Reminiscence bump

Why do people in old age recall their late teens and twenties far more vividly than the decades since?

Ask someone in their seventies for the memories of their life and the answers do not spread evenly. Almost nothing comes from before about the age of four. A good deal comes from last month. And then there is a hump that has no business being there at all — a dense band from roughly ten to thirty, peaking in the late teens and early twenties, standing well above what the surrounding decades supply.

The ready explanation is intensity: those were the years of first love, first freedom, first everything, and strong feelings make strong memories. It sounds right. But it makes a prediction we can check, and the check does not go the way you would expect.

If emotional intensity were doing the work, the bump should appear whatever the sign of the emotion. Distress is not a weak feeling. So ask people for their most painful or frightening memories rather than their most important ones, and the hump should still be there.

Largely, it is not. The bump is pronounced for positive memories and weak or absent for negative ones, which tend to distribute more evenly or cluster toward recent years. That asymmetry is hard to reconcile with a pure arousal account, and it pushes you to look for a different mechanism — or, as it turns out, for three of them, working at different stages.

Start with encoding. What is unusual about that stretch of life is not that things happened but that they happened for the first time. The first flight is a memory; the fortieth is not, and by then you cannot distinguish it from the thirty-ninth. A novel event has no existing script to be absorbed into, so it is stored as a particular episode rather than folded into a general one. Later decades are not empty of events — they are full of repeated ones, and repetition builds schema at the cost of episode. Tellingly, when researchers separate out first-time experiences, much of the bump goes with them.

Now consider what happens after encoding. That same window is when a person settles the questions that define them: occupation, partner, place, politics, faith. Memories that support the answer to who am I get rehearsed, told, and returned to for decades, and rehearsal is not neutral — retrieving a memory strengthens it. So the same events are also the most practised, which shows up in the cueing: ask for "most important" or "self-defining" memories and the bump is at its sharpest.

The third mechanism is about neither storage nor strength but access. Every culture carries a rough timetable of expected life events — finishing school, leaving home, first job, marriage, first child — heavily front-loaded into exactly this window. When you search your own past, that script functions as a search plan: it tells you where to look. And notice what it contains. It is a list of positive, expected milestones; there is no normative slot for a bereavement at thirty-four. Which is precisely why the bump shows up for happy memories and not sad ones — not because the sad ones are weaker, but because nothing indexes them.

One more observation suggests something broader than autobiography is going on. The same window governs preferences: favourite music, films and books cluster around what a person encountered in adolescence and early adulthood, and recall of public events does too. Whatever this period is, it is not only laying down personal episodes — it is setting the reference frame against which everything later is judged.

Two clarifications keep this honest.

The first is that the bump is a residual, not a raw fact. Lifespan recall is dominated by two ordinary effects: near-nothing before about age three or four, and a steep advantage for the recent past following the usual shape of forgetting. Draw those and you get a curve falling away smoothly with age. The bump is what sits above it — an excess of memories from a period old enough that they should have faded like the rest. That matters, because it says the explanation must be about something special in that window, not merely about time.

The second is that the three accounts are less rival than staged. Novelty acts at encoding, identity at rehearsal, the cultural script at retrieval, and all three can be true at once. But their relative weight genuinely is contested, and different cueing methods hand victory to different ones — word-cued free recall favours novelty, "most important memory" prompts favour identity, and the positive-negative asymmetry favours the script. Where a study looks determines much of what it finds; this is an open question, not a settled ranking dressed up as a debate.

• Why the same window sets lifelong musical and political preferences, not only memories.
• What childhood amnesia is actually caused by, given that young children plainly do remember at the time.

Every term the collection defines is gathered in the glossary.