Mind & Psychology · Memory & learning
Working memory and its limits
You can hold only a few things in mind at once. Learn how big that mental workspace is, how chunking stretches it and how its limits shape the way beginners learn best.

What you will learn
- Explain what Miller's memory span results showed and why the limit counts items rather than information
- Explain why tests that block rehearsal and grouping put working memory's limit at about four chunks
- Explain how grouping and renaming items turns many items into a few chunks
- Explain why chunking depends on knowledge already stored in long-term memory
Lesson 1 of 4 · 4 min
Seven things in mind, or closer to four
In 1956 the psychologist George Miller began a paper by saying he'd been "persecuted by an integer." The integer was seven, and it kept turning up in his data. That paper became one of the most cited in psychology. So what was he counting?
One of his sevens was memory span. That's the longest list you can repeat back in order right after hearing it, on half your tries. Young adults manage about seven digits. Psychologists call the small mental workspace that holds them working memory.
Miller noticed something odd. A binary digit, 0 or 1, carries one bit of information. A decimal digit carries about 3.3 bits, and a word about 10. Yet spans were roughly the same size for all three. So the limit wasn't on the amount of information. It was on the number of items, which Miller called chunks.
He didn't think seven was magic. He ended by suspecting that all the sevens in his data were "a pernicious, Pythagorean coincidence."
Later work showed that span isn't fixed even in chunks. It's about seven for digits, six for letters and five for words. People also use tricks. They rehearse, saying the list over silently, and they group items together. Both stretch the score.
In 2001 Nelson Cowan pulled together studies that blocked those tricks. Some flashed a whole set of items at once. Some had people keep saying a word like "the" aloud, which ties up the inner voice. Others used running span, where a list stops without warning and you report the last few items.
Under those conditions young adults held three to five chunks, about four on average. Cowan argues this is the real central limit. Children and older adults hold fewer.
Try to recallWhy do some memory researchers make people repeat a word like "the" aloud during a memory test?
Saying a word over and over ties up the inner voice people use to rehearse a list. With rehearsal blocked, the test shows what working memory holds without help. That's how studies find a limit of about four chunks rather than seven.
Miller's seven is the span you get when you can rehearse and group. Take those tricks away and you're working with about four chunks at a time.
Lessons in this topic
Seven things in mind, or closer to four
Learn what George Miller's famous magic number seven measured, and why later tests put the real limit nearer four.
How chunking packs more into a small memory
See how grouping and renaming turns many items into a few, and why experts seem to have bigger memories in their own field.
The inner voice and inner sketchpad of working memory
Learn the parts of Baddeley and Hitch's model of working memory and the experiments that show why some tasks clash and others don't.
Why beginners learn more from worked examples
See how working memory's limits shape good teaching, why worked examples help beginners and why the same help can hold back someone more skilled.
Three quick stories from this topic
The 79-digit runner
A student trained his digit span from 7 to 79 by hearing the digits as race times
It took more than 230 hours of lab practice. His working memory didn't grow. Groups of digits became running times he already knew, so each chunk pointed to digits stored in long-term memory.
A runner's code
The student, known as SF, was a long-distance runner. Strings of digits sounded to him like familiar race times, and he combined those into bigger groups.
Not a bigger memory
Researchers concluded his practice sped up storing and fetching digits from long-term memory. The few slots of working memory stayed the same.
Miller's doubtful seven
George Miller suspected that his famous magic number seven was only a coincidence
His 1956 paper opens with him joking that an integer had persecuted him. It ends by listing sevens from the seven seas to the days of the week and calling them a likely "Pythagorean coincidence."
What he measured
Young adults can repeat about seven digits back in order. Miller showed this limit counts items, or chunks, rather than bits of information.
A smaller number
When rehearsal and grouping are blocked, young adults hold about three to five chunks, according to Nelson Cowan's 2001 review.
A word that blocks rehearsal
Saying "the" over and over erases the memory advantage short words have over long ones
You normally hold more short words because your inner voice can repeat them before they fade. Keep your mouth busy with "the" and that silent rehearsal stops, so short and long words fade alike.
An inner voice
Working memory keeps about two seconds of speech alive by silent repetition. Long words take longer to say, so fewer fit in the loop.
Evidence for a loop
Because blocking speech removes the effect, Alan Baddeley and Graham Hitch took it as evidence for a sound store refreshed by rehearsal.
Sources
- The magical number seven, plus or minus two: some limits on our capacity for processing information, Psychological Review (George A. Miller, 1956), Classics in the History of Psychology, York University
- The Magical Number Seven, Plus or Minus Two, Wikipedia
- Memory span, Wikipedia
- The magical number 4 in short-term memory: a reconsideration of mental storage capacity, Behavioral and Brain Sciences (Nelson Cowan, 2001), PubMed abstract
- The magical mystery four: how is working memory capacity limited, and why?, Current Directions in Psychological Science (Nelson Cowan, 2010), PubMed Central
- Acquisition of a memory skill, Science (K. A. Ericsson, W. G. Chase and S. Faloon, 1980), PubMed abstract
- Chunking (psychology), Wikipedia
- Working memory, Wikipedia
- Working memory, Scholarpedia (Alan Baddeley and Graham Hitch, 2010)





