As presently conceived, “personalised learning” is just a ragbag of soundbites and good intentions. Designed presumably to chime with the Government’s feel-good obsession with “choice”, it lacks moral force, intellectual coherence, and anything much of an evidence base. Even the well-researched and useful “assessment for learning” has been co-opted and finessed into some kind of instant computerised read-out of student achievement. Whether things will improve with the arrival of the new Secretary of State remains to be seen. At this month’s North of England Education Conference, Ruth Kelly appeared to distance herself from personalised learning and the jargon associated with it.

Read the Government literature and tell me why I cannot systematically replace the word “learning” with old-fashioned “teaching”? And if I can’t, tell me what insights in personalised learning give sharp guidance to teachers to do something different from the status quo?

Categorising pupils by “learning styles” - visual, auditory or kinaesthetic - has been rightly rubbished by Professor Frank Coffield’s research review for the Learning and Skills Development Agency. “Our research suggests labelling a pupil as, say, a ‘visual’ learner may do them more harm than good”, he wrote in ‘The TES’ this month. “Moreover, as the tools used to split learners into different categories are so unreliable, most such lablels seem to be of dubious value.” Of 13 learning styles instruments examined, only one met the minimal criteria for a proper psychological test.

Meanwhile, “multiple intelligences” have been exposed as shallow and ill-founded by Professor John White, also of London University’s Institute of Education. In a recent public lecture he meticulously demolished the idea that “children come hard-wired with a whole array of abilities in varying strengths”. The “intelligences” that Howard Gardner has proposed are not scientifically “discovered”, do not have dedicated areas in the brain and are not universal across cultures. Professor White argues that they are merely one value-laden way among dozens of carving up the domains of intellectual life.

So what exactly is the Big Idea?

Is anything to be salvaged from this dog’s dinner? Perhaps we can help Ruth Kelly as she tries to sort out the mess.

Everyone is in favour of “improving students’ learning”, but they tend to slide between three quite different things that it might mean. Is it just another way of talking about raising standards? Is it “raising standards by helping students learn better” through better study skills and techniques (mnemonics, mind maps), different “learning styles” or “multiple intelligence profiles”, and creating optimal learning conditions (such as open-ended questions and bottles of water)? Or is it the third possible meaning - “helping students become better learners”? This means creating a school climate that cultivates generic, portable “learning dispositions and learning identities”. To be “disposed” to persist, for instance, means to show persistence in a broad range of occasions, and the face of obstacles.

A disposition to ask questions means it is second nature and is done in any subject, even in the face of discouragement from the teacher.

It does no good to try to fudge the differences between these three meanings. It is possible to raise standards (by spoon-feeding, for example) in ways that make young people more, not less, dependent on the teacher.

And you can help young people learn better by creating an ideal learning environment - safe, quiet, orderly, purposeful - that does nothing to prepare them to cope with the sub-optimal conditions they will find in the workplace.

Only the third aim widens the purpose of education, to embrace the rhetoric of, as the former education minister David Miliband put it, “learning how to learn in preparation for a lifetime of change”.

The language of personalised learning persistently confuses and conflates these three possibilities. Yet the rapidly accumulating weight of evidence is encouraging. If you help young people become more confident and capable learners and persistently encourage them to think of themselves not (just) as more or less successful “knowers”, but as continually improving finder-outers, the GCSEs and Sats go up as well. This third option, although it is still the least familiar, is beginning to look like the win-win option.

This line of thought leads to a credible and coherent way of thinking about personalised learning. Everyone in a gym is doing personalised learning: they are working at a level that’s right for them and that stretches them without causing damage. After a bit of initial help they are able to regulate their own progress. And they compare themselves not with the hunk across the room but with their own ability a month or a year ago.

Students can tune classroom activities so they set themselves an enjoyable level of challenge, keep track of their own developing learning power and create their own “learning to learn targets” for the next half-term. (Tom is working on asking good questions, while Kerry’s challenge is to plan more thoughtfully, and Naseem’s is to work more productively with different groups.) This approach does not mean chucking out the content. Just as the gym needs equipment that affords interesting challenges, so classrooms need topics that do the same. The teacher just needs to develop a “split-screen” way of thinking. On the left-hand is the content: learning about relative clauses, estimating or the properties of potassium. On the right is today’s learning-to-learn objective: persisting with difficulty, self-evaluation or critical thinking.

If personalised learning has a future, this is surely it. Seeing education as centrally about the cultivation of learning capabilities narrows the gap between the rhetoric of lifelong learning and the reality of students’

school experience. There is a good deal of detail to be worked out, but there is no doubt that learning itself is learnable - nor that teaching it is a key aim for 21st century education. And if it gets Ruth out of a hole, so much the better.

Guy Claxton is professor of the learning sciences at the University of Bristol . His latest book, ‘The Wayward Mind: an Intimate History of the Unconscious’, is published by Little Brown