
How Memory Works
Peterson and Peterson (1959): How Fast Short-Term Memory Fades
In 1959 two researchers showed that new information can drop out of short-term memory within seconds if you are stopped from rehearsing it. What Peterson and Peterson did, what they found, how the finding has been reread since, and what to do in the seconds after you hear something.
In How to Remember Anything I described an experiment I had read about in Thompson's The Brain, a neuroscience textbook: memories decayed in about ten seconds when people counted backwards from ten. It is a good line. It is also not quite right, and I want to put it straight.
The experiment is Peterson and Peterson, 1959. What it actually shows is simpler and more useful than my version. If you are distracted from rehearsing, new information can drop out of short-term memory within seconds. Not in exactly ten. Within seconds, and how fast depends on how much you were able to rehearse. Later work has also changed the explanation: much of that loss comes from interference, not just from time passing.
Here is what the study did, what it found, what it does not show, and what to do in those first few seconds after someone tells you a name, a number or an instruction.
Why this matters
Most forgetting that people complain about day to day is not slow. It is instant. You are told a name and it has gone before the handshake ends. Someone reads out a reference number and by the time you find a pen there is nothing there. A colleague tells you which room the meeting has moved to, the phone buzzes, and the room is gone.
That kind of forgetting feels like a personal failing. It is not. It is the normal behaviour of short-term memory when nothing is done with the information while it is there. Knowing that the window is a matter of seconds changes when you act. You do not wait until later to make it stick. You make it stick now.
What the study did
The research: Peterson and Peterson, 1959
- In short: People were tested on single verbal items after delays of a few seconds. When they could not rehearse, the items dropped out of short-term memory fast, and repetition helped them hold on.
- What they did: Each person heard three consonants, a meaningless letter trio, and then straight away a three-digit number. They counted backwards from that number in threes or fours, in time with a metronome, so they could not rehearse the letters. After 3, 6, 9, 12, 15 or 18 seconds they were asked to recall the letters. A second experiment gave some people a few seconds to rehearse first.
- Who was involved: Lloyd and Margaret Peterson at Indiana University, with introductory psychology students: 24 in the first experiment and 48 in the second.
- What they found: Recall fell steeply within seconds. In the second experiment, with no chance to rehearse, people got the letters right about 60 to 70% of the time after 3 seconds, and only about 15% of the time after 18 seconds. Rehearsing first helped: with three repetitions, recall after 18 seconds rose to about 30 to 34%.
- What they concluded: Unrehearsed information drops out of short-term memory fast, and rehearsal slows the loss.
This is one laboratory study with 72 students in total, using meaningless letters, and the rapid loss it found has been repeated many times since. Note the timing: the longest gap in the study was 18 seconds. You will often see "short-term memory lasts up to 30 seconds" quoted; that is a textbook rule of thumb, not a figure from this study. What is less settled is why it happens. The cause is debated: time-based decay, the memory simply fading, or interference. Later research points to interference as a large part of it. In plain words: much of the loss comes from other material getting in the way, not just the clock running down. So treat this as moderate evidence. The speed is solid; the reason is still argued over.
What I take from it
The headline is timing. The first few seconds are where most new information is won or lost. I used to frame this as a countdown, ten seconds and it is gone. I now think the more useful frame is a doorway. While something new is standing in the doorway of your mind, anything else that walks through can knock it out of the way.
That reading fits the interference explanation, but it is worth keeping the two apart. The research says the loss is fast and that interference is a large part of the cause. The practical conclusion here, which is not the paper's own, is that the gap after you hear something is the moment to protect. Whether time or interference does most of the damage, the advice is the same: act at once, and keep other things out of the gap for a moment.
It also matches how actors learn lines. The method in How Actors Learn Lines Fast: The Seven-Step Recipe (From Learning Your Lines) has you say a line aloud from the cue (the small picture you made to trigger it) straight after learning it, not later. That step was not built from this study. It came from the rehearsal room. But it is exactly what Peterson and Peterson would predict: do something with the line while it is still there.
What it does not show
Some things get attached to this study that it cannot carry.
- It does not show that short-term memory decays in ten seconds. That was my book's wording, and it overstates both the precision and the cause. "Within seconds" is as far as the evidence goes.
- The procedure was not counting backwards from ten. In the usual description the counting went backwards in threes, a deliberately awkward task designed to stop rehearsal. My earlier version simplified it.
- It is not proof that time alone wipes memories. The cause is debated, and interference appears to do a lot of the work.
- It does not give you a sample size. The abstract does not state one, so nobody should quote one from it.
- It is about seconds, not weeks. The study tells you what happens in the moments after you hear something. What keeps a memory for months is a separate question, covered in The Forgetting Curve (and How to Beat It).
- It says nothing bad about your memory. People in a task like this lose unrehearsed items fast. That is how the system works, not a sign that yours is failing.
What to do in the seconds after you hear something
Every piece of new information you care about arrives with the same short window. The skill is to use it on purpose. There are three moves, and you want at least one of them within the first few seconds: attend, rehearse, or turn it into a picture.
1. Attend before it arrives
You cannot rehearse what you never caught. When you know something is coming, a name at an introduction, a number on a call, directions from a stranger, decide in advance that you are going to catch it. Stop what you are doing. Look at the person. Let the next thing be the only thing for a moment. The full method for names, including this step, is in How to Remember Names: A Five-Step Framework That Actually Works.
2. Say it back, straight away
Rehearsal is what the study controlled, and rehearsal is what slowed the loss. The quickest version is to say it back in the first second or two. Repeat the number aloud. Read the instruction back in your own words: "So, second floor, the room on the left." Saying it aloud adds your own voice to what you heard, and it lets the other person correct you while it still matters.
This is the actor's version too. Learn the line, then go straight to the cue and say the line from it. Do not move on to the next page and plan to check later. Later is where it goes missing.
3. Turn it into a picture at once
Rehearsal holds something for as long as you keep repeating it. A picture gives it somewhere to live. For numbers, use the number shape system, where each digit becomes something it looks like: 1 is a pencil, 2 a swan, 3 a snake, 4 a sail, 5 a hook, 6 a bag, 7 an axe, 8 a snowman, 9 a balloon, 0 a Polo mint. A number you are told can become a quick scene in the time it takes the speaker to finish their sentence. How to Memorise Phone Numbers shows how to do this for longer numbers.
For an instruction, see yourself doing it, in the place where you will do it. Not a vague idea of the instruction, the actual action in the actual spot. Then say it as you see it. It is the same pairing actors use for scripts: see the cue, say the line aloud.
4. Protect the gap
In the study, counting backwards was the thing that stopped rehearsal. In your life, it is the notification, the next question, the second conversation that starts before the first has finished. Give the new information a few seconds of clear air. Do not reach for your phone while someone is still telling you their number. Do not start the next task until the last instruction is either said back, pictured, or written down.
5. If you cannot encode it, write it
Sometimes there is no time to make a picture and too much to say back. Then write it down, at once. Rehearse it, write it down or encode it (turn it into a picture) straight away: any of the three beats trusting it to still be there in twenty seconds.
A drill for this week
You can feel this effect at home, which is the best way to believe it. This is a home version, not a copy of the study.
- Round one, no help. Ask someone to say a five-digit number to you. Then have them talk to you about something else for about twenty seconds, so you cannot repeat it. Now try to give the number back.
- Round two, rehearse. Same again with a new number, but this time say it back aloud the moment you hear it, before the chat begins.
- Round three, picture it. A new number, and this time turn the digits into number shapes in a quick scene as you hear them, then say the digits as you see the scene.
- Compare. Expect round one to be patchy and rounds two and three far better. Notice which suited you.
- Take it into the day. For the rest of the week, every time someone gives you a name, a number or an instruction, use one of the three moves in the first few seconds. Keep a note of how many you still have at the end of the day.
A week of this changes the habit. You stop hoping things will stay, and start doing the small thing that makes them stay.
Where to go next
To see how small the space is that this information sits in, read The Magical Number Seven, and Why It Is Really Four. For the workspace and its parts, What is Working Memory? covers it. For interference, the other half of this story, Why Do I Forget Things? explains how memories crowd each other. And for what happens after the first few seconds, The Forgetting Curve (and How to Beat It) takes the timescale out to days and weeks.
Memory map
- My book said memories decayed in ten seconds; the study shows loss within seconds. Image: an author crossing out a line in his own book with a red pen.
- Peterson and Peterson, 1959, tested single verbal items after delays of a few seconds. Image: two researchers in a 1950s lab, one holding up a small card, the other holding a stopwatch.
- Without rehearsal, new items dropped out fast; repetition helped. Image: a soap bubble popping in mid-air beside one kept afloat by someone gently blowing on it.
- In the usual account, counting backwards in threes stopped rehearsal. Image: a person staring at a staircase and stepping down three stairs at a time.
- Much of the loss is interference, not just time passing. Image: a crowd of people pushing through a doorway, knocking a small parcel off the step.
- The first few seconds are the window: act at once. Image: a door just ajar with light spilling through, a hand reaching to catch it before it closes.
- Attend before it arrives. Image: a goalkeeper crouched and still, eyes fixed on the ball.
- Say it back straight away. Image: an actor on a bare stage turning to answer a cue the instant it is spoken.
- Turn it into a picture at once. Image: a swan sailing past a snowman holding a giant axe.
- Protect the gap from distractions. Image: a phone face down in a drawer with the drawer sliding shut.
- If you cannot encode it, write it. Image: a pen on the back of a hand.
Sources
- Mark Channon, How to Remember Anything (Hodder Education)
- Mark Channon, Learning Your Lines (Nick Hern Books)
- Peterson, L. R., and Peterson, M. J. (1959). Short-term retention of individual verbal items. Journal of Experimental Psychology, 58(3), 193 to 198. https://doi.org/10.1037/h0049234
- Thompson, R. F. (1993). The Brain: A Neuroscience Primer (2nd ed.). New York: W. H. Freeman. The secondary source behind the description in How to Remember Anything. https://vufind.lboro.ac.uk/Record/283897
- Miller, G. A. (1956). The magical number seven, plus or minus two: some limits on our capacity for processing information. Psychological Review, 63, 81 to 97. https://doi.org/10.1037/h0043158. Cowan, N. (2001). The magical number 4 in short-term memory: a reconsideration of mental storage capacity. Behavioral and Brain Sciences, 24(1), 87 to 114. https://doi.org/10.1017/S0140525X01003922. Broken down in full in The Magical Number Seven, and Why It Is Really Four.
Loci
Build the memory palace in this article inside Loci — store your locations, link images, and rehearse with spaced repetition.
Try it in LociMemory School community
Learn directly with Mark inside the Memory School community on Skool — monthly lessons, practice, and Q&A.
Join Memory SchoolAbout the author
Mark Channon
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 →