- A · Historical fact
- B · Empirical evidence
- C · Philosophical proposition
- D · Future scenario
The Trivium taught generations how to think and communicate. The Quadrivium taught them how to understand mathematical order. Neither was designed for a world in which machines read, write, calculate and reason. This manifesto argues that such a world calls not for less education but for a different idea of what education is for.
The origins of liberal education
Greek paideia meant the formation of a whole person for life in the city. Plato's Republic describes a higher curriculum of arithmetic, geometry, astronomy and harmonics leading to dialectic [1]. Aristotle argued that citizens should be educated for the noble use of leisure, scholē, the word from which "school" descends, and not only for useful occupations [2].
Rome named these studies the artes liberales: the arts fit for a free person. Even then they had critics. Seneca asked whether such studies made anyone good, and answered that they only prepared the mind for philosophy [4].
The tradition began with an honest contradiction. It aimed at freedom of mind, yet it was reserved for those freed from labour by the work of others. A post-labor society could, for the first time, offer that leisure to everyone. Whether it will is a political question, not a technical one.
The Trivium and the Quadrivium
In the fifth century Martianus Capella fixed the canonical list of seven liberal arts in an allegory that medieval schools copied for centuries [7]. Boethius used the word quadrivium, "four ways", for the mathematical arts of arithmetic, geometry, music and astronomy [6]. The verbal arts of grammar, rhetoric and logic came to be called the trivium. In the medieval university they formed the arts course that prepared students for the higher faculties of theology, law and medicine [9].
The list was never complete. Varro's earlier scheme had nine disciplines, including medicine and architecture [7]. Hugh of St Victor, around 1120–1130, added seven mechanical arts, from agriculture and navigation to medicine and theatre [8].
The genius of the scheme was not its list but its order: first the arts of language and reason that make all other learning possible, then the study of mathematical structure, then the great questions. We keep that order of dependence and abandon the idea that seven boxes can hold human knowledge.
Renaissance interdisciplinary knowledge
Fifteenth-century humanists built a new curriculum, the studia humanitatis: grammar, rhetoric, poetry, history and moral philosophy, joined in schools such as Vittorino da Feltre's with physical exercise and the formation of character [13]. The nineteenth-century historian Jacob Burckhardt made the "universal man", the uomo universale, the emblem of the age, with Leon Battista Alberti as his example [14].
Leonardo da Vinci, who called himself a man without (Latin) letters, joined painting, anatomy, engineering and the study of water and flight through drawing and observation [15].
Leonardo is an illustration, not a template. What he shows is that curiosity crosses disciplinary borders when it is driven by questions rather than by credentials, and that the hand can be an instrument of thought. Most of his notebooks went unpublished; interdisciplinary genius without institutions to receive it changes little.
Industrial education and specialisation
Mass schooling grew with the modern state and the industrial economy: Prussian school regulations from 1763 [10], compulsory attendance in Massachusetts from 1852 [11], England's Elementary Education Act of 1870 [12]. Universities split into specialised disciplines and professional schools. In 1918 Franklin Bobbitt proposed deriving the curriculum from the activities of adult working and civic life, applying the methods of scientific management to schooling [20].
Others resisted. Newman defended knowledge as its own end [18]; Dewey rejected the split between liberal and vocational education as a residue of class division [19].
It would be false to say that modern education became merely vocational. It produced mass literacy, the research university and enormous scientific progress. But its justification shifted. In public argument, education is defended above all by employability, credentials and economic competitiveness. That justification is about to be tested.
Artificial intelligence and cognitive acceleration
METR, an independent research organisation, measures the length of software and research tasks, timed by skilled humans, that AI agents can complete half of the time. In its November 2026 update it estimates that this time horizon is doubling about every 3.6 months (95% confidence interval 2.1 to 4.9 months) [43]. Models now often match or beat skilled humans on coding and optimisation tasks with clear feedback, but still show weaker judgment than experts where feedback is slow or ambiguous. A broader composite index shows no obvious general acceleration.
A task horizon is not general intelligence, and a benchmark is not a profession. But the direction is clear enough to plan for: work that consists of producing well-specified cognitive output is becoming automatable, and what remains distinctively valuable is judgment where feedback is poor, responsibility for outcomes and the ability to verify.
The possible decoupling of labour and production
This project adopts a working planning assumption: a 90% project-level confidence that advanced AI will substantially automate cognitive work and progressively challenge the necessity of human labour across much of the economy. This is our assumption for planning, not a probability any researcher has published, and the framework is built to remain useful if the transition is slower or uneven.
David Shapiro's post-labor economics names this decoupling of economic output from human labour and insists that post-labor does not mean post-scarcity: energy, land, materials and logistics stay constrained, and ownership of productive systems becomes decisive [58][56].
Economists remain divided. Autor shows how automation has historically complemented labour [53]; Acemoglu and Restrepo find that in recent US decades displacement has outpaced the creation of new tasks [54].
The problem of technological power
Every previous expansion of human power, from fire to fission, raised the stakes of human judgment. AI is different in degree and perhaps in kind: it amplifies the capacity to act in every domain at once, including the domain of thought itself. Recent scholarship names this the power–wisdom gap: the central educational challenge is shifting from a scarcity of knowledge to an abundance of power [42]. Bainbridge's "ironies of automation" add a sharper warning: the more we automate, the more we depend on human supervisors whose skills automation erodes [49].
The need for renewed human formation
If education was justified by work, the end of work's necessity can be read in two ways: as the end of education's purpose, or as the return of its original one. We take the second view. Pieper called leisure the basis of culture [21]; Arendt warned of a society of labourers about to be freed from labour, which no longer knows the higher activities for whose sake the freedom would be worth winning [23].
The warning is not only philosophical. When mass unemployment came to the Austrian village of Marienthal in the 1930s, researchers found not revolt but resignation and the collapse of daily structure [39]. Freedom from work without income, purpose or community is not liberation.
A new knowledge architecture
We therefore propose an expandable architecture of human knowledge and development:
- Ten domains of knowledge, each defined by a fundamental question, from LOGOS (meaning) to EPISTEME (knowing), each with a traceable ancestor in the historical curriculum. The domains
- Cross-domain fields such as engineering, medicine and education that span several domains rather than forming one. The fields
- Ten transversal capabilities, from independent judgment to lifelong learning, with AI fluency running through all of them. The capabilities
- One principle: the growth of technological power must be matched by the growth of human wisdom and institutional capacity. The power–wisdom gap
It is not a new list of school subjects. It is a map of what human beings can understand, create and become, and an invitation to argue about it. The number ten is not a dogma.
The next intellectual project must integrate everything humanity has learned, built, discovered and created, and ask what we should become next.