In 1956, psychologist George Miller published a paper with a title that became one of the most quoted phrases in psychology: "The Magical Number Seven, Plus or Minus Two." It described the apparent limit on how many discrete items a person can hold in working memory — the mental workspace used to consciously hold and manipulate information over short periods, distinct from long-term memory storage. Later research has substantially revised Miller's estimate downward, and the more it's been studied, the smaller the real number appears to be.
The trick that inflated the original number: chunking
Miller's original seven-item estimate turns out to depend heavily on "chunking" — the mental grouping of individual pieces of information into larger, meaningful units. A ten-digit phone number is genuinely ten separate digits, but a person familiar with area codes doesn't hold ten independent items in working memory; they hold something closer to three or four meaningful chunks (area code, exchange, line number), each of which can itself contain multiple digits. Miller's seven-item figure reflected this chunking capacity more than it reflected the true limit on truly independent, unchunked items — and that distinction matters enormously, because chunking effectively lets working memory pack more raw information into the same small number of "slots" by relying on prior knowledge to compress it.
Cowan's revision: closer to four, not seven
Psychologist Nelson Cowan's influential later work argued that once chunking and other strategies (like verbally rehearsing information, effectively refreshing it before it decays) are controlled for, the genuine core capacity of working memory for truly independent items is closer to about four, not seven — a figure now widely cited in cognitive psychology as a more accurate estimate of the mind's actual raw storage capacity. This smaller number has real practical implications well beyond psychology labs: interface designers, for instance, generally try to avoid asking users to hold more than a small handful of unrelated pieces of information in mind at once, precisely because working memory's raw capacity is this limited, however cleverly information gets chunked and organised to work around it.
What we're still unsure about
That working memory capacity is genuinely limited, and that chunking meaningfully extends effective capacity, are both well replicated findings — this isn't seriously disputed. What remains more actively debated is the precise nature of the underlying limit itself: whether working memory functions more like a fixed number of discrete "slots" that can each hold one chunk, or more like a continuously divisible pool of limited attentional resources that can be spread more thinly across more items at reduced precision for each. Different experimental paradigms have produced evidence supporting each model, and cognitive psychologists haven't yet reached a full consensus on which framework more accurately describes what's actually happening in the brain.
This sits inside Attention & Working Memory, one of seven topics in Cognitive Psychology, one of four domains in Psychology, one of seventeen subjects the app can quiz you on.