Ask what makes a mind good and every civilisation gives a plural answer. Aristotle named five intellectual virtues. Confucian China built the largest pre-modern meritocratic institution around moral character. Buddhist traditions refused to separate wisdom from compassion. Yoruba thought distinguished head-wisdom from belly-wisdom centuries before the West coined "emotional intelligence." Then Galton, Binet, and Spearman compressed all of it into a single number. This book reconstructs what was lost, maps the dimensions the traditions were actually tracking, and proposes a framework to replace the false unity of IQ — not with a new single score, but with something adequate to what humanity has actually valued.
It is a modern English-language compression of a much larger, older, and more plural human project. That project, pursued across civilizations and centuries, has been concerned not just with reasoning or learning speed but with judgment, character, emotional attunement, social competence, practical mastery, ecological skill, and wise conduct. Reducing all of that to a single word — and then building schools, hiring systems, AI benchmarks, and self-concepts around that word — has consequences.
The book makes a single argument in three movements, and each movement earns the next. The history establishes that the narrowing happened and was a choice. The map recovers what was lost and shows that the measurement gap is structural, not accidental. The replacement proposes what comes next — not a new single score but a framework adequate to what humanity has actually tracked, cultivated, and valued across civilisations and centuries.
Nineteen chapters reconstruct humanity's long, branching attempt to define the best kind of mind. Ancient Greece's five intellectual virtues. Rome's orator ideal. China's thirteen-century examination system built around moral-intellectual formation. India's five-step syllogism using a fundamentally different logical architecture. Buddhist insistence that wisdom without compassion is not yet wisdom. Yoruba distinction between head-wisdom and belly-wisdom. Ubuntu's relational challenge to the individual as the unit of analysis. Pacific navigators holding star compasses in memory across open ocean. Then the narrowing: Galton, Binet, Spearman, and the compression of all of this into a single measurable quantity. The book concedes what IQ gets right — cognitive abilities are real, measurable, and predictive — before arguing that the incompleteness is structural, not marginal, and excludes the dimensions most human traditions placed at the centre.
Thirteen chapters map twelve dimensions of human excellence — reasoning, learning, knowledge, practical judgment, emotional attunement, social-relational competence, self-regulation, creativity, wisdom, moral character, embodied skill, and collective capability — across traditions and modern measurement. The core finding: the dimensions psychometrics measures best (reasoning, knowledge, processing speed) are the ones most traditions treated as one component among many. The dimensions most traditions placed at the centre of excellence (practical judgment, wisdom, moral character, relational competence) are the ones modern psychology measures least well or not at all. This is not a technical gap. It is a structural consequence of building a measurement system around institutional convenience rather than around the full range of what humanity has valued.
"Intelligence" is not a single natural kind. It is a partly real, badly bundled construct — evidence from within psychometrics (the mutualism model), from across civilisations (convergent phenomena, divergent bundling), and from the measurement fragmentation documented throughout Movement II. The replacement is a twelve-dimensional framework applied to the domains where it matters most: assessment across cultures (where broadening the conception of excellence also broadens inclusion), human-AI comparison (where asking "is this AI smarter?" recapitulates the false unity), education (where the traditions provide existence proofs), institutional selection (where 74% of job performance variance is unexplained by IQ), and AI alignment (where dimensional profiles produce safer deployment than single benchmarks).
Every tradition was trying to answer some version of the question: what does it mean for a human being to think well, judge well, know well, and live well? Every one arrived at answers broader, more plural, and more entangled with ethics, emotion, and social life than the modern word "intelligence" suggests. Five civilisations independently developed ideals that integrate cognition with moral character. Three independently developed formal logical systems. Eight or more refused to separate intellectual excellence from ethical formation.
Five intellectual virtues: nous (intuitive reason), epistēmē (scientific understanding), technē (craft), sophia (theoretical wisdom), phronēsis (practical wisdom). Excelling at one does not guarantee the others.
Civic judgment, eloquence, and the educated person. Excellence defined through public service, rhetorical mastery, and the capacity to govern wisely — the orator as the ideal mind.
Xin (心, heart-mind) deliberately refuses to separate thinking from feeling. The imperial examination — the largest pre-modern meritocratic institution — tested literary mastery as a proxy for moral-intellectual formation.
Not "how intelligent is this person?" but "what makes a cognition valid?" The Nyāya five-step syllogism uses a fundamentally different architecture from Aristotle's. Emphasis on sources of knowledge (pramāṇas).
Prajñā (wisdom) paired with karuṇā (compassion). Wisdom without compassion is not yet wisdom. The Geshe degree: years of study, memorisation, public debate — reasoning inseparable from ethical transformation.
ʿAql (intellect) distinguished from ḥikma (wisdom). The madrasa system combined textual scholarship with jurisprudential reasoning, debate, and ethical development across centuries.
Wisdom as the ability to make good use of what one knows. Yoruba distinguishes intellectual wisdom (ọgbọ́n-ori) from emotional wisdom (ọgbọ́n-inú) — anticipating emotional intelligence debates by centuries. Character (omolúàbí) as the paramount quality.
Not "how smart is this individual?" but "how fully has this person realised their humanity through relationships with others?" The individual extracted from relational context is not a meaningful unit of analysis.
Pacific navigators hold star compasses in memory across weeks of open-ocean voyaging. Māori wānanga as knowledge transmission. Inuit ecological expertise. Culturally grounded assessment that no standardised battery captures.
Breadth beyond the major cases. The Nahua tlamatini (knower of things). Japanese concepts of mastery across martial, aesthetic, and contemplative traditions. Traditions that further confirm the pattern of plural excellence.
What counts as knowledge when there is no text? Proverb competence, elder judgment, demonstrated performance, apprenticeship. The politics of which knowledge systems get recognised and which get erased.
Method, reason, certainty — then Galton's anthropometric laboratory, Binet's school test, Spearman's g, the Army Alpha and Beta. A rich plurality compressed into a single measurable quantity to meet institutional needs: fast, cheap, scalable, defensibly objective.
Each chapter in Movement II takes one dimension and asks three questions: how has it been understood historically, how is it currently measured, and where is frontier science heading? The pattern that emerges is the core finding of the book: the dimensions the psychometric tradition measures best are the ones most traditions treated as one component among many. The dimensions most traditions placed at the centre of excellence are the ones modern psychology measures least well or not at all. This is not a technical gap awaiting a better test. It is a structural consequence of the narrowing.
Pattern recognition, logical inference, mathematical thinking, spatial reasoning. The core of what IQ tests capture. Tracked by every tradition but treated as one component among many by most. Aristotle's nous and epistēmē. The Nyāya inference engine. Spearman's g. Real, robust, replicable — and insufficient as a description of the whole.
Retained information, domain expertise, crystallised understanding. Pacific navigators holding star compasses in memory. The Confucian scholar's classical mastery. Measurable, but conflated with reasoning in composite IQ scores — the distinction between knowing things and reasoning about things is real and most traditions kept it separate.
Speed of acquisition, transfer across domains, responsiveness to instruction. What Vygotsky's zone of proximal development captures that static testing misses. Dynamic assessment — testing what someone can learn with help, not just what they already know — has strong theoretical support but limited adoption.
Perceiving, understanding, and regulating emotion in oneself and others. Yoruba ọgbọ́n-inú (emotional wisdom). Buddhist mindfulness traditions. The Western tradition was slowest to take this seriously — the concept of "emotional intelligence" arrived only in the 1990s, centuries after non-Western traditions had been cultivating it systematically.
Executive control, impulse management, sustained attention, delayed gratification. The dimension that predicts life outcomes as well as or better than IQ. Buddhist contemplative training, Confucian self-cultivation, and Stoic askēsis are all, in modern terms, systematic self-regulation training programmes that predate psychology by millennia.
Good decisions in ambiguous, real-world situations where rules underdetermine action. What Aristotle called phronēsis. What every human tradition placed near the centre of excellence. What IQ tests do not measure. Situational Judgment Tests (SJTs) have modest validity (r=.26) and crucially produce lower adverse impact than cognitive tests — broadening the conception of excellence also broadens inclusion.
Navigating relationships, building trust, reading social contexts, maintaining reciprocal obligations. Ubuntu personhood-through-others. Confucian relational ethics. The five relationships. Central to every tradition except the psychometric one. The Woolley/Malone finding: group performance predicted by average social sensitivity, not average IQ.
Novel, valuable production. Divergent thinking, recombination, the capacity to generate what did not previously exist. The Japanese aesthetic traditions. The Nahua tlamatini. Torrance Tests have been the standard for decades but capture only a fraction of what creative production actually involves. Product-based assessment is more valid but harder to scale.
Long-term perspective, tolerance of uncertainty, integration of knowledge with values, recognition of the limits of one's own knowledge. What five civilisations independently placed at the summit of human excellence. Aristotle's sophia. Buddhist prajñā. Confucian zhì. The Berlin Wisdom Paradigm is the most developed measurement approach but has not achieved widespread adoption. No standardised test captures this.
Reliable moral judgment, integrity, trustworthiness under pressure, the disposition to act rightly when no one is watching. The Yoruba omolúàbí. The Confucian junzi. The dimension most traditions refused to separate from intellectual excellence. Integrity tests exist but are contested. The gap between what traditions placed at the centre and what modern testing captures is widest here.
Sensorimotor mastery, environmental attunement, skill in material interaction with the physical world. Pacific wayfinding across thousands of miles without instruments. Surgical precision. Inuit ecological expertise calibrated across lifetimes. The dimension the desk-based testing paradigm structurally excludes by design.
Group-level performance that exceeds individual contributions and cannot be reduced to the sum of individual scores. The Woolley/Malone c-factor finding: group performance predicted not by the average or maximum IQ of members but by average social sensitivity and equality of conversational turn-taking. Ubuntu's challenge to the unit of analysis itself — the individual extracted from relational context may not be a meaningful unit. The most effective teams are not composed of the individually smartest people. They are composed of people who listen, take turns, and read each other. This dimension challenges whether "human excellence" should be measured at the individual level at all.
What most traditions placed at the centre of excellence is what modern psychology measures least well or not at all. This is not an accident. It is the structural consequence of building a measurement system around institutional convenience rather than around the full range of what humanity has valued.
Real capacities. Genuinely predictive within the systems designed around them. But treated by most traditions as one component among many — necessary but nowhere near sufficient for a good mind. Aristotle's epistēmē and nous without phronēsis. The Confucian scholar who can recite the classics but lacks moral character.
The dimensions most traditions placed at the centre: practical judgment, moral discernment, emotional attunement, relational competence, wisdom, collective capability. These are not secondary. For leadership, governance, parenting, friendship, and most of what constitutes a human life, they matter more than analytical reasoning alone. The measurement gap is not a technical limitation awaiting a better test. It is a structural consequence of the narrowing.
The book does not argue, as Gould did, that g is a fraud. It argues that g is a summary, and summaries leave things out. The question is how much they leave out, and whether what they leave out matters. The answer, across every tradition and every measurement system surveyed, is: they leave out most of what humanity has valued, and what they leave out matters enormously.
Three independent lines of evidence converge on this conclusion. From within psychometrics: the mutualism model offers an alternative explanation of the positive manifold (the fact that cognitive tests correlate positively) that does not require a single underlying factor — cognitive abilities may correlate because they develop together through mutual facilitation, not because they share a common cause. From across civilisations: convergent phenomena, divergent bundling — the traditions tracked the same capacities but bundled them differently, which is what you expect when the capacities are real but the bundling is cultural. From the measurement fragmentation documented throughout Movement II: the dimensions the traditions valued most are the ones where measurement is weakest, and this pattern is too systematic to be accidental.
The analogy: BMI is a real number from real measurements that correlates with health outcomes. No serious physician treats it as a complete description of health, because it does not distinguish muscle from fat, ignores cardiovascular fitness, and says nothing about mental wellbeing. IQ and g are compressions of the same kind. They summarise the shared variance among a set of cognitive tests. That shared variance is real and worth studying. But it is not the only variance, and it is not the whole person.
Not a new single score but a set of partly independent axes. Different traditions tracked them. Different measurement systems assess them. Different contexts weight them differently. The twelve dimensions proposed, with their separabilities and clusters mapped. Cognitive power and emotional attunement: clearly separable (r ≈ .20-.30). Practical judgment and wisdom: natural cluster. Self-regulation and moral reliability: related but distinguishable.
What should be measured separately, and how. A selection system that includes Situational Judgment Tests alongside cognitive tests produces a more diverse candidate pool without sacrificing predictive validity. Broadening the conception of excellence and broadening inclusion are not competing goals — they are aspects of the same correction. Kaupapa Māori and Indigenous-led assessment as existence proofs.
Asking "is this AI more intelligent than a human?" recapitulates the false unity. The dimensional framework makes it impossible to answer with a single yes or no — and that is a feature, not a bug. An AI system that exceeds human performance on reasoning while being absent on judgment, character, and wisdom is not "smarter." It has a different profile. The comparison should be dimensional, not scalar.
Not abandoning cognitive skills but situating them within a broader framework. The Confucian examination system, Buddhist contemplative training, and apprenticeship traditions are existence proofs that judgment, self-regulation, moral character, and relational competence can be systematically cultivated. Dynamic assessment as a fairer complement to static testing. The trade-offs are real: plural assessment is harder, more expensive, and less scalable.
Cognitive ability accounts for ~26% of job performance variance. The other 74% — judgment, social competence, self-regulation, moral reliability — collectively matters nearly three times as much. Cognitive tests produce the largest adverse impact of any major selection method. Different roles require different dimensional profiles. The practice of promoting the most technically able person into leadership, then being surprised when they struggle, is a predictable consequence of single-axis selection.
AI systems evaluated on dimensional profiles are deployed more safely than systems evaluated on single benchmarks. The framework makes it impossible to claim that a system is "generally intelligent" on the basis of high reasoning while being absent on judgment, character, and wisdom. It forces the question the single-axis model evades: does this system possess the specific capabilities and dispositions required for this specific deployment context?
The book concedes what IQ gets right — cognitive abilities are real, measurable, and predictive — before arguing that the incompleteness is structural, not marginal. It is not a new single score. It is a set of partly independent axes that different traditions have tracked, different measurement systems assess, and different contexts weight differently. Forty-one chapters across history, cartography, and reconstruction. Early draft, open for review.
"Humanity has never had one concept here. What we have is a plural field of capacities — cognitive, regulatory, relational, creative, moral, embodied, and collective — that different civilisations have tracked, cultivated, bundled, and validated in different ways. The task ahead is not to find the one true definition of intelligence but to build frameworks worthy of the full range of human excellence." — The Varieties of Human Excellence, Early Draft 2026