“Daniel O’Connell - three ls, three-fold aims”. Trouble is, I cannot for the life of me recall a thing about the aims, and even the memory of who exactly O’Connell was is beginning to slip away. But that was O-level history, circa 1961: wars usually had four causes and five results (or sometimes five causes and four results) while a Victorian municipal act would generally lead to six improvements (which, it must be said, is six more than most of our current legislation achieves). All this had to be memorised and reproduced as “model answers”. It was perfectly possible to get an A grade in history without the slightest interest in or grasp of the subject.

Latin was even more extraordinary. One paper involved translation of passages from Virgil, detailed in the syllabus. In class, we were made to chant the English translation of these passages until we knew them off by heart; when we took the exam paper, the passages could have been printed in Swahili and it would have made no difference. Nobody thought this unusual or improper; it was just one of many incomprehensible things that teachers assured us we had to do to secure our O-levels. I passed Latin comfortably; two years later, I couldn’t understand the simplest Latin phrase.

So why is anybody surprised that model answers and rehearsed phrases are now so widespread? Researchers have gone to some trouble to establish that this is the case, but they might as well have spent the time proving that mice will eat cheese and cats fish. If you set tests, people will find a way of drilling for them. The stakes are far higher now; O-levels may have mattered to the pupils, but the teachers continued serenely on their way whatever the results, untroubled by targets, tables, Ofsted inspections or performance pay.

The Qualifications and Curriculum Authority proposes to tell schools there should be no cramming until a month before the tests. I wonder who is supposed to police this prohibition and how. If schools don’t cram, better-off parents will get private tutors to do it. Then the complaint will be that children from poor homes are disadvantaged.

If we have lots of tests, we shall have lots of cramming. Yet I doubt that employers and universities will stop complaining that young people lack basic skills. On the contrary. Passing tests and examinations is a skill in itself, like driving, typing or bricklaying. The mind can do it on autopilot. This is why employers complain that school-leavers with high GCSE grades in maths and English seem to be innumerate and illiterate. The pupils have learnt maths and English for the purpose of getting through an exam. They file all the information and technique away in a mental compartment labelled “passing the test and pleasing the teachers”. It does not occur to them that the material has any wider applicability, and it drains away once the exam or test is completed.

None of this is to argue that tests and exams have no role at all, only that they should be kept in their place, as a way of reinforcing, supporting and checking children’s learning. Their present role is almost exactly the opposite. Handed down from the centre, the tests and their results determine not only the child’s future, but the teacher’s and the school’s. They are dominant in everyone’s minds. Far from teaching transferable skills, schools end up teaching skills directed at a highly specialised goal.

But this won’t do, will it? Sorry. I meant to write: excessive cramming has four causes and five results.

Peter Wilby is editor of the New Statesman