Bartlett's War of the Ghosts: Why Memory Is Reconstruction

    How Memory Works

    Bartlett's War of the Ghosts: Why Memory Is Reconstruction

    In 1932 Frederic Bartlett had people retell unfamiliar stories again and again, and watched the stories bend towards what each teller already knew. What Bartlett's War of the Ghosts work shows about memory, what it does not, and how to use that bending on purpose.

    By Published Updated

    Try a small exercise called Fact or fiction? Think of a novel you read a while back, then a factual book from around the same time. Try to tell yourself what each one was about. The novel usually comes back far more easily: who was in it, where it happened, what happened next. A story has a shape your mind already knows how to hold.

    That exercise is the start of creative reading. When you read a novel you are not trying to remember anything. There is no tension; the scenes simply come to life in your mind. So when you read a factual book, read it as if it were fiction. Read to experience, not to remember, and let the ideas play out as pictures. On dense technical subjects, the shift in how much stays with you can be big.

    Bartlett's War of the Ghosts research shows the other side of that coin. In 1932 the Cambridge psychologist Frederic Bartlett had people retell unfamiliar stories again and again. Their versions drifted towards what made sense to them. Remembering, he concluded, is rebuilding, not replaying a stored copy.

    That one idea sits under the memory techniques that work. The thing that makes a story stick is the same thing that quietly rewrites it.

    Why Bartlett still matters

    It is easy to picture memory as storage. Something happens, a copy goes into a drawer, and later we take the copy out. It is a comfortable picture, and it is wrong. If memory were a set of copies, a story told back to you a week later would come out the same, give or take a few missing words.

    Bartlett found that it does not. Stories changed, and they changed in a direction: towards the teller. That matters for anyone who learns, teaches or relies on what they remember, because it means your memory is always working with what you already know, and it will happily smooth things over to fit.

    It also matters because a great deal of later research grew from it. The work on false memories and on how recalling a memory changes it both build on the idea that remembering is reconstruction.

    The research

    The research: Bartlett, 1932

    • In short: People retold stories that were unfamiliar to them, and redrew pictures, again and again. Their versions shifted towards what made sense to them, so Bartlett concluded that memory rebuilds rather than replays.
    • What they did: Bartlett had people repeatedly retell unfamiliar stories and reproduce pictures, then compared the versions over time. One of those stories, The War of the Ghosts, is the one his name is now tied to.
    • Who was involved: Sir Frederic Bartlett at the University of Cambridge and his participants. How many took part is not given here, because the check against the original did not record it.
    • What they found: The retellings shifted towards what made sense to the teller, in line with their own knowledge and culture. They were not exact copies.
    • What they concluded: Remembering is an act of reconstruction, shaped by what a person already knows, not a replay of a stored copy.

    This is a book, Remembering: A Study in Experimental and Social Psychology, published by Cambridge University Press in 1932. It is a classic and hugely influential, and its central idea underpins much of the later work on reconstruction and false memory. The individual experiments in it have not been checked in detail, which is why you will not see figures, sample sizes or percentages here. Treat it as a landmark idea with a long line of research behind it, not as a single measured result.

    What I take from it

    Start from one line: memory is an act of creation. When you remember something, you make it again from the connections you built. Bartlett is arguably one of the earliest people to show that in the lab. His tellers were not losing the story at random. They were rebuilding it out of what they already had.

    That cuts two ways.

    It is why memory techniques work. A good memory technique gives new material a shape your mind already knows. A chain turns a list into a story, with one thing leading to the next. A memory palace hangs information on a route you can walk in order. Creative words turn an abstract word into a picture of something you can see. In psychology the word for an existing framework like this is a schema. Memory techniques are schemas you build on purpose, instead of leaving your mind to pick one for you.

    It is why you always check. If your memory leans towards what makes sense to you, then the parts that do not make sense, the odd detail, the unusual word, the step you did not expect, are the parts most likely to be smoothed away, and the smoothed version can feel completely right. That is why, when you learn lines, every recall should be followed straight away by a check against the script.

    So the Fact or fiction? exercise is half the lesson. The novel comes back easily because it fits a shape you know. The question Bartlett makes you ask is: how much of what came back is what the author wrote, and how much is what you would have written?

    What Bartlett's work does not show

    • It does not show that memory is useless. Reconstruction is not the same as invention from nothing. His tellers still remembered stories; they remembered them in their own terms.
    • It does not give you numbers. You will not see a percentage of change, a rate of drift or a count of participants, and if you see confident figures quoted elsewhere, check where they came from.
    • It does not show that every detail changes. The finding is about the direction of change, towards what makes sense to the teller, not that nothing survives.
    • It does not support the recording or filing cabinet idea of memory. Quite the opposite. If someone tells you memory works like a video you play back, Bartlett is one of the first places that picture breaks.
    • It is one book from 1932. It is influential and well supported by what followed, but it is a starting point for the science, not the last word.

    How to use it this week

    You cannot stop your memory reconstructing. You can give it a good structure to rebuild from, and then check its work.

    1. Run Fact or fiction? on yourself. Pick a novel and a factual book you read around the same time. Spend two minutes on each, out loud or on paper. Notice which comes back more easily, and how much of it is in your own words rather than the author's.
    2. Read your next factual chapter as if it were a story. Drop the effort to remember. Picture the people, places and ideas as scenes, and let them play out. Then close the book and tell yourself what happened.
    3. Turn a list into a story. Take ten items you need, a shopping list or the points of a talk, and link them into one ridiculous chain where each thing leads to the next. You are handing your memory a story shape on purpose.
    4. Give it a place. Put the key ideas of a chapter on a memory palace, one picture per location, in order. If you would like help with this, LOCI turns a book or your own material into key ideas and images, builds a palace with AI and adds flashcards with the real information on the back.
    5. Connect it to something you know. Before you learn a new fact, ask what it reminds you of. Material tied to what you already know is remembered through that knowledge. Just remember it can be reshaped through it too.
    6. Run your own Bartlett test. Read a short passage today and write it out from memory. Tomorrow, write it out again without looking at either version. Then put both next to the original. What crept in, what fell out, and which way did it drift?
    7. Check against the source, straight after. Every time you recall something that has to be exact, a quote, a figure, a line, compare it with the original immediately and fix the picture that let it slip.
    8. Hunt for the odd details. When you check, look hardest at the unusual word or the step that surprised you. Those are the parts your memory is most likely to have tidied away.

    Where to go next

    For the full picture of how memory is built and rebuilt, start with How Memory Works (A Practical Guide). To see how far reconstruction can go, Lost in the Mall: How False Memories Are Planted covers a study in which people came to remember a childhood event that never happened. Memory Reconsolidation: Why Recalling a Memory Changes It picks up where Bartlett leaves off, with what recall does to a memory in the brain. And to put a story shape to work straight away, The Chain Method: How to Remember Lists Without Building a Memory Palace walks you through it.

    To give a real piece of work a structure to rebuild from, How to Memorise a Sales Pitch shows you how to turn a pitch into a short run of beats and pictures that come back in order.

    Memory map

    1. A novel comes back more easily than a factual book: Fact or fiction? Image: two books on a bedside table, one with characters climbing out of its pages, the other closed and grey.
    2. In 1932 Frederic Bartlett had people retell unfamiliar stories again and again. Image: a line of people in a Cambridge college garden, each whispering to the next.
    3. The retellings drifted towards what made sense to the teller. Image: a paper boat on a stream, slowly turning towards the bank where the teller stands.
    4. Remembering is rebuilding, not replaying a stored copy. Image: a builder reassembling a house from the same bricks, the front door now in a different place.
    5. It is a classic book, so no numbers are put on it here. Image: a heavy old volume on a desk beside an empty set of scales.
    6. Memory is an act of creation. Image: a potter shaping clay on a wheel, a new pot rising from the same lump.
    7. Techniques work because they give material a shape the mind knows: chain, palace, picture. Image: a string of linked objects, a winding corridor and a bright painted image hanging on a wall.
    8. The odd details are the ones most likely to be smoothed away. Image: a sanding block gliding over a carved wooden figure, its sharp little features fading.
    9. Recall, then check straight away against the source. Image: an actor speaking on a bare stage, then glancing down at an open script in their hand.
    10. Run your own Bartlett test: write it today, write it tomorrow, compare both with the original. Image: three sheets of paper side by side on a kitchen table under a lamp.

    Sources

    Are you old school? Get Mark's book Improve Your Memory and dive in depth.

    Loci

    Build the memory palace in this article inside Loci — store your locations, link images, and rehearse with spaced repetition.

    Try it in Loci

    Memory School community

    Learn directly with Mark inside the Memory School community on Skool — monthly lessons, practice, and Q&A.

    Join Memory School

    About the author

    Grand Master of Memory (1995). Author of three books on memory published by Hodder and Nick Hern. Creator of BBC1's Monkhouse's Memory Masters. Memory coach on Channel 4's Can I Improve My Memory?

    Read full bio →