Britain has already tried reading a child's future off a form

Ministers want primary schools in England to pick out the children most likely to end up out of work. It was reported this week as early help. What it does is make a forecast about a child, and forecasts about children have a way of arranging their own proof.

We know that because Britain ran the experiment already. The break at eleven, the one that still decides when a child changes school, came out of a report of 1926 which said education should divide on the cusp of adolescence. The eleven-plus grew out of the changes to English and Welsh schooling in the years up to 1944, and it rested on an honest belief: that a test could find the strand a child was suited to and send them to the school that fitted. Nobody set out to do harm. That is worth saying plainly, because nobody is setting out to do harm now either.

Look at the age, though. Eleven was not picked because that is when a mind settles. Children were changing school at eleven anyway, so that is where the test went - an accident of the timetable, dressed up afterwards as science.

Nor did it stay a way of allocating. It turned into passing and failing. A child learned on one morning which sort of person the country had decided they were, and so did everyone who taught them afterwards. And the thing the exam caught best is the part nobody wanted to look at: in 1957 a study of the borderline cases, the children who could have gone either way, found that the ones who got the grammar place were more likely to come from middle-class homes. What the exam measured, in the cases that actually mattered, was the house the child went home to.

They say you cannot help a child you have not identified. But nobody is short of names. Any head teacher in a struggling town can list those children before you have finished the question, and list the reasons alongside them. What runs out at the end of that list is the help, not the identification.

And a label will not sit still in a file. It travels - into the group the child is put in, the work they are handed, the tone a tired adult uses on a Friday afternoon, the thing a parent quietly stops expecting. It ends up in the sentence the child says about themselves. Mark a child as a likely failure and you have not spotted the future. You have supplied it.

The goal behind this is right and worth defending. A country with a mass of young people out of work is running up a bill it will still be paying in forty years, and saying so out loud is not cruelty. But the things that actually move it are dull and expensive: an employer paid to take a young person on, a course that ends in a job that exists, a town with somewhere to be at eight in the morning. Ministers were talking about paying employers in the same week as this. Only one of the two costs money, which is usually how you tell which one is meant seriously.

Britain spent decades unpicking an exam that told children at eleven what they were for. The answer was never to move the telling earlier. Tell a child early enough what they are going to be, and they will do you the courtesy of proving you right.

An opinion of the house.