The Questions Children Stop Asking

AIn a much-cited classroom study carried out in Leiden in 1998, an observer counted the questions asked by five-year-olds and by eleven-year-olds during an ordinary school morning. The younger group asked, on average, twenty-six; the older, four. The finding is usually reported as a loss, and the explanations offered for it are familiar ones: examinations crowd out wondering, or children learn that a question is an admission of ignorance. Yet the same figures can be read in a second way. The youngest children ask about almost everything, while the older ones ask about the few things they already know enough about to find puzzling, so the fall in the number of questions may record a change in the kind of curiosity rather than its disappearance. Which reading one prefers turns out to matter a great deal, because the two lead to opposite advice about how lessons should be built. The dispute among researchers is not over whether curiosity helps learning, but over what curiosity is.

BIris Halvorsen, of the Kelso Centre for Learning Research, treats curiosity as a form of discomfort. On her account it is triggered not by ignorance but by a gap: a learner who has just enough information to see the shape of what is missing feels the absence as an irritation, and acts to remove it. The prediction that follows is counter-intuitive. If curiosity came from ignorance, it would be strongest where knowledge was thinnest; on the gap account it should be weakest at both extremes and strongest in between. Halvorsen tested this in 2011 with 118 undergraduates, who rated their curiosity about 60 trivia questions and were then asked, without warning, to recall the answers a fortnight later. Items the students had rated themselves most curious about were recalled 46 per cent of the time, against 29 per cent for the rest. Curiosity, she concluded, does not merely accompany learning; it alters what is retained.

CAnselm Weiss, who studies children in museums rather than undergraduates at screens, thinks the laboratory has smuggled in its own answer. A trivia question hands the learner a ready-made gap; outside the laboratory nobody supplies the gap, and the interesting question is how a learner finds one worth having in the first place. Between 2014 and 2019 his team recorded 209 children aged six to nine at a science centre in Ravensburg, coding how long each stayed at an exhibit and what was done there. The children did not linger at the displays they already understood, nor at those they could make nothing of; they went back to the ones where each attempt taught them a little more than the last. Weiss calls the driver learning progress: what rewards a child is not the closing of a gap but the rate at which understanding improves. On this view curiosity is a pleasure rather than an itch, and it seeks out difficulty instead of fleeing it.

DThe two positions were brought together in a joint paper in 2018 which conceded more ground on each side than either author had expected. Halvorsen accepted that her paradigm measured only one variety, which the paper labelled deprivation curiosity: sharp, uncomfortable, aimed at one missing fact, and extinguished the moment that fact arrives. Weiss accepted that his own variety, labelled interest curiosity, was slow and pleasant, and poor at producing the urgency her experiments reliably generate. The paper's central claim is that the two are not stages of a single process but separate states with different signatures, and that many disagreements in the literature dissolve once experiments are sorted by which state they provoke. Both authors insisted on one further point: neither state is a trait of the learner. The same child will show one in a quiz and the other at a rock pool, and a classroom arranged to suit one may well suppress the other.

EThe distinction has an uncomfortable implication for teaching, which a large trial in 2016 made plain. Twenty-four schools in the Marlow district replaced a term of conventional science lessons with a curriculum built around questions the pupils themselves had proposed. On the end-of-term examination the trial classes scored no higher than the controls. On a separate test six months later, covering material the pupils had never been examined on but had asked about, they scored substantially higher. Halvorsen reads the first result as the decisive one and the second as unreplicated elsewhere; Weiss reads the second as the only result that bears on his claim, since a conventional examination samples the syllabus rather than the pupils' questions. What both concede is that the pleasant form of curiosity works slowly, and that a single term is a short window in which to detect it. Enthusiasm, as Weiss puts it, is not evidence.

FThe advice each offers therefore differs in kind. Halvorsen's is a matter of sequencing: withhold the answer, but only briefly, and only once enough has been given for the class to feel the absence, since a gap that cannot be closed within the lesson turns into frustration. Weiss's is a matter of supply: stock a room with material at many levels of difficulty and let pupils settle at the level where their own understanding is moving fastest. The two prescriptions are compatible, and a handful of schools now run both. What neither can yet do is settle the question that matters most to ministries, namely whether curiosity produces adults who go on learning once examinations stop. Answering it would mean following pupils for decades, and the studies that have attempted this have been unable to separate curiosity from the family circumstances that so often accompany it.