The Man Who Forgot Tomorrow

Henry Molaison was born in Connecticut in 1926, and by the time he was a young man his life had been taken over by epilepsy. The seizures had begun in childhood — a knock from a bicycle was blamed at the time, though nobody was certain — and they had grown steadily worse until he could not hold down work. In 1953, when he was twenty-seven, he was operated on by a neurosurgeon named William Scoville, who removed tissue from deep inside both temporal lobes, including most of a small seahorse-shaped structure called the hippocampus. The operation was experimental and Scoville had said so. It worked, in the narrow sense that the seizures improved. It also destroyed Henry's future.

When he woke, Henry could still talk, reason, joke and complete a crossword. His childhood was intact and his intelligence was undamaged. But from that morning onwards, nothing new was ever stored. He read the same magazine again and again with fresh interest. He was told of his father's death and grieved, and then was told again, and grieved again. A psychologist called Brenda Milner began studying him in the mid-1950s and went on doing so for the next fifty years; every single time she walked into the room, she was a stranger who had to introduce herself. Henry knew something was wrong. He described it once, with terrible clarity, as being like waking from a dream — every day alone, whatever sorrow or joy he might have had.

Then came the experiment that changed the science. Milner sat him down with a five-pointed star to trace, but he could only see his hand in a mirror, which turns the movement backwards and makes fools of everybody. Henry was hopeless on the first day, as everyone is. She brought him back the next day, and the day after. His accuracy improved steadily, exactly like a healthy person learning a new skill — and each morning, picking up the pencil, he assured her politely that he had never attempted anything like this before. His hands had been learning while he had been forgetting. In that one result, memory split in two: the memory of facts and events, which you can state, and the memory of skills, which you can only perform. They were not one faculty after all, and they did not live in the same place. Almost everything now known about human memory rests on a foundation that starts with Henry. He was known to the world only as "H.M." until his death in 2008; afterwards his brain was cut into more than two thousand thin sections, an operation that was broadcast live on the internet.

If Henry showed what memory is built from, a psychologist called Elizabeth Loftus spent her career showing what it is built out of — which turns out to be far less solid than anyone assumed. In an experiment in the 1970s, students watched a film of a car accident and were then asked how fast the cars had been going. The question was worded differently for different groups. Those who were asked how fast the cars were going when they smashed into each other estimated around forty-one miles per hour; those asked when the cars hit each other said around thirty-four. One verb, seven miles an hour. A week later they were asked whether they had seen any broken glass. The "smashed" group were more than twice as likely to say yes. There had been no broken glass in the film. The question had not merely influenced their answer; it had gone back into the memory and left something there.

Loftus then asked the harder question: could you install an entire event that never happened? In the mid-1990s she and a colleague recruited volunteers and, with the help of an older relative, presented each of them with four childhood stories — three true, one invented, in which the participant had been lost in a shopping centre at about five years old, had cried, and had been rescued by an elderly stranger. Roughly a quarter came to remember it. Not vaguely: they added details nobody had given them — what the man was wearing, how the shop had smelled, how frightened they had been. When it was explained afterwards that the event had never occurred, some of them argued. The memory was as vivid as any other memory they owned, because by then it was made of the same material.

This work has made Loftus enemies, and the reason is that it does not stay in the laboratory: it walks into courtrooms, where she has testified in cases that people feel very strongly about, and where the implications for those who have genuinely been harmed are painful and contested. So the honest summary matters. The finding is not that memory is worthless or that people who remember things are lying — most memory is broadly serviceable most of the time, which is why we survive. The finding is narrower and sharper: memory is reconstructive rather than reproductive, it is edited by every question that is asked about it, and — the part that should genuinely worry you — confidence and vividness are not evidence. A witness may be certain, detailed, entirely sincere and completely wrong, and there is no feeling from the inside that distinguishes that state from being right.

Long before either of them, a German psychologist named Hermann Ebbinghaus had decided to measure forgetting, and had done it the hard way: on himself. Because ordinary words carry meanings and associations that would contaminate the measurement, he invented thousands of meaningless syllables, learned lists of them by heart, and then tested how many survived after an hour, a day, a week. When he plotted the results in the 1880s he produced what is still called the forgetting curve: a steep collapse in the first hours, then a slower and slower decline. Most of what you learn today will be gone within days — and this is not a defect. A brain that kept everything would be unable to think at all, drowning in a detail it could never generalise from. Forgetting is not memory failing. It is memory doing its job, which is to throw away almost everything and keep what looks likely to matter.

The practical question, then, is how to change what "looks likely to matter". Ebbinghaus found the answer in his own data and a century of research has confirmed it. What defeats the curve is not more time at the desk and it is certainly not rereading — rereading is the most popular study technique in the world and one of the least effective, because familiarity feels exactly like knowledge and isn't. What works is retrieval: shutting the book and dragging the information out of your own head, which is unpleasant, feels like failing, and is the entire point. The effort of the search is what marks the memory as worth keeping. And the second half is timing — spacing those retrievals out, coming back just as the thing begins to slip, at wider and wider intervals. Difficulty at the moment of recall predicts strength later. Ease predicts nothing.

Which returns us to Henry, who could not do any of this, and to the small strange fact that his hands could. Every time you shut your eyes in this lesson and told the story back before pressing the button, you were running Ebbinghaus's experiment on yourself, and the reason it was uncomfortable is the reason it worked. There is a link, too, to the conversation elsewhere in this collection about attention: nothing can be retrieved that was never properly encoded, and nothing is properly encoded while your attention is somewhere else. A phone in the room does not merely interrupt the evening. It quietly decides which parts of it you will still own in ten years' time. That is the real stake of this topic — not trivia, not exams. What you attend to is what you will have been.