Everything you've been told about learning is a lie.

The Concise Oxford English Dictionary defines learning as 'knowledge acquired by study'

This is not true. Learning is not the acquisition of knowledge. Not in the sense of 'storing information in your head'. Learning is the encoding of reactions. A collection of reactions (what I have been calling 'affective context') make up an episode - or a story. We use these stories to generate 'knowledge' when it is needed.

In 1974, in what is now a classic experiment, Elizabeth Loftus investigated eyewitness testimony and found it to be unreliable. People shown a video of a car crash would later recall different speeds of impact depending on whether the word 'collided' or 'smashed' were used.

It is sometimes thought that this experiment demonstrates the malleability of information storage. It does not. The information (or knowledge) was never stored, instead people encoded a story - an integrated set of reactions - about what they witnessed. When questioned they used this story to generate a response. In later experiments Loftus found that leading questions could cause people to remember things they had never seen; they may feel shocked at the collision, even more so when it is described as a 'smash'.

A story is never accurate: it is constructed from affective elements. It will often include elements that were never actually present simply because they fit the story - because they were affectively congruent. We remember balloons at birthday parties whether or not they were there. The stories we tell ourselves about the things we have done, the places we have been turn out to be constructed from our affective responses, conjuring up elements that fit the pattern of the way we felt.

Not only does information arise on demand, it can also be affected by the context of that demand: when Elizabeth used the words 'collided' and 'smashed' - which differ principally in their affective context - very different speeds were given.

And affective reactions are more subtle than we imagine: we have a reaction to faces, to gaits, to inferred intent - and eventually to raindrops on roses and whiskers on kittens. The power of this approach is that we can compare concepts based not on their semantic features but on their affective features. We can say "I wandered lonely as a cloud" and it makes sense.

In the 1970s, Based on his work in Artificial intelligence, Roger Schank proposed that humans store information in the form of schema - in a fashion quite unlike computers - by linking related items, sometimes in the form of scripts, sometimes as mental maps. But Roger was only half-way right: certainly humans do not store information like computers do - they do not store information at all. The thinking of the time was, and still is strongly influenced by the Cartesian and computational paradigms. A world where the mind is rational machine.

Psychologists still differentiate semantic and episodic memory. But this is also a lie. Since we do not store information it makes no sense to talk of semantic memory. It is true that we will sometimes abstract from several episodes - for example a child will learn that fire burns but forget the times they were burned. But strangely this is the nature of all episodes: they are all abstractions to some degree. I imagine my train journey to work but cannot recall a specific episode. If you ask me, I can make one up for you. So-called 'semantic memory' is only a particularly revealing case of episodic memory - the story of fire burning, derived from episodes, can be used to conjure memories. All memory works this way, though we don't like to think so. An experience of something we have not experienced first hand is an experience nonetheless. It is encoded in the same way. We may easily remember as episode things that we were merely told.

It seems a great shame that every year, in private and public education billions of people are educated in a way which is accidentally effective at best. An approach which depends for its successes on the unrecognised intuitions of caring educators.  Perhaps a greater shame that this will continue for decades in all likelihood. It is reminiscent of times when people believed the world to be flat and for that reason did not explore - or when civilisations remained ignorant of probability theory, believing chance to be the will of the gods.

But it is also a reminder that we always live in an era of flat earth beliefs; that we are foolish to scoff at the misguided beliefs of previous generations. Generations to come will laugh at us just the same. They will think us silly to have thought that because we have constructed a world of facts and figures, reason and reductionism, we imagined ourselves to be constructed in the same way. What we missed was the subtlety - the sheer complexity - of our affective processing. Distracted by the moving hands, we failed to ever look beneath the dial.