Small children are question-generating machines; the folk observation and the research literature agree on the phenomenon if not on the exact rate. Then formal schooling begins, and the flow reverses. The institution asks; the child answers; the questions on the test are the only questions with standing. By adolescence, the machine has largely shut down, not because curiosity died but because the market for it closed. The curriculum is a supply-side artifact: answers manufactured in advance, distributed on a schedule, to customers who never placed the order.
The alternative has a precise name: demand-driven learning, in which the child's actual question is the unit of curricular organization, and the adult's job is to receive it, take it seriously, and engineer it into work worth doing. This is routinely caricatured as letting children study whatever they like, which misses the design entirely. A question is not a preference. "Why do bridges not fall down?" is a question that, honestly pursued, drags load-bearing mathematics, materials, history, and drawing behind it. The craft is in the dragging.
What the motivation evidence says
The empirical case for the inversion rests on the best-replicated body of findings in educational psychology. A meta-analysis spanning 144 samples and 79,079 students found that autonomy-supportive teaching, taking the learner's perspective, offering choice, providing rationale, is among the strongest predictors of intrinsic motivation, well ahead of parental influence (Bureau et al., 2021). A companion synthesis of 153 studies ties autonomy support to engagement, mastery goals, self-regulated learning, and achievement (Mammadov and Schroeder, 2023). And intervention studies show the lever is causal-adjacent, not merely correlational: 36 experiments based on self-determination theory, nearly 12,000 participants, moved intrinsic motivation and the needs beneath it (Wang et al., 2024). A question the child owns is autonomy in its most concentrated form: the topic, the reason, and the finish line all originate in the learner.
Schools that run on questions
None of this is hypothetical. The Escola da Ponte in Portugal, the model that has most influenced Brazilian experiments, abolished both grade-years and subject divisions; students organize around projects of their own choosing and, crucially, indicate themselves when they are ready to be assessed, replacing the scheduled exam with a declared one (Marchelli et al., 2008). Public schools in São Paulo, Amorim Lima and Campos Salles among them, imported the architecture into ordinary state systems, evidence that question-driven structure is not a private-school luxury.
Ricardo Semler's schools ran the same inversion for two decades. As he has told it: some 3,000 children across six countries, choosing what to learn, with adults as tutors rather than bosses, no rows, no bells, no grades for memorization, everything organized as a project, a tutor, and a deadline, "which is the shape of work too." Even the institutional mainstream has conceded the direction of travel: the OECD's Learning Compass framework places student agency, the capacity to set goals and act on them, at the center of its 2030 vision, with "co-agency" between students and adults as the operating relationship (OECD, 2019). When the club of industrial-era education ministries adopts the vocabulary of the free schools, the argument has moved from whether to how.
The adult does not disappear
The strongest objection deserves its due: children do not know what they do not know, and a curriculum of pure demand would orbit dinosaurs and video games. The objection defeats a version nobody serious proposes. In every working implementation, the adult remains decisive, not as the source of the questions but as the metallurgist of them: sensing when a question can bear rigor, connecting it to the disciplines it secretly contains, refusing to let it end in a poster when it could end in understanding. Demand-driven learning is more mediated than the lecture, not less. The lecture requires knowing the material; the tutor's craft requires knowing the material and the child.
The industrial curriculum made one great bet: that answers delivered before the question arrives will be there waiting when it does. A century of forgetting-curves and a planet of adults who remember almost nothing of what they were taught have graded that bet. The demand-driven wager is the opposite one, that knowledge sticks to questions the way it never sticks to schedules, and it is the only wager in education whose mechanism the motivation science actually endorses. Children arrive at school already asking. The shortest description of reform is: stop interrupting.
Sources
- Bureau, J. S., Howard, J. L., Chong, J. X. Y. & Guay, F. (2021). Pathways to Student Motivation: A Meta-Analysis. Review of Educational Research. 144 samples, 79,079 students; teacher autonomy support strongly predicts intrinsic motivation. https://doi.org/10.3102/00346543211042426
- Mammadov, S. & Schroeder, K. (2023). Autonomy support and positive learning outcomes: a meta-analytic review. Contemporary Educational Psychology 75. 153 studies; autonomy support linked to engagement, mastery, self-regulation, achievement. https://doi.org/10.1016/j.cedpsych.2023.102235
- Wang et al. (2024). SDT-based educational interventions: a meta-analysis. Learning and Motivation 87. 36 studies, 11,976 participants; interventions improve intrinsic motivation and need satisfaction. https://www.sciencedirect.com/science/article/abs/pii/S0023969024000572
- Marchelli, P. S., Dias, C. L. & Schmidt, I. T. (2008). Autonomia e mudança na escola. Revista Psicopedagogia 25(78). Escola da Ponte: no grade-years or subject divisions; students choose projects and declare readiness for assessment; São Paulo public implementations. https://pepsic.bvsalud.org/scielo.php?script=sci_arttext&pid=S0103-84862008000300011
- OECD (2019). Learning Compass 2030 / Student Agency for 2030. Student agency and co-agency as the center of the OECD's 2030 education framework. https://www.oecd.org/en/data/tools/oecd-learning-compass-2030.html