future has to be renewed.
That renewal rarely arrives with a ribbon-cutting ceremony. Sometimes it looks like an experienced person making room beside them and saying, “Watch carefully. Then you try.”
Reading routes · Contents · Next chapter
2. Why a Circle Does Not Necessarily Become a Spiral
Picture two people practising the same task every week.
The first repeats it, makes a familiar mistake, receives a correction and forgets the correction before the next attempt. The second also makes mistakes, but begins to recognise their causes, changes an approach and can use the improvement later. Both have completed another week. Both have been busy. Only one has clearly changed the starting conditions of the next week.
This is the difference between a circle and an upward spiral.
A circle returns. A spiral returns with a change in level. The metaphor becomes useful when we ask what survived the previous round. Did an explanation become available when it was needed? Did a repair change the process that produced the fault? Did a successful experiment become something other people could reproduce? Did an institution retain the reason for a decision, rather than merely retaining the decision?
The essay on retained learning makes this distinction especially clear: a correction is not yet cumulative improvement. Useful change must remain retrievable and capable of influencing a later encounter. A report saved somewhere is not the same as a lesson that changes what people do.
An imagined bakery gives us a simple example. On Monday, the afternoon batch is disappointing. The owner compensates by working late. On Tuesday, the same thing happens. By Friday, everybody has become very good at rescuing the afternoon. The bakery has developed a response, but it has not necessarily learned why the problem recurs.
Another bakery records the conditions, checks plausible causes, tests one change and compares the result. Suppose it finds a scheduling problem rather than a failure of effort. It alters the sequence and teaches the revised procedure. If the improvement survives different staff and an unusually busy day, the next week begins with less avoidable trouble.
The first bakery may have shown admirable commitment. The second has converted experience into an improved operating arrangement. The distinction is not between hardworking and lazy people. It is between expenditure that restores the moment and learning that changes subsequent moments.
There is also a downward spiral. A mistaken explanation can survive just as effectively as a correct one. Suppose the owner concludes that staff are unreliable and responds by removing their discretion. Employees then stop reporting small concerns because every report becomes an accusation. The owner sees fewer warnings, interprets the silence as improvement and tightens control further when the next failure arrives.
Now each round changes the conditions of the next round, but in an unhelpful direction. Information becomes worse. Trust becomes thinner. Correction becomes more difficult. The system has learned something, but it has learned how to protect a bad interpretation.
This is why feedback is not automatically wisdom. A result tells us that something happened under particular conditions. It does not arrive carrying a complete, infallible explanation. We still have to distinguish cause from coincidence, persistent patterns from unusual episodes, and changes that help one part of the system from changes that harm another.
Nor is an upward spiral simply a demand for permanent acceleration. Sometimes the intelligent next step is to slow down, check the work or consolidate a skill. A learner who takes time to understand equivalent fractions may appear to be covering fewer chapters. Yet that learner may be improving the foundation from which later chapters become accessible.
There can even be progress in learning what not to continue. Retiring an unnecessary procedure, abandoning an unsupported claim or stopping a damaging practice can increase future capability without adding another visible object to the world.
Three questions make the metaphor more precise. What is returning? What is being retained? And why should the retained change count as better?
Without the first, we have only isolated events. Without the second, we have activity without accumulation. Without the third, we may become increasingly effective at producing an outcome we should not have wanted.
The spiral earns its upward direction. It does not receive that direction merely because time has passed, expenditure has increased or the newest version has a more impressive name.
Previous chapter · Contents · Next chapter
3. The Two Returns: Feeding the Foundations or Feeding on Them
Imagine a small town with a useful surplus after a good year.
Its businesses have done well. Essential services have remained dependable. People have a little more room to think beyond the immediate emergency. The town now faces a choice that is less dramatic than a crisis but may be just as consequential: what happens to the benefits of success?
Some of the surplus might improve homes and ordinary enjoyment. Some might maintain shared facilities. Some might support training, better tools, a library, a reserve for difficult years or a more accessible route into learning. People can enjoy today while making tomorrow easier to inhabit.
But another arrangement is possible. The surplus might be extracted while maintenance is postponed. Training might be reduced because experienced staff are still managing. People doing the most demanding work might absorb the growing strain privately. The year’s accounts could look pleasing because the costs have moved somewhere the accounts do not adequately reveal.
These are two different kinds of return.
In the regenerative version, the outcome of the current round replenishes the conditions that made a good outcome possible. Capability produces benefits; some of those benefits support learning, care and renewal; the next round begins with a stronger base.
In the self-consuming version, current performance is protected by weakening that base. A service appears inexpensive because its staff donate unsustainable amounts of effort. A classroom appears successful because adults complete too much of the learner’s thinking. A company appears efficient because it has stopped training the people it will later need.
These examples are illustrations, not a claim that every low-cost service or strong examination result conceals damage. The point is to ask where the cost of the result has gone.
An educational version is easy to imagine. A student struggles with a piece of work. An adult supplies the answer so the deadline can be met. Occasionally, this may be a reasonable response to exceptional circumstances. But if it becomes the standing arrangement, the next assignment may arrive before the child has acquired the missing capability. The adult then supplies more. Completion remains visible. Dependence grows underneath it.
A more regenerative response might still include immediate support. The important difference is that support also repairs the underlying difficulty and gradually gives the learner more control. The next assignment is not simply completed; it becomes less dependent on rescue.
The same logic applies to organisational emergencies. It would be absurd to tell a team not to restore an essential service because “rescue is not learning.” Of course the service must be restored. The question is whether the organisation then protects time to investigate, train and prevent recurrence, or treats permanent emergency response as its normal operating model.
There is nothing morally superior about never needing help. We are interdependent creatures. The desirable change is not the disappearance of support but the development of support that enlarges capability rather than quietly replacing it.
This also means that consumption is not automatically self-consumption. A family’s holiday, a beautiful meal, a concert or a comfortable home need not be defended as an investment in future productivity. Enjoyment is part of what prosperity is for. A society that permits no present pleasure because everything must serve a future objective has lost sight of the people that future is meant to benefit.
The distinction concerns the overall relationship between use and renewal. Can people enjoy what they have without systematically exhausting the human, institutional and material conditions on which their enjoyment depends? Are the less visible contributors able to recover, learn and continue? Are future obligations acknowledged rather than passed forward as surprises?
The discussion in Civilisation | Abundance provides a natural companion. Abundance becomes more durable when it means more than a large quantity of attractive things. It includes the ability to sustain access, repair what fails and keep widening the circle of people who can participate.
The two routes can be stated simply. In an upward loop, learning supports useful capability; useful capability produces benefits; some benefits return to learning, care and renewal. In a downward loop, an unresolved problem invites short-term compensation; compensation consumes the time or resources needed for repair; the next problem arrives under worse conditions. These arrows describe possible relationships to investigate, not fixed rates of growth or decline.
An upward spiral is therefore not merely success followed by more success. It is success that helps reproduce the conditions for worthwhile success. That small addition changes the entire story.
Previous chapter · Contents · Next chapter
4. Wealth, Stability and Direction Are Different Questions
We are tempted to draw civilisation as a single staircase. At the bottom are poor places. At the top are rich places. Every invention becomes another step, and every difficulty becomes a fall.
It is an appealing picture because it is easy to read. It is also too simple for the work we need it to do.
A wealthy society can be stable in one system and struggling in another. A community with modest resources can improve its teaching, trust and practical resilience while remaining materially constrained. A technologically sophisticated organisation can become less honest about its failures. A recovering institution can still have poor current results while making changes that improve its future.
These statements are possibilities within the model, not a ranking of particular countries. Their purpose is to keep three questions separate: what does the system currently possess, how well does it currently function, and what direction are its underlying conditions moving?
Possession, condition and direction are not synonyms.
Absolute improvement and relative position are different as well. A community can become better able to provide a service while falling behind another community that is improving faster. Conversely, moving up a ranking need not mean that conditions have improved if other places have deteriorated more. A rank describes a comparison, not the whole direction of a life.
This is why a difficult present and an improving trajectory can coexist. It is also why a comfortable present should not end the inquiry. We need to know what is happening to the foundations, not merely where the place stands in a photograph of everyone else.
An imagined school makes the distinction concrete. School A has a handsome campus, extensive equipment and strong inherited results. Its teachers, however, have little time to plan together, and the expertise of departing staff is not being replaced. School B has fewer facilities and weaker current results, but has recently established reliable attendance, clearer diagnosis and protected teacher collaboration.
We cannot responsibly conclude that School B is already better in every respect. Families still experience the actual quality of teaching available now. But neither should we conclude that School A’s impressive stock of resources guarantees the healthier trajectory. To judge direction, we need evidence of what is changing underneath current performance.
Stability has a similar ambiguity. A system can be stable because its foundations are strong and its errors are corrected early. It can also appear stable because complaints are suppressed or an exhausted group absorbs every disturbance. The surface movement may look alike; the underlying mechanisms differ.
Conversely, visible disagreement is not always evidence of decline. A newly open discussion of previously hidden problems can make an institution look less comfortable while improving its ability to respond. What matters is whether the disagreement becomes useful diagnosis and fair action, not whether everyone appears cheerful in a photograph.
Technology adds another complication. A powerful tool can spread into settings with very different institutional conditions. Its presence does not tell us that the surrounding system has acquired the skills, safeguards, legitimacy or resources to use it well. Nor does a difficult surrounding environment imply that nobody there can use the tool intelligently.
The appropriate unit of observation may also be smaller than a country. One hospital, workshop, school network or community can improve while another deteriorates. The national average may conceal both movements. A person’s life is shaped by the institutions they actually encounter, not only by the label attached to the country around them.
This is why “some civilisations rise and some fall” should begin an investigation rather than end one. What rose? Income, safety, knowledge, freedom, reliability, health, access or the concentration of power? What fell? For whom, over what period, and relative to which starting conditions?
The earlier essay on advancing or regressing asks how to judge such movement. Education adds a further question: are the people and institutions needed for the next stage becoming more capable, or is present performance being sustained by an inheritance that is not being renewed?
We do not need a single civilisational score to ask this well. In fact, a single score can conceal precisely the relationships that need attention. A useful assessment can acknowledge that a place is prosperous but strained, less affluent but improving, secure in one domain and vulnerable in another.
That is not indecision. It is a more accurate description of a complicated world. It also leaves room for action. A society does not need to wait until everything is failing before repairing one important system, and it does not need to claim perfection before celebrating a genuine improvement.
Previous chapter · Contents · Next chapter
5. Education Lets a New Mind Begin Somewhere Other Than Zero
There is something quietly astonishing about a child learning to read a paragraph written by a person they will never meet.
Marks on a surface become someone else’s thought. The reader can agree, object, ask for evidence or carry the idea into a different situation. Distance has been crossed. Time has been crossed. The author’s experience is no longer limited to the author’s lifetime.
Education helps make that encounter usable. The letters themselves are only the beginning. A reader also needs vocabulary, background knowledge, attention to context and some ability to recognise the difference between a claim and its support. Access to a text and access to its meaning are related but distinct achievements.
The National Academies’ How People Learn II discusses how prior knowledge supports interpretation and learning, while also noting that existing knowledge can mislead when it is inaccurate or applied inappropriately. This is a useful corrective to two extremes: learners are neither empty containers waiting to be filled nor fully independent discoverers who can dispense with inherited knowledge. What they already understand matters, including when it needs revision. [2]
Consider an imagined child hearing the word “pressure” in three places. At home, someone is under pressure to finish a task. In a Science lesson, pressure concerns a physical relationship. In a conversation about friendship, pressure can mean an attempt to influence a decision. Recognising the familiar word does not guarantee that the child has selected the right meaning.
A good explanation supplies more than a definition. It helps the learner separate cases, notice relationships and decide which meaning belongs in the present situation. Later, when the child encounters a new sentence, the earlier teaching can become a tool rather than an answer stored for one exercise.
The civilisational equivalent is an inheritance that includes context. A technical rule should travel with enough information about its purpose and conditions to remain interpretable. A historical account should not be detached from the evidence that supports it. A professional procedure should not become an incantation that nobody is allowed to examine.
Of course, we cannot attach the entire history of human thought to every instruction. Useful knowledge has to be compressed. A road sign is valuable precisely because it does not require an essay at the roadside. A well-designed checklist can prevent omissions without retelling every incident that motivated its creation.
The challenge is to compress without severing the route back to understanding. The immediate user needs a clear action. The trainer, investigator or future reviser may need the reason. A mature knowledge system makes these different depths available to the people who need them.
This is one reason an archive alone cannot guarantee continuity. Imagine a laboratory whose records are carefully preserved but whose remaining staff cannot interpret an important set of measurements. The information exists, yet practical access to it has weakened. Conversely, a team may have excellent tacit knowledge but no reliable way to transmit it when its members leave.
Education connects these forms of memory. Reading, demonstration, supervised attempts, discussion and later independent use allow stored information to become living capability. Documentation can support teaching; teaching can reveal where documentation is unclear; experience can show what both have missed.
There is no need to romanticise a past in which knowledge supposedly passed perfectly from master to apprentice. Any method of transmission can exclude people, preserve errors or depend too heavily on one person. The aim is not to choose between old and new forms of learning on the basis of their age. It is to ask whether the arrangement preserves useful knowledge, opens access and permits correction.
Nor should every generation be asked merely to imitate the previous one. The child learning to read has acquired the means to encounter inherited ideas, but also the means to write something different. Education is most alive when receiving and contributing belong to the same process.
An upward spiral therefore has two movements that must coexist. It reaches backward for knowledge worth keeping, and forward for problems that knowledge has not yet solved. Remove the first and each generation wastes effort rediscovering what could have been taught. Remove the second and inheritance hardens into repetition.
A new mind does not have to start at zero. It also does not have to finish where its teachers stopped.
Previous chapter · Contents · Next chapter
6. A Small Gap Can Become a Large Detour
Let us bring the discussion down to a school desk.
Imagine a student looking at the equation 3x + 6 = 21. The student has seen similar questions before and remembers that something should “move to the other side.” On a good day, that remembered phrase produces the expected answer. On a difficult day, the student changes a sign incorrectly and cannot explain why the next line should follow.
It is tempting to call this a lack of effort or a careless mistake. But neither description tells us enough to teach the next lesson.
Perhaps the student does not yet understand an equation as a statement of equality. Perhaps negative numbers are unstable. Perhaps the procedure was learned as a sequence of visual moves rather than as operations that preserve a relationship. Different gaps can produce the same wrong answer.
A useful repair might begin by subtracting six from both sides, leaving 3x = 15, and then dividing both sides by three, giving x = 5. Substituting five into the original expression confirms that 3 × 5 + 6 equals 21. The point is not that this is a revolutionary technique. The point is that the learner can see why the steps are valid.
Now change the surface. Ask a word problem that requires forming the equation. Later, introduce a negative value or an unfamiliar arrangement. The teacher is not simply asking whether the student remembers this answer. The teacher is investigating whether the relationship remains available when the question no longer looks exactly the same.
This diagnosis-first approach is developed in eduKate’s explanation of the transition into Secondary 1 Mathematics. Its wider significance is that an apparent difficulty in the present may be a missing connection from an earlier stage. Repairing that connection is not a humiliating retreat. It is an intelligent route forward.
In our imagined example, understanding equation balance can make later learning easier to approach. The student does not have to invent a separate rule for every new arrangement. A useful principle carries some of the load. Each successful use also offers another opportunity to check and strengthen the principle.
This is one way educational gains can accumulate. A learned relationship becomes a tool for learning another relationship. A better vocabulary makes an explanation easier to understand. Clearer reading helps a learner identify what a question actually asks. An improved checking habit catches an error before it travels through several lines of work.
But this does not mean every learner follows a smooth exponential curve. Knowledge is uneven. A student may reason beautifully in geometry and struggle with fractions, or explain a Science idea well while misreading a graph. Some improvements transfer widely; others remain narrow. Progress can pause while a new representation becomes familiar.
The sensible image is therefore not a perfectly rising line. It is a network becoming more connected and usable. Some connections need strengthening. Some need correcting. Some earlier shortcuts need replacing because they no longer fit the complexity of the task.
The downward version is also a network story. An unresolved gap makes current work harder. Harder work may invite avoidance or excessive dependence on examples. Less productive practice leaves fewer opportunities to repair the gap. The next topic then arrives on top of an unstable connection. This is a possible mechanism, not a diagnosis of every child who dislikes a subject.
A humane response does not make the child carry a civilisational metaphor as another burden. The learner does not need to be told that the future of society depends on tonight’s worksheet. The learner needs someone to identify the problem accurately and offer a manageable next step.
Often, that next step is pleasingly small. One distinction becomes clear. One misleading phrase is replaced. One old skill is reconnected to its new use. The work has not become effortless, but it has become less mysterious.
At the scale of a person, an upward spiral may begin with that change: “I am bad at this” gives way to “I can now see the part I did not understand.” It is a shift from an identity sentence to a repairable problem.
Previous chapter · Contents · Next chapter
7. Attendance, Certificates and Capability Are Not the Same Thing
A classroom can be full while learning remains incomplete. A certificate can be genuine while the holder still needs practice. An examination can be useful while failing to measure everything that matters.
These distinctions should not become an attack on schools or qualifications. They are reasons to take education seriously enough to ask what its activities achieve.
The World Bank’s World Development Report 2018 drew attention to the difference between schooling and learning. Expanding access is essential, but attendance alone does not guarantee that students acquire foundational skills. The report also treats the problem as systemic: teaching, assessment, preparation, resources and political incentives can all affect whether educational participation becomes learning. [1]
The key word is “becomes.” A school place offers a route. The existence of the route does not prove that every learner can travel it successfully. But the answer is not to remove the route. It is to improve the conditions that allow students to use it.
Imagine a school that wants to improve reading. It can count books purchased, library visits, lesson hours and completed worksheets. All of these may be useful pieces of information. Yet none, alone, tells us whether a child can understand a new passage, identify its main claim or explain why a conclusion follows.
Equally, one reading test cannot tell us everything about that child. Performance may vary with the language of the passage, its assumed background knowledge, the instructions or the conditions of the test. The sensible response is not to reject measurement, but to use several appropriate observations and investigate what they mean.
A qualification serves a related but different purpose. It helps people communicate achievement, meet entry requirements and move between institutions. In a large society, strangers cannot personally observe every skill of every applicant. Some forms of certification are useful ways of reducing that uncertainty.
The danger appears when the signal becomes detached from the capability it is supposed to indicate. An institution might then become better at producing successful-looking records than at developing understanding. Learners may respond rationally to the incentives they face, concentrating on what is rewarded. Blaming them alone misses the design problem.
Suppose a fictional course assesses only whether a student can reproduce a familiar explanation. Reproduction will become an important study strategy. If the course also asks students to apply the idea, explain a limitation and recognise an error in an unfamiliar example, the incentive changes. The institution has made a broader form of learning visible.
This is not an argument that every lesson should become an elaborate research project. Practice, repetition and accurate recall have legitimate roles. The question is what they are preparing the learner to do. Remembering a useful fact can free attention for a more demanding task. Repeating an unexplained routine can also leave the learner stranded when the task changes.
We should therefore distinguish support for performance from evidence of learning. During instruction, prompts, examples and collaboration can be appropriate. During a later check, the teacher may reduce those supports to discover what the learner can now manage. Both stages belong in education. Confusing them produces an unreliable picture.
The same distinction extends beyond school. An organisation can record that every employee completed a training module without knowing whether employees can recognise the situation the module addresses. A public campaign can reach a large audience without producing the intended understanding. A professional association can award attendance credits without demonstrating changed practice.
None of this makes formal learning meaningless. It makes the follow-through important. What did participants understand? What can they now do? Where are they still uncertain? What further support is needed? What changes when the setting is different?
An upward educational spiral does not abolish certificates or examinations. It keeps them connected to a richer account of capability. It uses measures as instruments for seeing, not as replacements for the world they are meant to reveal.
The result is a more honest kind of encouragement. We can celebrate an achievement without pretending it settles every question. We can identify a gap without treating the learner as a failure. A useful educational record is not a verdict on a person’s worth. It is information that helps the next stage begin well.
Previous chapter · Contents · Next chapter
8. Learning Needs Somewhere Useful to Go
Now imagine a young person who has worked hard, acquired a valuable skill and is ready to contribute.
What happens next?
The answer depends on more than education. There must be an opportunity to use the skill, tools that make the work possible, institutions that recognise competence and conditions under which effort can lead to a fair result. A society cannot responsibly train people and then treat the absence of those connections as a personal defect in the graduate.
This is where an educational spiral becomes an institutional one.
A trained technician needs access to equipment, suitable work and a way to keep learning. A teacher needs a functioning school, time to prepare and conditions in which teaching can occur. A researcher needs more than a qualification: there must be some route from a good question to investigation, criticism and application. The relevant complements differ by field, but the principle is the same.
Esther Duflo’s research on Indonesia’s large school-construction programme used differences across regions and birth cohorts to study its effects. The 2001 paper found increases in schooling and wages associated with greater exposure to the programme. It offers evidence that educational expansion can improve important outcomes; it does not establish that schooling alone explains an entire country’s development. [3]
A later analysis of the programme’s medium-run labour-market effects examined a more complicated adjustment. Its findings suggested that physical capital had not adjusted to the faster growth of educated labour supply, and the effects were not uniformly favourable for all workers. The useful lesson is not that education was a mistake, but that economies do not transform every increase in skill into an immediate, evenly shared gain. [4]
Consider an imagined district that trains many people in a new technical field. Its training quality is good. Yet local employers have not invested in the necessary equipment, small firms cannot access suitable finance, and recruitment still rewards personal connections over demonstrated ability. The district has improved one part of the chain while leaving other parts obstructed.
If graduates struggle, the diagnosis should include those obstructions. “Study harder” cannot repair a missing market, discriminatory selection or a procurement system that excludes competent new suppliers. Education can help people navigate and challenge such conditions, but it should not be used to excuse them.
The opposite mistake is to assume that opportunity makes education unnecessary. An employer may have a vacancy and equipment but still need people who can learn the work safely and reliably. A rapidly expanding industry can encounter a shortage of suitable training, mentors or experienced supervisors. Demand and capability need a way to meet.
An upward spiral becomes more plausible when these connections improve together. Learning develops usable skill. Institutions create fair routes into work and contribution. The resulting value supports decent livelihoods and further investment. Employers and educators exchange evidence about changing needs without reducing education to a list of today’s vacancies.
That last qualification matters. Education has to prepare people for a life, not merely for a position that happens to exist now. A narrow course can be useful, but a whole education also needs language, reasoning, adaptability, civic understanding and the ability to learn again when circumstances change.
There is a difference between coordination and captivity. Coordination helps a learner understand possible routes. Captivity trains a person for one arrangement while denying them the means to leave, question or adapt. A healthy relationship between education and work should expand choices, not quietly remove them.
We should also recognise contributions that do not appear neatly as higher wages. Care, civic participation, cultural work, volunteering and better judgement in everyday life can matter even when their value is not captured by a pay packet. Earnings are an important outcome, especially for people under financial pressure, but they are not the complete purpose of learning.
The practical question is therefore broader than “Does education make a country rich?” It is: what must be connected so that learning can become a better life, useful work, fair participation and renewed public capability?
The spiral does not live inside the school alone. It passes through the doorway into the rest of society. That doorway needs to open onto something real.
Previous chapter · Contents · Next chapter
9. Singapore Makes the Connections Easier to See
A useful way to think about Singapore is not as a magical exception, but as a setting in which many forms of interdependence are close enough to notice.
A port is connected to engineering, logistics, finance, communications and training. A housing estate is connected to water, transport, maintenance and community services. A school is connected to families, further education and the organisations in which its learners will eventually participate. No single institution explains the whole arrangement.
The newer Making Singapore Rich article on port cranes and container handling is valuable because it takes an impressive economic activity apart. It invites us to look beyond the ship and the skyline towards scheduling, equipment, controls, maintenance and the people who understand how the pieces work together.
MPA’s account of Tuas Port describes a design involving electrified automated yard cranes, automated guided vehicles and remote management from a control centre. Its discussion of automation includes collaboration with technology partners and the development of more advanced operating arrangements. The important point here is the combination of physical equipment, software and human organisation—not a claim that any one machine can independently run a port. [9]
From that description, an educational question follows. What kinds of understanding must be available when an activity becomes more automated? The answer is not simply “more computer skills.” It includes the ability to understand equipment, interpret data, coordinate work, recognise abnormal conditions and communicate across specialisms. The precise training required must be established by the organisations responsible for the work.
Consider a purely imagined container terminal. It has a fast crane, but vehicles arrive at the wrong times. Buying a faster crane may not solve its main problem. Another terminal has good scheduling but repeated equipment failures. A third has reliable equipment but poor handovers between teams. Similar headline output can conceal different constraints.
The educational lesson is that component knowledge and system knowledge have to meet. A person may understand a machine in detail while needing colleagues who understand the wider sequence. Another person may coordinate the sequence while relying on specialists for safe technical judgement. Good collaboration does not require everybody to become everybody else.
Singapore’s Institute of Technical Education provides another concrete connection. Established in 1992, ITE describes a role that includes pre-employment education, continuing education and training, and workplace-related learning. Its institutional remit makes visible an important idea: learning is not confined to one academic route or completed once a young person leaves school. [8]
The broader inference is that a society needs several routes for developing competence. Some knowledge is learned in academic study, some through practical training, and much through the combination of structured instruction and experience. Treating one route as inherently dignified and another as a consolation prize would be a poor way to sustain a complicated society.
None of this proves that every Singaporean system works equally well or that education explains every economic achievement. It would be equally misleading to overlook pressures experienced by learners, teachers, workers and families. A strong national story should not make it difficult to discuss the work needed to keep improving it.
The useful question is how the relationships are maintained. Can training respond to real changes without chasing every passing trend? Can experienced workers update their skills without being treated as obsolete people? Can new entrants obtain supervised opportunities rather than being expected to arrive already experienced? Can practical knowledge move between institutions rather than remaining locked inside them?
These are questions about a continuing process, not a trophy already won.
They also change the way a student might look at the world. A crane is no longer only a large object. It becomes a meeting place for Mathematics, Science, language, design, human judgement and organisation. A journey through a working city becomes a chance to notice what knowledge is doing.
That kind of attention can make education more meaningful without promising every child a particular career. The child is learning that school subjects are ways of understanding relationships that exist beyond the classroom.
The city’s capability becomes more intelligible. And once a system becomes more intelligible, a learner can begin to imagine participating in its future rather than merely consuming its present.
Previous chapter · Contents · Next chapter
10. A Gold Standard Is Something Another Person Can Reliably Do
There are two very different ways to use the phrase “gold standard.”
One is as a decorative claim: a way of saying that something is excellent without showing how its excellence is produced. The other is operational: a demanding account of what good performance involves, how it is checked and how it can remain dependable when circumstances change.
Only the second meaning helps a civilisation reproduce capability.
Imagine a brilliant employee who can resolve almost every difficult case. People praise the organisation because this person keeps saving the day. But what happens during an absence? What happens when demand grows? What happens when a new employee tries to learn the work?
Individual excellence is valuable. It becomes a shared asset when enough of its reasoning, practices and limits can be taught, examined and supported by a functioning system. This does not mean reducing every judgement to a rigid script. It means making competence less dependent on luck, private memory or the permanent presence of one exceptional person.
The Gold Standard Of Maintenance Management makes this distinction practical. It directs attention away from the volume of breakdown activity and towards deliberate priorities, planned work, verification, failure history and improvement. The question is not simply whether people are busy fixing things. It is whether important assets remain useful and whether recurring trouble becomes less normal.
As an educational principle, this is powerful. A learner should not only be shown what a good answer looks like. The learner needs access to the decisions that make it good. Which information matters? What would count as a mistake? How do we check? Under what circumstances would this method not be suitable?
A standard becomes teachable when these features are visible. It becomes trustworthy when performance is tested rather than assumed. It becomes sustainable when people can learn it without relying indefinitely on extraordinary individual effort.
There is a danger at the other extreme. An organisation can produce so many procedures that nobody can find the one that matters. It can confuse compliance with a document and competence in the world. It can punish intelligent questions because those questions reveal that the document is incomplete.
That is not a higher standard. It is a failure to keep the standard connected to its purpose.
Consider an imagined training exercise. A learner follows every step of a checklist but notices that the situation differs from the example used in class. A poor learning culture rewards finishing the list without interruption. A stronger one teaches the learner how to recognise the boundary of their competence, pause safely and seek the right assistance.
In this case, asking for help is evidence of judgement. The learner has not failed to become independent; the learner has learned when independent action would be inappropriate. Civilisational capability includes knowing how to coordinate uncertainty, not merely how to appear confident.
Standards also need memory. When a procedure changes, future users benefit from knowing what problem the change addressed. Otherwise, a later team may remove an inconvenient step without understanding the consequence it was designed to prevent. Equally, retaining every old step forever can burden the system with protections against conditions that no longer exist.
The appropriate combination is continuity with revision. Preserve what remains useful, make reasons accessible and test proposed changes against the actual work. Neither blind obedience to the past nor a permanent appetite for novelty is sufficient.
This gives education a role in quality that extends beyond the initial course. Teaching must help people perform, interpret, question and update. Organisations must then create a place where these abilities are useful rather than inconvenient.
A gold standard, understood this way, is not a badge fixed to the institution’s door. It is a promise supported by an ongoing practice: another suitable person can learn this; the result can be checked; the system can recover when something goes wrong; and evidence can improve the method.
That is how an exceptional performance begins to become a civilisational capability. Excellence acquires a route into the future.
Previous chapter · Contents · Next chapter
11. The Words We Can Use Change the Tools We Can Use
Imagine two learners opening the same powerful digital tool.
Both have access. Both can type. Both can receive a fluent response. Yet access alone does not tell us whether they can identify the problem, recognise a useful answer or notice when the answer is wrong.
One learner may ask for a finished result without understanding what would make it valid. Another may be able to describe the situation, separate known facts from assumptions, specify a constraint and compare the output with independent evidence. Their encounters with the same tool can therefore become different learning experiences.
This is not a claim that the second learner is innately more capable. The relevant abilities can themselves be taught. A well-designed tool can also help a beginner acquire them. The important distinction is between possessing an interface and being able to use the relationship behind it productively.
Language matters because many tools respond to descriptions of what we are trying to do. But the solution is not simply to memorise impressive phrases or write longer requests. Clear language begins with clearer distinctions. What is the objective? Which information is reliable? What is missing? What would count as success? Which part should be checked by a person with relevant expertise?
The newest eduKate article on reading research abstracts without overclaiming offers a particularly useful example. A reader who distinguishes an observed association from a demonstrated cause can protect the meaning of a study when summarising it. A reader who removes that distinction may produce a more confident sentence and a less accurate understanding.
Now imagine using an AI system to discuss a fictional school survey. The survey finds that students who report a certain study habit also report greater confidence. An uncritical summary might announce that the habit causes better examination results. But the survey has not necessarily measured examination results at all, and a reported relationship does not by itself settle causation.
The educational skill is not merely finding a fault in the machine. It is knowing what the original evidence supports. That knowledge helps the learner formulate a better question, evaluate the response and decide what further information is required.
This creates a possible upward loop. Better understanding produces a better question. A useful response expands understanding. The learner tests it, retains what survives the test and becomes able to ask a more precise next question. The tool supports the development of its user rather than merely completing tasks on the user’s behalf.
A downward loop is also possible. A plausible response is accepted without checking. The learner builds another question on its unsupported claim. More fluent material accumulates around the mistake. Because the output looks organised, confidence grows faster than understanding. Later work begins from a weakened foundation disguised as a polished one.
Neither outcome is guaranteed by the presence of AI. Tool design, teaching, subject knowledge, incentives and the consequences of error all matter. A task completed for amusement may tolerate uncertainty that would be unacceptable in an important technical decision. The right amount of checking depends partly on what could happen if the answer is wrong.
We should also resist confusing language proficiency with intelligence. A person may understand a problem deeply while struggling to express it in the language preferred by a particular interface. Translation, accessible design, diagrams and other forms of representation can help. A fair educational system expands these routes rather than treating one style of expression as the sole entrance to technological participation.
The same applies to vocabulary teaching. Its purpose is not to make every person sound like an academic. It is to offer distinctions that help people think and act. Words such as “estimate,” “evidence,” “constraint,” “exception,” “reliable” and “uncertain” can matter because they mark differences in how a claim should be used.
Education, in this sense, improves the connection between people and tools. It helps us describe the world without pretending our description is the world itself. It helps us recognise when assistance is useful and when a response needs to be questioned.
The most promising technological future is not one in which only already-confident experts can participate. It is one in which more people are helped to become informed participants. Better tools and better education can strengthen one another, provided that learning remains part of the design.
Previous chapter · Contents · Next chapter
12. AI Can Raise Performance Without Raising the Same Kind of Capability
Two questions that sound similar can lead us to very different conclusions.
Did the person produce a better result while using the tool? And did the person become more capable in a way that matters beyond that assisted task?
Both questions are valuable. Neither should be silently substituted for the other.
A 2025 study by Erik Brynjolfsson, Danielle Li and Lindsey Raymond examined the introduction of a generative-AI assistant among 5,172 customer-support agents. Productivity, measured as issues resolved per hour, increased by 15% on average. Less-experienced and lower-skilled workers improved more strongly, and the researchers also reported evidence consistent with learning. This was a particular workplace and implementation, not a universal estimate for every occupation or AI system. [5]
That finding is an important counterweight to the idea that powerful tools must always widen the advantage of people who are already expert. Under some arrangements, a tool can help distribute practices that beginners would otherwise take longer to acquire.
A different 2025 study, by Hamsa Bastani and colleagues, investigated nearly a thousand students in one Turkish high school during autumn 2023. AI assistance improved performance during Mathematics practice. Yet students using the standard-style assistant performed worse on subsequent unassisted tests; the reported reduction was 17%. A teacher-informed tutor largely avoided that negative effect, but did not establish a positive unassisted-test advantage over the control group. The study concerns those designs, tasks and short-term conditions, not the inevitable future of AI in education. [6]
Together, the studies invite a design question rather than a slogan. What is the human activity the tool is supporting, and what do we need the human to learn?
Consider two imagined revision routines. In the first, a learner receives completed solutions, copies the relevant steps and moves quickly through the assignment. The visible work is tidy. But nobody checks whether the learner can identify the relationship in a new question or explain why a step is valid.
In the second, the learner first attempts the problem, receives limited help at the point of difficulty and then explains the reasoning. A later question changes the surface. The teacher uses that later attempt to decide whether the explanation has become usable. The tool is part of a learning sequence rather than the entire sequence.
These routines are illustrations, not a promise that one exact prompt will guarantee learning. They show why the surrounding practice matters. The same broad technology can be placed inside a system that rewards completion or inside a system that investigates understanding.
For a workplace, producing a good tool-assisted result may itself be a legitimate objective. We do not require an architect to give up every design instrument before recognising professional competence. But if future responsibility includes checking unusual cases, adapting the workflow or responding when the tool is unavailable, some capabilities must remain accessible elsewhere in the human-and-technical system.
For a learner, the appropriate objective depends on the lesson. Sometimes the aim is to understand a method personally. Sometimes it is to compare approaches, communicate an idea or use a tool responsibly. An assessment should match that aim. Otherwise, it may test the wrong thing and then call the result a judgement about intelligence.
There is a second useful distinction: the tool can make an expert’s knowledge available without making every user an expert. That can still be a major benefit. The question is whether users understand enough to apply the assistance appropriately and obtain qualified help when the situation exceeds the tool’s dependable range.
We should not make beginners wait until they are already excellent before allowing them helpful tools. Nor should we tell them that access removes the need to learn. A constructive approach offers support while developing the ability to use, question and sometimes decline that support.
The upward spiral is therefore not “more AI, therefore more intelligence.” It is a sequence in which assistance improves worthwhile performance, teaching protects the capabilities needed for future tasks, and evidence is used to revise the arrangement.
The downward spiral is not “using AI is laziness.” It is an arrangement in which attractive immediate results conceal a loss of understanding that later becomes costly. That is a design failure we can look for, study and change—not a reason to abandon curiosity about what the technology can help us do.
Previous chapter · Contents · Next chapter
13. We Do Not Need to Know Everything the Machine Knows
At this point, a reasonable objection appears.
If civilisation has always allowed us to rely on other people and tools, why insist that knowledge must remain inside individual heads? Is the whole point of technology not to free us from having to do everything ourselves?
Yes. That freedom is one of technology’s great gifts.
The purpose of education is not to make every person independently reproduce the entire technical world. Nobody needs to understand every component of a train before becoming a competent passenger. A writer need not manufacture a pen. A mathematician can use a calculator without making the calculator a moral problem.
The issue is not reliance itself. It is whether the larger arrangement of reliance remains trustworthy, interpretable and capable of repair.
Imagine a community that adopts a useful automated service. Some knowledge now sits in software. Some sits with the supplier. Some sits in the people who supervise the service, maintain its equipment, evaluate its outputs and make decisions about its use. The relevant capability is distributed.
The community’s educational task is to develop an appropriate relationship with that distribution. People need to know enough for their responsibilities. Those commissioning the service need a way to assess its suitability. Those operating it need relevant training. Those affected need intelligible routes for questions and complaints. Specialists need the competence and authority to investigate failure.
No individual has to contain the whole system. But the whole system should not be arranged so that nobody can explain what matters when something goes wrong.
This perspective also guards against nostalgia. Some old skills can reasonably become less common when a dependable new method makes them unnecessary. The goal is not to preserve every former task forever. It is to decide deliberately what can be delegated, what remains strategically important and what new capabilities the delegation requires.
An early evaluation of the One Laptop per Child programme in rural Peru illustrates the difference between acquiring equipment and achieving an educational objective. The study covered 319 schools and found increased access to and use of computers, but no evidence of gains in the measured Mathematics and language outcomes after fifteen months. It also reported improvements on some general cognitive measures. The mixed findings do not mean computers cannot help learning; they show why hardware distribution is not a complete theory of educational change. [7]
A useful next question is therefore not simply whether a school has technology. It is what learners do with it, what teachers can see, which barriers it removes and how progress is assessed. An accessible device may enable a child to participate in a lesson at all. That benefit should not disappear merely because a narrow test does not capture it.
This is particularly important when discussing independence. Independence should not mean being deprived of legitimate assistive technology, language support or the accommodations that make participation possible. A person who uses a screen reader is not less entitled to be recognised as a reader. An assessment that removes an essential access tool may reveal exclusion rather than lack of understanding.
The better question is whether a person has appropriate agency within the supported arrangement. Can they pursue a purpose, understand relevant choices, recognise when help is needed and participate without unnecessary dependence on another person’s permission?
At civilisation scale, the same question concerns institutional agency. Can a society maintain access to important capabilities, negotiate with providers, train relevant specialists and recover from foreseeable disruptions? These abilities do not require complete national self-sufficiency. They require a thoughtful relationship with interdependence.
We can therefore welcome a world in which machines hold more knowledge and perform more tasks. We should also ask what kinds of human understanding become more valuable because of that change. Interpretation, judgement, coordination, accountability and the ability to learn new arrangements do not become irrelevant simply because some procedures are automated.
An upward spiral uses delegation to enlarge worthwhile human possibilities. A downward one delegates so carelessly that people lose the ability to govern the dependence they have created.
The distinction is not whether the tool is powerful. It is whether the relationship remains capable of learning.
Previous chapter · Contents · Next chapter
14. When Private Wealth Becomes a Public Starting Point
The Crazy Rich stories are enjoyable because they make abundance visible. They invite us into extraordinary rooms, unusual collections and places where people can purchase experiences far removed from ordinary budgets.
For this essay, their most interesting question is not how much somebody can spend. It is what wealth can become after it has been accumulated.
Does it remain a private collection of advantages? Does it support work, culture and care? Does it help create institutions that other people can enter? Does it finance the next round of capability, or merely display the results of an earlier round?
One of the newer articles, Crazy Rich Japan’s account of Izumi City and the Kuboso museum, offers a distinctive example. According to the museum’s own history, after the Kuboso business closed in 1977, Sotaro Kubo III donated an art collection, museum building, site and funds to the city. The museum opened in 1982, and later donations expanded its facilities. [10]
The case gives us something more useful than a general celebration of generosity. It shows a change in the form of an inheritance. Private assets can be organised into a public institution through which people beyond the original family encounter art, knowledge and cultural memory.
The donation is not the end of the story, however. A museum has to be cared for, interpreted and made accessible. Its collection requires people who know how to preserve objects, study them and help visitors understand what they are seeing. The gift opens a possibility; continuing human work keeps the possibility usable.
This is where wealth and education meet without needing to reduce culture to economic output. A child encountering an unfamiliar artwork may not become a museum professional or earn a higher salary because of that afternoon. The experience can still matter. It may enlarge attention, curiosity, historical understanding or the range of things the child finds possible to love.
A civilisation worth improving has room for benefits that are not immediately converted into a productivity statistic. Music, play, conversation, beauty and the freedom to explore are not embarrassing extras attached to a serious economy. They belong among the reasons we want material security in the first place.
The same conversion can occur in less spectacular settings. Imagine a successful local business supporting a well-run apprenticeship, a neighbourhood reading room or a practical workshop. The contribution need not carry a famous name. Its importance may lie in making a first opportunity available to someone who would otherwise not have had one.
But private generosity is not a complete substitute for dependable public provision. Access to essential education should not depend entirely on whether a donor happens to prefer a particular town, subject or group of children. Public institutions, fair funding and durable responsibility remain important.
The healthiest picture includes several forms of contribution. Families care and teach. Communities create opportunities. Businesses train and invest. Public institutions protect broad access. Cultural organisations preserve and reinterpret. None should be made an excuse for the others to withdraw from their proper responsibilities.
We should also be cautious about treating wealth as proof of wisdom. A person can accumulate resources while misunderstanding the educational needs of others. A large donation can be poorly designed. Good intentions do not eliminate the need to listen, evaluate and maintain what has been created.
The useful question is what happens after the announcement. Can people access the opportunity? Does it meet a real need? Are the people running it supported? Can it adapt? Does it create dependence on one benefactor, or does it help build a more durable institution?
Seen this way, abundance becomes an input into the next starting point. It can buy a larger private room, which may be perfectly enjoyable. It can also help build a room in which many other people begin learning.
The upward spiral is not a demand that every pleasure become a public project. It is an invitation to notice that prosperity has a second life when it helps others acquire possibilities of their own.
Previous chapter · Contents · Next chapter
15. Success Can Make the System Harder to Carry
An uncomfortable thought now enters the cheerful story: success can create new difficulties precisely because it succeeds.
A useful service attracts more users. More users justify expansion. Expansion encourages specialisation. Specialisation makes greater efficiency possible, but also creates more handovers. More handovers require coordination. The system may become more capable while also becoming harder to understand and maintain.
Nothing in that sequence makes decline inevitable. It simply means that the learning required to operate the system changes as the system grows.
Imagine a small organisation in which everybody once knew everybody else’s work. A problem could be solved by a conversation across the room. Years later, the organisation has several departments, specialised software and a much larger customer base. The old informal arrangements no longer carry the same load.
The organisation now needs clearer documentation, appropriate responsibilities and new ways for consequences to travel between departments. If it fails to develop them, it may continue expecting a small-company style of coordination inside a much more complicated structure.
The resulting mistakes can be misread as a decline in personal commitment. Yet the people may be working harder than before. What has changed is the relationship between the complexity of the work and the means available to coordinate it.
Education has to respond at several levels. New employees need enough understanding to perform their tasks. Experienced employees need to learn how the surrounding system has changed. Managers need to recognise interfaces and unintended consequences. Specialists need ways to communicate with people whose expertise is different from their own.
The answer is not to demand that every employee understand every detail. That would be another form of overload. The aim is to make the system intelligible at the appropriate level: what am I responsible for, what do I rely on, whom does my decision affect, and where can I obtain the knowledge I do not personally hold?
An imagined school can face the same problem. It adopts several useful programmes, each with its own reporting requirements, resources and training. Every addition may have a reasonable purpose. Together, however, they can consume the time teachers need to plan lessons, interpret student work and coordinate support.
A programme designed to improve learning can then interfere with the conditions of teaching. This is not an argument against innovation. It is a reason to examine the combined load rather than approving each addition in isolation.
Success can also hide the work on which it depends. A reliable service becomes taken for granted. Because failures are rare, the organisation may see training and maintenance as expenses with little visible return. It can then reduce the activities that made failure rare in the first place.
The danger is a delayed one. Experienced people may continue compensating for a while. Existing assets may continue working. The loss becomes obvious only after enough knowledge, time or resilience has been consumed.
This is the mechanism behind a possible inversion: the system’s visible success begins to weaken its willingness to fund the invisible work that supports success. It is not a law of prosperity. It is a pattern to investigate when a supposedly efficient organisation becomes increasingly dependent on exhausted experts and last-minute intervention.
The appropriate response is not to worship redundancy or spend without limit. Resources are finite, and some procedures genuinely can be simplified. The task is to distinguish waste from the reserves, skills and checks that allow a system to remain dependable under realistic conditions.
A mature upward spiral includes this kind of subtraction and renewal. It removes unnecessary burden while protecting the capacity to learn and recover. It asks whether the next improvement makes the system easier to carry, not only whether it makes the current dashboard look better.
Civilisation does not advance simply by adding more layers. Sometimes it advances by making the relationships between its layers clearer—and by giving the people who carry them enough time to understand what has changed.
Previous chapter · Contents · Next chapter
16. The Downward Spiral Often Begins With a Misread Signal
A result deteriorates. Someone must explain it.
This is a small moment with large possibilities. The explanation can open a route to repair, or it can become the first turn of a downward spiral.
Imagine a department whose work is taking longer. One explanation is that employees have become less committed. Another is that cases have become more complicated. A third is that the department’s new software creates extra steps. A fourth is that experienced staff have left and new staff are still learning. More than one explanation may be relevant.
If management chooses a cause before investigating, it can make the wrong intervention. Tighter deadlines may help in one setting and make another worse. More monitoring may reveal a useful pattern or consume the time needed to do the work. Reorganisation may solve an interface problem or destroy a relationship that was keeping the system functional.
The downward spiral begins when the intervention worsens the conditions, and the resulting difficulty is then treated as confirmation of the original mistaken explanation.
This pattern is explored in How Intelligence Works | When the Spiral Goes Down. A system can receive feedback, change in response and still become worse. The missing ingredient is not necessarily activity. It may be sound interpretation, proportionate updating or a willingness to test the preferred explanation.
A school example is equally revealing. Suppose students perform poorly on unfamiliar questions. The response might be to supply more examples of exactly the same question format. Familiar-format scores then improve. The school concludes that the repair worked, but later discovers that performance still collapses when the wording changes.
The intervention improved recognition of the surface without necessarily repairing the underlying understanding. If the response is then even more dependence on supplied patterns, a self-reinforcing arrangement can develop. Adults do more of the interpretation; students receive fewer opportunities to practise it; the next unfamiliar task requires still more adult interpretation.
Again, this is a possible mechanism, not a verdict on every use of model answers. Examples can be excellent teaching tools. The question is what the learner is helped to notice and whether responsibility for the reasoning eventually shifts towards the learner.
At national scale, claims about decline require the same discipline. The OECD’s 2024 release of results from its 2023 Survey of Adult Skills reported stagnation or decline in literacy and numeracy in many participating countries compared with earlier rounds, alongside important differences across countries and groups. Such evidence challenges the assumption that skills automatically keep improving in affluent societies. It does not, by itself, identify a single cause or justify declaring an entire civilisation in decline. [11]
A responsible response asks how the measures were constructed, who was included, which comparisons are appropriate and what additional evidence would distinguish explanations. Demographic change, educational experience and opportunities to use skills should not be collapsed into a convenient story before they are investigated.
There is also a social condition for getting the diagnosis right: people must be able to report unwelcome information. A teacher needs room to say that coverage is ahead of understanding. A worker needs a credible route to report a recurring problem. A researcher needs permission to publish a result that does not support the preferred policy.
This does not mean accepting every complaint as a complete account. Reports need checking. But if reporting a problem reliably harms the reporter, the system should not be surprised when its information becomes less useful.
The upward alternative is a learning culture in which evidence can revise an explanation without requiring someone to defend their entire identity. “Our first interpretation was incomplete” should be a usable sentence.
That sentence can save considerable effort. It allows the organisation to stop making the same problem worse with increasingly determined versions of the wrong response.
Civilisation does not need leaders, teachers or citizens who never make mistakes. It needs arrangements in which mistakes can become visible, explanations can change and correction is not treated as humiliation. A spiral turns upward more readily when reality is allowed to interrupt the story we would prefer to tell.
Previous chapter · Contents · Next chapter
17. The Future Can Develop a Shortage Before the Present Notices
Imagine a service with one hundred fully qualified workers.
Each year, ten leave and six newly qualified people join. If those simple flows remain unchanged, the workforce falls to ninety-six after one year, ninety-two after two and eighty-eight after three. The numbers are hypothetical; real workforces have more varied entries, exits, hours and skills. But the arithmetic reveals a planning problem.
At the beginning, the service may still appear to function. Experienced staff absorb the gap. Some work is postponed. Overtime increases. The public sees the service continuing and may not see the weakening route through which future staff are supposed to arrive.
Now suppose the organisation launches a training programme to supply four additional qualified workers a year. If training takes three years, announcing the programme does not create four qualified workers tomorrow. There is a delay between recruiting learners and having people ready to carry the responsibilities of the role.
Even when the additional four begin arriving, they merely close the annual gap between ten departures and six previous entrants. Under the simplified assumptions, the organisation would then stabilise at a lower headcount. Recovering the earlier level requires a further period in which suitable entries exceed departures, or a justified redesign that changes how much capacity is needed.
This is why replacement and expansion are different tasks. It is also why preventing unnecessary departures can matter alongside training new entrants. A system cannot recruit its way out of every retention problem while leaving the reasons for departure untouched.
UNESCO’s 2024 teacher-planning guidance described an estimated need for forty-four million additional primary and secondary teachers to reach universal education goals by 2030. It identified replacement of people leaving the profession as a large part of the secondary-teacher requirement. These are planning estimates for a stated goal and period, not a live count of empty classrooms today. They illustrate why sustaining education includes renewing the people who provide it. [12]
The numerical example still leaves out an important detail: one person is not interchangeable with any other person. Subject knowledge, experience, working hours, language, location and supervisory ability affect what a workforce can actually do. A total headcount can look adequate while a particular responsibility has no suitable successor.
The departure of a skilled trainer can be especially consequential. The organisation loses not only someone who performs the work, but someone who helps others become able to perform it. Its present capacity and its ability to create future capacity can weaken together.
This gives rise to a second downward loop. Fewer experienced staff carry more work. Because they are carrying more work, they have less time to teach. Because there is less teaching, new staff become capable more slowly. The experienced staff then carry even more of the difficult work.
The solution cannot simply be a moral appeal for those experts to work indefinitely harder. A serious plan protects time for supervision, distributes knowledge, improves conditions and recognises that training is part of the organisation’s production of future capability.
Technology may help. Better tools can reduce unnecessary work, support training or change staffing needs. But those gains should be demonstrated in the relevant setting rather than assumed as a convenient way to erase an emerging shortage from a spreadsheet. New systems can also introduce new maintenance and training requirements.
Nor should workforce planning turn young people into interchangeable units assigned to vacancies. Their interests, rights and aspirations matter. A healthy route into a profession makes the work understandable, accessible and worth choosing. It does not merely announce that society needs people and expect individuals to absorb every cost of becoming them.
The lesson extends beyond employment. Families, voluntary organisations, artistic traditions and community institutions also depend on people learning roles before the current carriers can no longer perform them. Continuity is easier when a handover begins while there is still time to teach.
A civilisation can therefore develop a future shortage while its present still looks complete. Education is one of the ways we notice and repair that gap before the missing person is urgently needed.
Previous chapter · Contents · Next chapter
18. Sometimes the Door Is Closed From the Outside
We must now make an essential distinction. Not every downward educational spiral begins with poor study habits, mistaken teaching or an institution’s reluctance to learn.
Sometimes people are prevented from learning. Sometimes a school is inaccessible. Sometimes conflict, displacement, a disaster or an exclusionary policy removes the conditions in which education can continue.
In those circumstances, advice about trying harder is not merely insufficient. It can obscure where responsibility belongs.
In their joint statement of 24 January 2026, UNESCO and UNICEF described the exclusion of girls and women from secondary and higher education in Afghanistan and warned about the consequences for future opportunities and the supply of qualified women teachers and health workers. Their statement identifies both an immediate denial of education and a threat to the future provision of essential services. [13]
The first concern must remain the people whose lives are restricted now. A girl’s education does not become important only because she might later provide a service to somebody else. She deserves the opportunity to learn, develop interests, make choices and participate in a wider life.
The long-term system effect is an additional consequence, not the only justification. When a route into advanced learning is closed, fewer people can progress through that route into professions that require it. If some of those professions are themselves needed to teach the next cohort, the restriction can reproduce its own damage.
This is a particularly clear example of how one generation’s opportunities become another generation’s conditions. A barrier placed in front of today’s learner can later appear as a shortage of suitable teachers, mentors or professionals. The effects travel through time and across institutions.
That analysis must not become a judgement about the inherent capacities of a population. People living under restrictive policies are not the same thing as the policies imposed on them. A nation is not a single mind with a single intention. The existence of exclusion tells us about access and power, not about the potential of the people being excluded.
Environmental disruptions offer a different route into the same continuity problem. UNICEF’s review of 2024 estimated that climate-related events disrupted schooling for at least 242 million students in eighty-five countries or territories. Disruption included different events and durations; the figure is not a count of children permanently leaving school. It shows how education can be interrupted by conditions beyond the classroom. [14]
The educational consequences of an interruption depend on what follows. Is there a safe alternative place to learn? Can teachers stay connected with students? Are materials usable without a stable internet connection? Can children return without being treated as if missed instruction were a personal failure? Are families also dealing with lost housing, income or transport?
These questions point to a joined-up response. Reopening a building is important, but continuity may also require transport, accessible materials, teacher support and a realistic way to identify what learners have missed. A timetable cannot erase an interruption simply by resuming at the chapter that would otherwise have been reached.
The same principle applies to displaced learners entering a new school. Their previous knowledge may not align neatly with the receiving system’s language or sequence. A capable student can appear behind when the assessment assumes a different curriculum or form of expression. Diagnosis should distinguish missing knowledge from a difficulty in demonstrating knowledge under new conditions.
A civilisation that spirals upward after disruption is not one that denies the disruption. It is one that prevents temporary interruption from becoming permanent exclusion wherever possible. It creates ways back in.
This may require resources from beyond the affected institution or country. Interdependence is not only a source of vulnerability; it can also carry assistance, expertise and opportunities for recovery. The quality of that assistance depends on listening to the people involved rather than arriving with a ready-made story about their deficiencies.
Education can help rebuild a damaged society, but it should not be asked to compensate alone for violence, exclusion or the loss of basic conditions. Sometimes the first educational intervention is to reopen the door that should never have been closed.
Previous chapter · Contents · Next chapter
19. Recovery Begins by Finding the Actual Starting Point
There is a particular kind of relief in discovering that a large, frightening problem contains a smaller problem that can be worked on.
“I cannot do Mathematics” becomes “I do not yet understand this relationship.” “This group cannot read” becomes a more careful account of what different learners can recognise, decode, understand and explain. “The organisation is failing” becomes a map of particular gaps, constraints and responsibilities.
The smaller description does not make every problem easy. It makes action more precise.
Teaching at the Right Level, developed by Pratham and evaluated in a series of studies, offers a practical educational example. It uses assessment to identify children’s current foundational reading and Mathematics levels, provides focused instruction and checks progress. J-PAL’s account stresses that successful implementation involves more than grouping children: clear objectives, support for teachers, regular use of learning information and continuing mentoring matter. The evidence concerns defined interventions, not a guarantee that any programme calling itself “personalised” will work. [15]
The important connection to the spiral is that repair changes the point from which the next learning cycle begins. Instead of repeatedly confronting a learner with material they cannot yet use, teaching reconnects the learner to an attainable next step.
This requires care. Assessing a current level should not become assigning a permanent identity. A child’s present performance is information about what to teach, not a final judgement about the height of the child’s future. Grouping arrangements should be revisable and used for a clear educational purpose, not as a quiet mechanism of exclusion.
Imagine three learners receiving the same difficult paragraph. One needs help recognising some words. Another can read the words but lacks the background knowledge to understand the situation. A third understands the passage but struggles to explain the author’s reasoning. More of the same worksheet would not necessarily provide the same benefit to all three.
A useful teacher can offer different entry points while keeping a shared ambition: each learner should become better able to understand and communicate. The target is not to make everybody perform the same small exercise forever. It is to build the capability required for a wider range of work.
The organisational parallel is straightforward. A recurring problem should be examined at the point where the process becomes unreliable. The solution may involve training, but it may instead involve unsuitable equipment, unclear responsibility, an impossible workload or a policy that encourages the wrong action. Calling every problem a training need can be as imprecise as calling every school mistake careless.
Diagnosis therefore needs to include the environment. What can the person do under reasonable conditions? What changes when the task is presented differently? Which resources are missing? What happens when the relevant constraint is removed? The aim is to locate a useful intervention, not to collect ever more labels.
Recovery also benefits from an order. First, stop avoidable damage where possible. Then restore enough stability for learning to occur. Repair the important gap, practise the revised approach and check whether the change survives. Only then does it make sense to add further complexity at scale.
This sequence is a practical heuristic, not a rigid rule for every situation. Emergencies can require several actions at once. Some repairs must occur while ordinary work continues. But the principle protects us from a familiar mistake: adding an ambitious new layer before the earlier layer can carry it.
An upward spiral can begin with modest outcomes. A learner starts a task without immediate surrender. A teacher can identify which explanation helped. A team reduces one recurring failure. An institution becomes better at noticing the difference between a temporary fix and a durable improvement.
Such gains are not trivial merely because they lack drama. They create evidence that a problem is changeable. That evidence can support confidence, better participation and the next round of work. Encouragement becomes more credible when it is attached to something the learner has actually managed.
The hopeful conclusion is not that every difficulty can be repaired quickly. It is that a current difficulty need not be mistaken for a permanent destiny. A good starting point is not always the most advanced-looking one. It is the one from which the next meaningful step is genuinely reachable.
Previous chapter · Contents · Next chapter
20. Learning Has a Physical and Human Floor
It is easy to discuss education as if it consisted only of information moving into minds.
But imagine trying to complete a careful piece of work while the room is unusable, the instructions are inaccessible, the journey has become unsafe or the people responsible for helping you have no time left. The problem is not simply a shortage of information.
Learning happens inside a life. It depends on arrangements that allow people to be present, attend, practise, ask questions and return. Some of those arrangements are educational. Others belong to housing, transport, health, safety, accessibility, family life and working conditions.
There is no universal checklist for an ideal household here. The practical mistake is to treat a learner as if they were separate from the circumstances in which learning is expected to occur.
Consider an imagined adult returning to study. The course is well designed, and the person is motivated. Yet the timetable conflicts with caring responsibilities, the workplace offers no flexibility, and the course assumes uninterrupted evenings. The learner’s difficulty may be a collision between otherwise reasonable demands, not a lack of interest.
A more useful response could involve a different schedule, accessible materials, recognition of prior learning or a realistic pace. The exact solution would depend on the learner’s circumstances. What matters is that the institution considers whether its route is actually usable by the people it claims to serve.
Teachers have a corresponding human floor. An educational system can demand careful feedback, differentiated support, communication with families and professional learning. If it also fills every available hour with competing obligations, the quality of these activities may depend increasingly on personal sacrifice.
An upward spiral should not rely on educators continually donating the time they need for recovery and a life outside work. Nor should it treat every limit as evidence of insufficient dedication. Sustainable care includes the conditions of the people providing care.
The same reasoning applies to families. Parents can support learning without becoming unpaid substitutes for every institutional function. A household should not have to possess expert knowledge in every subject to give a child a fair chance. Advice that assumes unlimited money, time, space and confidence can inadvertently turn inequality of resources into a judgement about parenting.
A practical educational approach asks what support is proportionate and possible. It offers clear information, manageable next steps and routes to qualified help when needed. It does not make anxious families believe that every spare minute must be converted into an academic advantage.
Material systems belong in the picture too. A digital resource depends on devices, connectivity, power and maintenance. A science room depends on suitable equipment and responsible supervision. A library depends on access, curation and people who can help readers find their way. Calling something a knowledge resource does not make its physical requirements disappear.
The ecological relationship is similar. We cannot sensibly celebrate an educational or technological gain while refusing to examine the resources and environmental burdens through which it is sustained. This does not mean every activity must have no footprint. It means that costs should remain visible enough to be governed and reduced where possible.
The self-consuming spiral hides these supporting conditions. It presents a polished output while transferring strain to families, workers, communities or the future. The regenerative spiral asks what must be restored so that participation remains possible after the current round of effort.
There is a cheerful version of this principle, too. A functioning learning environment can release energy that was previously spent navigating avoidable difficulty. A clear instruction, a reliable journey, an accessible text or a protected period for teaching can make the work feel less like a contest against the surroundings.
Not every improvement needs to be a grand innovation. Sometimes the most enabling change is to remove an unnecessary obstacle between a person and a worthwhile activity.
Education becomes stronger when we stop treating these conditions as external to the educational story. The floor is not beneath the subject. It is part of how the subject becomes learnable.
Previous chapter · Contents · Next chapter
21. Better at What—and Better for Whom?
An upward arrow looks innocent until we ask what it points towards.
A system can become more efficient at an activity that harms people. An organisation can become more skilled at protecting itself from criticism. A learner can become better at manipulating a measure without understanding the subject. Increased capability and improved human life are not automatically the same thing.
This is why education cannot be discussed only as a means of producing more competent workers or more powerful institutions. We also have to ask what purposes competence serves and who has a voice in deciding them.
Suppose an imagined organisation develops an extremely effective monitoring system. It can identify deviations quickly and enforce compliance. Whether this counts as an improvement depends on what is being monitored, why, with what safeguards, and how the people affected can challenge mistakes. Technical success does not settle the ethical question.
The same applies to a school. Perfectly orderly classrooms would not necessarily demonstrate that students were learning to think. The order might support attention and respectful participation. It might also be produced by making questions unwelcome. We need to examine the educational relationship, not merely admire the absence of noise.
A worthwhile upward spiral should enlarge the conditions in which people can live with dignity, develop capabilities, participate in shared life and exercise meaningful choice. This is a statement of the values guiding this essay, not a neutral mathematical discovery. Other people may phrase those values differently, and disagreements require discussion rather than being hidden inside a score.
Education contributes by helping people understand consequences and arguments, encounter perspectives beyond their own and distinguish evidence from persuasion. But education itself must remain open to scrutiny. A curriculum can transmit useful knowledge while also carrying assumptions that deserve examination. An institution does not become exempt from criticism because its mission is educational.
There is a difference between teaching people how to reason and telling them that reasoning is acceptable only when it arrives at an approved conclusion. There is also a difference between respecting experience and making authority impossible to question.
A strong learning culture can hold respect and criticism together. Students need not humiliate a teacher to challenge an explanation. Teachers need not surrender their expertise to acknowledge uncertainty. Institutions need not regard every disagreement as disloyalty. The quality of the exchange depends on reasons, evidence, care and appropriate responsibility.
Human purposes also include things that do not resemble upward economic movement. A person may choose a quieter life, more time with family or work that is less prestigious but more meaningful to them. It would be strange to call every such choice regression because it does not maximise earnings or status.
Similarly, a society might decide not to pursue an available technology in a particular setting because the risks, costs or effects on agency are unacceptable. Refusing one form of expansion can protect other forms of progress. The presence of a capability does not create an obligation to use it everywhere.
This protects the spiral metaphor from becoming an endless demand. “Upwards” should not mean that nobody is ever allowed to be satisfied, rest or enjoy an ordinary day. A civilisation that turns every person into a permanent improvement project may increase pressure while diminishing the life it claims to advance.
The more generous ambition is to create room. Room to learn, room to contribute, room to recover, room to disagree, and room to choose a worthwhile life that does not have to look identical to somebody else’s.
Education helps when it offers more than a route to competitive advantage. It can give a person the language to understand a difficult situation, the knowledge to recognise a misleading claim, the confidence to ask for help and the imagination to see an alternative.
Those are forms of capability, too. Some will produce measurable economic benefits. Others matter because they allow people to be less easily confined by fear, confusion or someone else’s narrow account of what their lives should become.
The question is therefore not simply whether civilisation can do more. It is whether more of its ability can be placed in the service of lives people have reason to value.
Previous chapter · Contents · Next chapter
22. There Is More Than One Route Into a Capable Society
A city does not work because every child grows up to do the same prestigious job.
It works because different forms of knowledge can meet without some of their carriers being treated as less worthy people. The person who designs, the person who maintains, the person who teaches, the person who organises and the person who cares do not contribute in identical ways. Their differences are part of the arrangement.
This is an ethical and practical argument for taking multiple educational routes seriously. It does not require pretending that every course leads to the same opportunities or that all qualifications are interchangeable. It requires refusing to turn those differences into a hierarchy of human worth.
Imagine a student who enjoys understanding how physical things work. Another enjoys writing and interpretation. Another is drawn to caring roles, organisation, design or mathematical abstraction. A good educational system can help each explore strengths while also developing foundations that keep later choices open.
The danger is premature closure. A temporary difficulty can become a permanent track. A narrow test can be mistaken for the full extent of a person’s potential. A child can learn that one route is supposedly for people who failed at the route that matters.
Such messages are not necessary for maintaining standards. We can require competence in a demanding field without treating other forms of competence as lesser lives. We can acknowledge different prerequisites without pretending that people develop every ability at the same time.
An upward spiral benefits from permeability: sensible routes between stages, opportunities to return, recognition of relevant experience and ways to repair a missing prerequisite. This does not mean removing all entry requirements. It means making requirements intelligible and offering appropriate ways to meet them.
Consider an imagined adult who has spent years organising complex work but lacks a formal credential required for advancement. A responsible institution should not simply assume that experience proves every relevant skill. Nor should it assume that the absence of one certificate means no competence exists. Assessment, recognition and targeted learning can help distinguish what is already present from what still needs to be developed.
The same principle applies to learners whose abilities are difficult to see through a standard presentation. Accessible assessment, appropriate accommodations and multiple ways of demonstrating understanding can reveal competence that a single format obscures. Fairness is not always achieved by giving everybody an identical obstacle.
Access also has a timing dimension. An opportunity offered at the wrong stage, with no support for entry, may exist formally without being practically usable. Information about a course is less helpful if the applicant discovers it only after a deadline, cannot understand the requirements or has no way to meet the associated costs.
This is why guidance matters. Good guidance does not decide a person’s life for them. It helps them see the routes, constraints and possible next steps clearly enough to make a more informed decision. The educational system becomes easier to navigate rather than more dependent on private insider knowledge.
We should be especially careful with the phrase “talent pipeline.” It can be useful for planning, but it can also make people sound like material being fed into someone else’s production process. Learners have purposes of their own. A society’s need for skilled people does not erase the learner’s right to choose among meaningful possibilities.
The best arrangement allows those interests to meet. Work is made worth doing. Training is made accessible. Standards remain real. People can see how to develop and move without being required to sacrifice their entire identity to one route.
A capable society is not one that sends everybody through the same doorway. It is one that keeps enough well-supported doorways open, and enough connections between them, for diverse people to contribute and continue learning.
This matters for the spiral because exclusion wastes possibility before it becomes visible. We cannot know every contribution that would have emerged from a person who never received a usable opportunity. The reasonable response is not to claim that everyone will become exceptional. It is to stop making unnecessary barriers the reason ordinary and extraordinary capabilities fail to develop.
Previous chapter · Contents · Next chapter
23. Civilisations Learn Across Borders, Not Only Within Them
The image of a civilisation rising alone can be misleading.
A more useful starting point is that knowledge can travel. People move, texts are translated, methods are adapted, instruments are shared and institutions learn from work done elsewhere. A society’s capacity need not be limited to what originated within its current borders.
This is also a conceptual point about the spiral. If useful knowledge enters from outside, the next round can begin from a better starting point. If a local discovery becomes available elsewhere, its benefits can extend beyond the place that first developed it. The loop can cross organisational and geographical boundaries.
There is no need to turn this into a romantic claim that all exchange is fair. Knowledge can be appropriated without recognition. Skilled people can be welcomed for their labour while excluded from full participation. Institutions can adopt an imported programme without understanding the conditions that made it work in its original setting.
The quality of exchange matters as much as its existence.
Imagine a school system interested in an approach that produced promising results abroad. A weak transfer copies the visible materials and assumes that the result will follow. A stronger transfer asks about the original learners, language, teacher preparation, timetable, resources and implementation. It then tests a suitably adapted version rather than confusing a borrowed label with a reproduced mechanism.
This distinction is familiar from individual learning. A student who copies the surface of an example may miss the relationship that made the example work. Institutions can make the same mistake at a much larger scale.
Migration raises an additional set of questions. People should not be treated as possessions of the place that educated them. They have legitimate reasons to move, including safety, family, opportunity and freedom. At the same time, a community can face real continuity problems when important expertise leaves faster than it can be replaced.
A serious response holds both concerns. It does not blame individuals for seeking a better life. It asks how working conditions, professional development, recognition and institutional quality can make continued participation or future collaboration more attractive.
The possible benefits of movement are not limited to permanent return. People can share expertise, teach, collaborate, support networks or connect organisations across places. These are possible routes, not automatic compensation for every loss. Whether they emerge depends on relationships and opportunities that need to be maintained.
A receiving institution also has responsibilities. It should not rely on skilled newcomers while ignoring the conditions that allow them to practise effectively, have their qualifications fairly assessed and participate with dignity. The relevant contribution is made by a person, not by a credential detached from a life.
The same care is needed when outside organisations offer educational assistance. A programme can bring resources and useful expertise while still misunderstanding local constraints. Listening to learners, teachers and communities is not a ceremonial addition to implementation. It is part of discovering what the programme needs to work.
An upward spiral across borders therefore includes translation in a broad sense. Not only translation between languages, but translation between contexts, institutions and practical conditions. What exactly is being transferred? Which assumptions travel with it? What must change? How will the receiving setting know whether the transfer is useful?
These questions make global connection more demanding and more promising. They move us beyond the choice between total self-sufficiency and uncritical dependence. A society can learn from others while developing the judgement to adapt what it receives.
Education has a double role here. It helps people participate in wider networks of knowledge, and it helps them avoid mistaking outside prestige for local suitability. A famous method still needs to work for the learners in front of us.
The hopeful consequence is that no community has to be the original inventor of every capability it hopes to use. The demanding consequence is that receiving knowledge remains active work. A civilisation rises through connection when it can understand, adapt, reciprocate and keep learning—not merely when it acquires the latest imported vocabulary.
Previous chapter · Contents · Next chapter
24. Truth Needs Somewhere Safe to Arrive
A child says, “I do not understand.”
A worker says, “This keeps happening.”
A teacher says, “The students can repeat the explanation, but they cannot yet use it.”
A researcher says, “The result does not support the claim.”
Each statement offers a chance to improve the next round. Each can also be made expensive to say.
The spiral depends not only on whether information exists, but on whether useful information can reach a place where it changes action. A society can possess excellent instruments and still make poor decisions if the resulting evidence is ignored, distorted or prevented from reaching the relevant people.
Imagine an institution that announces a strong commitment to learning from mistakes. Its official language is encouraging. Yet the first person to report a problem is blamed for damaging confidence. Other people notice. The institution has now taught a practical lesson that contradicts its stated lesson: keep difficulties private until concealment becomes impossible.
This is education too, although not the education the institution claims to provide. People learn from consequences as well as instructions. A poster about openness cannot reliably overcome a repeated experience that openness is punished.
The reverse arrangement is possible. A concern is heard, checked and handled proportionately. The person reporting it is not automatically treated as either a hero or a troublemaker. The institution separates the quality of the evidence from the status of the messenger. Over time, people gain a reason to believe that reporting a difficulty can lead to useful work.
This is not the same as abandoning accountability. Deliberate misconduct, negligence, honest error and a reasonable decision under uncertainty are different situations. A system that treats them identically may become less capable of distinguishing what actually needs to change.
Education can help by developing the language of those distinctions. “We observed this” is different from “we have established why it happened.” “The evidence is incomplete” is different from “nothing can be known.” “This claim needs checking” is different from “the person making it is worthless.”
These are not merely polite phrases. They preserve room for inquiry. When disagreements immediately become total judgements about people, it becomes harder to investigate the claim itself.
The article on reading research abstracts offers a small but important version of this discipline. A careful reader keeps a claim connected to what was measured, how it was studied and what remains uncertain. That habit becomes useful wherever confident language might outrun evidence.
It also protects hope. Honest uncertainty is not the enemy of action. It allows us to choose an intervention proportionate to what is known, identify what to observe next and revise when better evidence arrives. Pretending to know everything can make an institution less able to notice the part it has misunderstood.
Of course, endless questioning can also become a way to avoid responsibility. Some decisions must be made before every uncertainty is resolved. The aim is not permanent hesitation. It is to make the reasoning, assumptions and consequences sufficiently visible that the decision can be evaluated and improved.
A good learning culture therefore needs both inquiry and execution. Ask what matters, act within appropriate limits, observe what follows and retain the useful change. Neither discussion without action nor action without reflection closes the loop well.
At the personal level, this can begin with a small change in response. When a learner admits confusion, the adult does not treat the admission as an inconvenience to the lesson plan. It becomes information about where teaching should go next.
At the institutional level, the same principle becomes more demanding. People need credible channels, proportionate protections and responsible decision-makers. The details differ by setting, but the underlying purpose remains: reality must be allowed to arrive before it becomes an emergency.
A civilisation’s intelligence is therefore partly a quality of its relationships. It is not enough to have people who can see a problem. The system needs to be capable of hearing them.
Previous chapter · Contents · Next chapter
25. Seven Questions That Make the Direction Easier to See
How, then, can we tell whether a spiral is moving upwards or downwards without reducing civilisation to one grand score?
Start with a smaller object: a school, a service, a profession, a community programme or a particular public capability. Define the period under discussion and the people affected. Then ask questions that connect current results to the conditions of future results.
The following seven questions are a practical framework for inquiry, not a validated national index. They should guide the collection and interpretation of evidence, not replace it.
Are people actually learning?
Look beyond participation and completion. What can learners understand, explain or do that they could not do before? Which observations support that conclusion? Where appropriate, does the improvement remain visible after a delay or in a changed task? Different purposes require different checks; a written test is not the answer to every educational question.
Can capability travel beyond the original expert?
Ask whether useful knowledge can be taught, located and applied by others. Does the organisation depend on one person’s private memory? Can a new participant find the relevant explanation? Are the conditions and limits of a method preserved alongside its instructions? Distribution matters because continuity cannot depend indefinitely on the same carriers.
Are successors becoming ready in time?
Examine entry, development, retention and departure together. Is there enough supervised opportunity for people to acquire the responsibilities they will later inherit? Are trainers themselves being replaced? Do plans recognise the time between beginning a course and becoming ready for a demanding role?
Do benefits return to the foundations?
Consider what happens after success. Are some gains used to support maintenance, care, access, learning and future capacity? Or does present performance increasingly depend on postponed work and hidden sacrifice? The answer does not require every benefit to be reinvested. It requires a sustainable relationship between enjoyment and renewal.
Who is gaining—and who is carrying the cost?
Averages can conceal different trajectories. Ask whether opportunity is widening, whether important groups face unnecessary barriers and whether apparently inexpensive performance transfers unreasonable burdens elsewhere. A system can improve its headline output while making participation harder for the people who keep it working.
Can errors change the system?
Inspect the route from an observed problem to a decision. Are reports checked? Can assumptions be revised? Do repairs address recurring causes, or are the same incidents repeatedly resolved without learning? A system’s response to unwelcome evidence can reveal more about its direction than its most polished success story.
Will the arrangement remain usable under realistic change?
Consider changing demand, staff turnover, new technology and plausible disruption. This is not an invitation to invent endless disasters. It is a check on whether the system has enough adaptability, reserves and access to expertise to continue when familiar conditions shift.
Taken together, these questions discourage two easy mistakes. The first is declaring progress because something looks advanced. The second is declaring decline because one difficult result appears. Both conclusions can outrun the evidence.
A proper assessment also separates observations from explanations. “More staff are leaving” is an observation if measured appropriately. “They are leaving because standards have fallen” is a causal claim that needs further support. “The current programme caused the improvement” requires more than a before-and-after photograph of favourable numbers.
Where possible, compare relevant groups, examine alternative explanations and keep track of changes in measurement. If the population being assessed has changed, that matters. If the assessment has become easier or harder, that matters. If a new policy changes who can enter the system, a shift in the average may not mean the same thing as a change in individual capability.
We should also distinguish early signals from final outcomes. Better reporting may initially produce more recorded problems. A stronger intake of learners with additional support needs may change headline results. Such movements need interpretation before they are celebrated or condemned.
The framework is most useful when it produces a next action. A question about successors might lead to protected mentoring time. A question about access might reveal a timetable barrier. A question about retention might lead to a better later assessment. The purpose is not to produce an elegant diagnosis and stop there.
A direction becomes more credible when several kinds of evidence point towards the same conclusion: people are learning, capability is spreading, essential work is being renewed and the benefits are reaching the lives the system exists to support.
Previous chapter · Contents · Next chapter
26. Two Towns, Two Very Different Returns
Let us put the pieces together in a deliberately fictional example.
Two towns begin with similar resources, a working secondary school, several local employers and an essential public service maintained by a small technical team. Both receive funding for a new learning and technology initiative. Neither town is meant to represent a real country or community. The comparison is a thought experiment about decisions, not evidence that one policy package guarantees a particular outcome.
Town A begins with the visible purchase. New equipment arrives. A launch takes place. The programme is judged mainly by installation, participation and the number of completed activities. These measures are easy to report, and the first report looks impressive.
Teachers are asked to integrate the equipment, but little time is protected for preparation. Employers are invited to a meeting, but no continuing route is created for supervised student experience. The technical team gains another system to maintain without a clear plan for training or replacement parts.
None of the participants needs to be foolish or selfish for the arrangement to struggle. Each may be responding to the responsibilities and incentives placed in front of them. The problem is that the initiative’s parts have not been connected into a dependable learning system.
When difficulties appear, Town A responds by increasing the number of required activities. Completion becomes more urgent. Teachers provide more step-by-step help so students can meet the target. The technical team fixes problems after hours. The next report still looks respectable because the extra effort remains largely invisible.
A year later, the town has more equipment and a growing burden. Students may have benefited in some ways, but the programme has weak evidence about what they can now do. The people carrying it have less spare capacity for improvement. An initiative intended to create an upward spiral risks becoming a repeated demand for rescue.
Town B begins with a less photogenic exercise. It asks which learning outcomes matter, what students can currently do and which barriers prevent progress. Teachers identify places where technology might help and places where clearer instruction or better materials would be more useful. The technical team describes what it can maintain responsibly.
The town starts with a manageable scope. Learners receive instruction, opportunities to attempt the work and later checks of what they understand. Teachers compare student responses and revise explanations. Employers offer a small number of properly supervised experiences with clear expectations, rather than promising an unmanageable placement for everybody immediately.
Maintenance is included in the plan. So is the time required to teach new staff. Some activities are simplified because they add reporting burden without helping the objective. The programme’s success is not judged solely by how many things it contains.
When a result disappoints, Town B does not assume that the whole idea is worthless. It asks what happened. Perhaps the task was poorly matched to the learners’ starting point. Perhaps access was uneven. Perhaps the chosen tool was not suitable. A revision is tested, and an ineffective component can be discontinued without destroying the programme’s identity.
After a year, Town B may have fewer impressive photographs. What matters is whether it has better evidence of learning, a more manageable operating arrangement and people who are becoming able to continue the work. The next round begins with lessons that have survived rather than with a fresh demand to repeat the launch.
Even this comparison is incomplete. Town B could still face an external shock, lose funding or make a mistaken decision. Town A could recognise its difficulties and change direction. There is no permanent moral label attached to either town.
The point is the mechanism of return. Town A’s immediate success depends increasingly on hidden compensation. Town B tries to convert experience into reusable capability and to include renewal in the definition of success.
Now imagine asking a child in each town what the programme has meant. A complete evaluation would listen to that answer, too. Did the child become curious about something new? Did access improve? Was the work understandable? Did the programme make school feel more possible or merely add another demand?
A civilisational lens should not become so absorbed in systems that it forgets the person inside them. The town’s future is not separate from that child’s present experience. A worthwhile upward spiral connects the two: today’s opportunity develops capability while also making today a better time to be learning.
Previous chapter · Contents · Next chapter
27. What a Family Can Do Without Turning Home Into a Factory
A parent reading an essay about civilisation might reasonably think: this is a very large discussion, and I still need to help a child with homework tonight.
Fortunately, the household contribution need not be grand. It also need not become another exhausting project.
The useful starting point is to replace a vague judgement with a more precise conversation. Instead of “Why are you always careless?” try to discover what happened in this piece of work. Did the child misunderstand the question, lose track of a step, lack a prerequisite or rush because the task felt overwhelming? The answer may not be obvious immediately, and a teacher’s help may be needed.
This is not a demand that parents become subject specialists. It is a way of making the next conversation with the child or teacher more informative. A parent can notice that a child consistently struggles to begin a task without personally knowing the full solution.
Imagine a learner who says, “I understood it in class, but I cannot do it now.” That sentence contains a useful distinction. Understanding an explanation while it is present is not identical to being able to reconstruct and use it later. The response can include a small opportunity to explain what the learner does remember, followed by appropriate help at the missing point.
The goal is not to withhold help in the name of independence. It is to make help responsive. Sometimes an explanation is needed. Sometimes the child needs a question that reveals a connection. Sometimes the work is unsuitable for the moment and the sensible response is to communicate with the teacher rather than force completion through escalating distress.
A later check can be modest. After suitable teaching, can the learner explain the idea in their own words or attempt a related example? Can they identify the part they still do not understand? These observations can be more useful than insisting on a large volume of nearly identical work.
No particular home routine will suit every family or learner. Age, access needs, school demands, health and available support matter. The principle is to protect a relationship in which difficulties can be discussed and useful learning can become visible.
Parents can also help children notice knowledge in ordinary life. A bus journey raises questions about routes and timing. A recipe invites attention to quantities and sequence. A museum object opens questions about materials, history and interpretation. A news claim offers a chance to ask what evidence would support it.
These encounters do not all need to become lessons. Sometimes a journey should simply be a journey, and a meal should be enjoyed without an examination. Curiosity grows more naturally when the world is offered as interesting rather than converted into an endless test.
The adult’s example can matter as well. A child can see an adult check a fact, ask for expertise, repair a mistake or admit that an explanation was incomplete. This presents learning as an ordinary part of adulthood, not a temporary burden imposed on children.
There is particular value in showing that competence and uncertainty can coexist. An adult can be responsible without knowing everything. A child can be capable without answering immediately. Asking for appropriate help need not become a confession of inadequacy.
The family can also resist making comparison the entire meaning of education. Another child’s achievement may be informative, but it does not reveal this child’s whole starting point, circumstances or needs. The most useful question is often whether the learner is becoming more able to handle the next relevant task.
This does not mean ignoring examinations or opportunities. It means relating them to a larger purpose. A result can help identify what to repair, celebrate what has improved and plan what comes next. It should not become the sole language through which a child learns to describe themselves.
At home, the upward spiral may therefore look rather gentle. Confusion is noticed earlier. A small gap is repaired. An appropriate challenge becomes manageable. The learner gains evidence that understanding can be built. The next difficulty is approached with a little less fear and a little more knowledge about how to begin.
That is a meaningful contribution. Families do not have to carry the whole civilisation. They can help make one person’s relationship with learning more honest, humane and sustainable—and expect schools and public institutions to carry their responsibilities too.
Previous chapter · Contents · Next chapter
28. What Institutions Can Do Before the Crisis
Institutions have a different scale of responsibility. They control timetables, resources, staffing arrangements, incentives and the routes through which information becomes action. It would be unfair to ask individuals to compensate indefinitely for poor decisions at that level.
A useful institutional response begins by identifying the capability the organisation exists to provide. “Run more courses” is an activity. “Help learners acquire and use these important skills” is an objective. “Buy a new system” is an activity. “Make this service more dependable and accessible” is an objective.
The distinction does not make activities unimportant. It keeps them answerable to a purpose.
Next, examine the chain that connects the activity to the desired outcome. What must learners already understand? What do staff need? Which other institution controls a critical condition? Where does work become delayed, misunderstood or repeatedly repaired? A modest map of the actual process can be more useful than a grand declaration of transformation.
Then protect a manageable opportunity to improve it. A pilot should be large enough to encounter realistic conditions but bounded enough that problems can be studied without exposing the whole system to unnecessary risk. Not every intervention requires an experiment, but significant uncertainty should be acknowledged rather than disguised by an ambitious launch.
Evaluation should include both the desired gain and important costs. A faster process may create more downstream rework. A new reporting system may consume teaching time. A productivity improvement may depend on an unreasonable transfer of burden to a small group. These effects are part of the result, not irritating exceptions to it.
Institutions should also decide what would justify continuing, revising or stopping an initiative. Without such a decision rule, programmes can survive because ending them would be embarrassing rather than because they remain useful. Learning includes the ability to discontinue a method that does not earn its place.
Knowledge retention deserves an explicit home. When a team discovers an effective repair, where does that discovery go? Is the relevant procedure updated? Are people taught the change? Are its conditions recorded? Can someone later understand why the change was made? A folder full of reports is not enough if the work continues as before.
The people who perform this connecting work need time and authority. An institution cannot simultaneously demand careful learning and treat all reflection, mentoring and coordination as nonproductive overhead. Some of today’s resources must be assigned to the capability needed tomorrow.
That does not require an unlimited budget. It requires visible trade-offs. What is the consequence of postponing this maintenance? What is lost when mentoring is cancelled? Which administrative activity could be simplified to release time? Which apparent saving is simply a future obligation moved out of view?
A particularly useful question is whether the organisation can explain its dependence on exceptional people. Their contributions should be recognised, but dependence should not become a permanent substitute for developing successors and reliable systems. An expert’s greatest long-term contribution may include helping others become less dependent on that expert.
Public institutions have the additional responsibility of considering access and distribution. A programme can work well for people who already know how to enter it while leaving others behind. Information, language, disability, location, cost and timing can all affect whether a nominal opportunity becomes a real one. The relevant barriers must be investigated in context.
The institution should remain open to the experience of those it serves. Students, families, workers and service users can reveal problems that a central dashboard misses. Their accounts need not be treated as complete causal explanations to be treated as important evidence.
An upward spiral at this level is therefore a disciplined relationship between purpose, observation, action and renewal. It is not a succession of increasingly exciting announcements. It is a system becoming better at doing worthwhile work, noticing its limits and preparing people to continue.
The time to build that relationship is not only after failure. A functioning institution has a precious opportunity: it can use present stability to strengthen the next handover. Success is most valuable when it creates the room to prevent the crisis that would otherwise consume that room later.
Previous chapter · Contents · Next chapter
29. The Questions That Keep This Essay Honest
A good account of education and civilisation should survive several objections.
The first is straightforward: can people be well educated and still use their abilities badly? They can, in principle, use knowledge for purposes that are manipulative, exploitative or destructive. Nothing in the ability to calculate, organise or persuade logically guarantees a humane objective. This is why the essay treats the direction of progress as an ethical question as well as a technical one.
A second objection asks whether an improving school system guarantees an improving society. It does not. Education can expand capability while other conditions obstruct its use. Conflict, exclusion, weak institutions or a lack of suitable opportunities can interrupt the route between learning and a better life. The research discussed earlier offers evidence about specific parts of that route, not a universal promise.
A third objection concerns people with little formal schooling. Must they be regarded as less capable, less wise or less valuable? Absolutely not. Formal education is one important route through which knowledge is developed and recognised. It is not an exhaustive measure of practical judgement, care, creativity or the ability to learn. The model would contradict itself if it confused a credential with the whole person.
A fourth objection asks whether the spiral is merely a complicated way of saying “good things lead to good things.” It would be, if we stopped there. Its usefulness lies in specifying the connections: what is learned, how it becomes usable, who can access it, how results are evaluated, where benefits and costs go, and whether the conditions of the next round change. A missing connection gives us something to investigate.
A fifth objection concerns measurement. Can we really observe all these relationships? Not completely. Some outcomes take years to appear, some are difficult to compare, and some are obscured by simultaneous changes. That limitation requires modest claims, mixed evidence and continued inquiry. It does not make every interpretation equally good.
A sixth objection asks whether repeated improvement must eventually become self-consuming. No such inevitability is claimed here. Growth can introduce new coordination and maintenance demands, but people can recognise and respond to them. The point of examining those mechanisms is precisely to make adaptation more possible before decline becomes severe.
A seventh objection concerns historical collapse. Can this essay explain why every empire, state or civilisation fell? No single educational mechanism could responsibly do that. The terms themselves refer to different kinds of historical continuity and political change. A regime can end while knowledge survives elsewhere. A city can lose prominence while people adapt. A headline story of “rise and fall” may combine several different processes.
This essay therefore offers a lens for examining renewal and depletion, not a master key that replaces historical investigation. Applying it to a particular case would require careful evidence about that case: institutions, material conditions, population, conflict, environment, exchange and the relevant periods of change.
An eighth objection asks whether technology might eventually perform so much that human education becomes unnecessary. The answer depends on the tasks, the reliability of the arrangement and the purposes we want human beings to retain. Even when technology performs an activity well, people may still value understanding, choice, participation and the capacity to govern their relationship with it. Education is not justified only by tasks machines cannot yet perform.
A ninth objection is more personal: does this framework turn life into a permanent audit? It should not. The purpose of a useful framework is to clarify where attention is needed, not to make every meal, friendship and afternoon accountable to a performance indicator. People need room to live without constantly proving that their existence is an investment.
The final objection is perhaps the most important. Does talking about civilisation make the individual disappear? It can, unless we take care. A system-level improvement is not enough if its benefits require treating some people as expendable. The ordinary life is not a minor detail beneath the civilisational story. It is where that story becomes real.
These objections do not dismantle the spiral. They give it boundaries. Within those boundaries, it remains a useful question: what does this arrangement carry forward, and does that inheritance make worthwhile life more possible?
A model becomes stronger when it helps us see what would challenge it. Education should do the same for the ideas it teaches—including this one.
Previous chapter · Contents · Next chapter
30. The Next Generation Should Inherit More Than Our Buildings
Return to the child walking through the city in the morning.
The station is open. The familiar bridge is still there. Somebody has prepared a classroom. Somewhere, a person who understands a complicated piece of work is beginning to show another person how to do it.
These are ordinary scenes. Together, they contain a large part of civilisation’s future.
The child will inherit objects, institutions, problems and possibilities. Some will be wonderful. Some will be confusing. Some will need repair. The most generous inheritance is not a world in which every decision has already been made, but a world in which the child can acquire the means to understand and improve what has been received.
That is why education matters beyond the next examination. It allows experience to become available to people who were not present when the experience occurred. It gives a new generation ways to recognise a pattern, question a claim, use a tool, care for a system and imagine an alternative.
But the connection runs in both directions. The world adults create affects how that education can happen. A classroom needs more than good intentions. A teacher needs more than praise. A learner needs more than a message about the importance of working hard. Institutions, resources, access, care and opportunity have to make those intentions usable.
The upward spiral appears when these relationships support one another. Learning develops capability. Capability contributes to useful work and richer lives. Some of the resulting benefits strengthen access, teaching, maintenance and renewal. Knowledge is tested, retained and made available to others. The next round begins with more than the previous round consumed.
The downward spiral appears when those connections weaken. A gap is misdiagnosed. A warning is silenced. A route into learning is closed. A success is maintained by exhausting its contributors. A tool supplies answers while the surrounding system forgets to ask what people are learning. The costs return later, often to people who did not choose to create them.
Neither direction belongs permanently to a particular people or place. A society can carry both at once. An institution can change direction. A learner can find a better starting point after a long period of difficulty. Progress is neither automatic nor forbidden.
There is comfort in that uncertainty, if we read it correctly. We are not guaranteed a better future, but neither are we trapped inside a story in which decline must arrive on schedule. The quality of the next turn is partly shaped by decisions that can still be made.
Some are public decisions about access, institutions and the distribution of resources. Some are organisational decisions about training, evidence and maintenance. Some are ordinary decisions about how we respond when another person says they do not understand.
We should not pretend these decisions have equal power. A kind conversation cannot replace a functioning education system, and a diligent teacher cannot individually repair every surrounding injustice. Responsibility must remain proportionate to the authority and resources people possess.
Yet small acts can still belong to a larger pattern. A clear explanation can make a subject less frightening. A protected apprenticeship can keep a useful skill from disappearing. An honest report can prevent a repeated failure. A well-supported return to study can reopen a route that once seemed closed.
There is joy in these beginnings. Not the forced cheerfulness that refuses to look at difficulty, but the more durable pleasure of discovering that something can be understood, shared or repaired. A problem that had seemed like a wall becomes a door with a difficult but visible handle.
Perhaps this is one of civilisation’s most worthwhile forms of abundance: more people encountering a world they can participate in, rather than a world they must merely endure or admire from outside.
We can appreciate the monument and still ask who will care for it. We can welcome the machine and still ask what its users need to learn. We can celebrate wealth and still ask what possibilities it opens. We can honour the past without asking the future to remain identical to it.
The next generation should inherit more than our buildings. It should inherit the knowledge to maintain what is valuable, the freedom to question what is not, and the support to build something we have not yet imagined.
An upward spiral begins whenever today’s learning becomes a better place for someone else to begin.
And education, at its best, is how we keep offering that beginning.
Previous chapter · Contents · Sources
Sources and reading notes
Research and institutional information checked on 12 October 2026. Publication dates and study periods are stated separately. The empirical sources support particular claims and examples; they do not establish the essay’s overall spiral framework as a universal law. Imagined people, organisations and towns are explanatory examples rather than observations or testimonials.
1. World Development Report 2018: Learning to Realize Education’s Promise
World Bank. 2018. Read the source. Used for the distinction between educational participation and actual learning, and for the importance of the surrounding education system. This is a synthesis, not a causal estimate for the essay’s spiral model. Return to the discussion.
2. How People Learn II: Learners, Contexts, and Cultures — Chapter 5, Knowledge and Reasoning
National Academies of Sciences, Engineering, and Medicine. 2018. Read the source. Used for the role of prior knowledge and its possible benefits and biases. The book’s online URL labels this chapter differently from its printed chapter number; the cited chapter is Knowledge and Reasoning. Return to the discussion.
3. Schooling and Labor Market Consequences of School Construction in Indonesia: Evidence from an Unusual Policy Experiment
Esther Duflo. 2001. Read the source. American Economic Review, 91(4), 795–813. The regional and cohort-based study concerns a specific Indonesian programme, not education as a sufficient explanation of national development. Return to the discussion.
4. The Medium Run Effects of Educational Expansion: Evidence from a Large School Construction Program in Indonesia
Esther Duflo. 2002; published version 2004. Read the source. NBER Working Paper 8710; subsequently published in Journal of Development Economics, 74(1), 163–197. Used to qualify the assumption that labour-market adjustments and the benefits of expansion are immediate and uniform. Return to the discussion.
5. Generative AI at Work
Erik Brynjolfsson, Danielle Li and Lindsey Raymond. 2025. Read the source. Quarterly Journal of Economics, 140(2), 889–942. This essay uses the published study’s 5,172-agent sample and 15% average productivity estimate, not the different figures reported in an earlier working-paper version. Return to the discussion.
6. Generative AI without guardrails can harm learning: Evidence from high school mathematics
Hamsa Bastani and colleagues. 2025. Read the source. PNAS. The experiment took place in autumn 2023 at one Turkish high school. The guarded tutor mitigated the negative unassisted-test effect; the study did not establish an unassisted-test gain for that tutor. Its short-term findings should not be generalised to every subsequent model or design. Return to the discussion.
7. Technology and Child Development: Evidence from the One Laptop per Child Program
Julián Cristia, Pablo Ibarrarán, Santiago Cueto, Ana Santiago and Eugenio Severín. 2012. Read the source. Inter-American Development Bank working paper. The fifteen-month, 319-school evaluation reported mixed outcomes: greater computer use, no detected Mathematics or language-test gains, and gains on some general cognitive measures. Return to the discussion.
8. About ITE
Institute of Technical Education, Singapore. Institutional overview. Read the source. Used for ITE’s establishment in 1992 and its pre-employment, continuing and workplace-related educational roles. It does not establish the causal contribution of ITE alone to Singapore’s prosperity. Return to the discussion.
9. Port of the Future
Maritime and Port Authority of Singapore. Institutional page; updated 6 August 2026. Read the source. Used for the described Tuas Port automation design, including electrified automated yard cranes, automated guided vehicles and remote management. Future plans on the page are not treated as completed operating capacity. Return to the discussion.
10. About the Museum
Kuboso Memorial Museum of Arts, Izumi. Institutional history. Read the source. Used for the 1977 donation of art, building, land and funds, the museum’s opening in 1982, and later expansion through donations. The wider discussion of public inheritance is the essay’s interpretation. Return to the discussion.
11. Adult skills in literacy and numeracy declining or stagnating in most OECD countries
Organisation for Economic Co-operation and Development. 10 December 2024. Read the source. Release concerning the 2023 Survey of Adult Skills. Country and group differences matter; the results are not used to claim that every affluent society is declining or to establish a single cause. Return to the discussion.
12. The world needs 31 million more secondary teachers: new guidance out now to fill the gap
UNESCO International Institute for Educational Planning. 4 October 2024; updated 2 October 2025. Read the source. The cited global teacher requirement concerns achieving universal primary and secondary education by 2030. It includes replacement needs and should not be read as a current vacancy count. Return to the discussion.
13. UNESCO and UNICEF urge action to protect the right to education in Afghanistan
UNESCO and UNICEF. 24 January 2026. Read the source. A dated joint statement used for educational exclusion and its implications for future teachers and health workers. No historical count in the statement is presented as an October 2026 total. Return to the discussion.
14. Nearly a quarter of a billion children’s schooling was disrupted by climate crises in 2024
UNICEF. 24 January 2025. Read the source. The 242-million estimate covers school closures or significant timetable interruptions in 2024. It does not mean that all affected students left school permanently or experienced the same duration of disruption. Return to the discussion.
15. Teaching at the Right Level to accelerate learning
Abdul Latif Jameel Poverty Action Lab. Evidence synthesis. Read the source. Used for Pratham’s assessed-level approach, its evaluation history and the importance of implementation support. The evidence does not validate every form of grouping or any unrelated programme using similar language. Return to the discussion.
Continue with the sibling essay Education and Civilisation | Drift for how learning, support and succession can quietly separate before a downward spiral becomes obvious. The foundational drift article defines that separation, while the CivOS Drift Map locates its wider consequences.