Unused curriculum

Why does a curriculum keep teaching material almost no student will ever use?

Ask an adult when they last factorised a quadratic, labelled the parts of a flower, or used the subjunctive of a language they studied for five years. Usually the answer is never. Yet these are not fringe lessons — they sit at the centre of the timetable, and survive every reformer arriving to teach mortgages and wiring instead. Either schooling is stupid in a remarkably durable way, or it is doing something other than transmitting content that later gets used. Which?

Test the complaint on its own terms first. "Almost no student will use this" is a claim about an average, and an average is the wrong instrument here. Picture a cohort of a thousand thirteen-year-olds, of whom perhaps thirty end up somewhere trigonometry is load-bearing. The designer is not choosing whether to teach the average student trigonometry; they are choosing whether to leave those thirty without it — and cannot tell which thirty. Neither can the thirteen-year-olds.

So ask a different question. Not what fraction of students use this, but what does it cost the student who needed it and did not have it? Those are separate quantities, and the second is what breadth buys. This is option value: you hold something you will probably not exercise, because the cases where you would want it cannot be identified in advance and the loss in those cases is severe.

Why severe? Because school subjects are chains, not items. Missing trigonometry does not merely leave a hole shaped like trigonometry; it quietly removes physics, engineering and much of economics from what you can start at eighteen without going back several rungs. A door closes without a sound, and the student who never opens it never learns it was there. Cheap to hold, expensive to reacquire in sequence, invisible when lost — that asymmetry is what makes the option worth its premium.

The premium also rises with uncertainty. If we knew which skills a cohort would need in twenty years, breadth would be waste and we should specialise at ten. We do not: the confidently vocational curricula of the past obsolesced faster than the useless ones, and it is the abstract material that has aged best.

That is the strongest argument, but not the only thing going on, because the credential does work the content does not. An employer cannot observe patience, trainability, or willingness to finish a dull task on a deadline. A qualification that required years of exactly that is informative about those traits whether or not the material was used again. Economists call this signalling, and the tell is the sheepskin effect: the year of study that completes a degree carries a substantially larger earnings gain than an equivalent year that does not, which is hard to explain if only accumulated knowledge were being paid for. How much of the return to education is signal and how much genuine capability is properly contested — some argue most, others a modest slice — and the honest position is that both are present and the split unsettled.

Then the argument people reach for first, which is the weakest. "Latin trains the mind; mathematics teaches a rigour you will apply everywhere." That is a claim about far transfer, and the evidence against it goes back to Thorndike a century ago and keeps arriving. Skills learned in one setting mostly stay there. Near transfer, within a domain sharing structure, is real — algebra genuinely shows up in a spreadsheet — but general mental strengthening cannot carry the weight put on it.

And finally the unglamorous cause: inertia. Syllabi are written by people trained on them, examined by boards invested in them, and frozen in place by university entrance requirements. Some unused content is there because option value justifies it. Some is there because nobody has the appetite to remove it.

Three defences, doing different work and deserving different confidence. Option value is the strong one and it is not cynical: uncertainty about who needs what, plus prerequisite chains that make late acquisition expensive, genuinely justifies teaching more people more than will use it. Signalling is real, and explains why difficulty and completion are rewarded independently of content, but its magnitude is argued over. Mind-training is the one to hold most loosely.

One refinement: "unused" is measured badly. The adult who cannot factorise a quadratic may still recognise a problem as one about rates, know a statistic can be built to mislead, and know when to fetch an expert. That residue never shows up in a survey asking when you last used the quadratic formula.

The cost side belongs here too. Because hours are rival, option value argues for breadth in general, not for any particular topic. The defensible test is roughly: how badly is a student hurt if they need it and lack it, and how hard is it to pick up later? Content scoring low on both has a weak case however traditional it is.

• How countries that track students into vocational and academic streams early trade option value for fit, and what happens to the ones sorted wrongly.
• Whether credential inflation is a signalling equilibrium that ratchets, and what would stop it.

Every term the collection defines is gathered in the glossary.