Psychology › Thinking, memory & biases › Memory and how it works
Memory systems: short-term, long-term, episodic, procedural
1968 (Atkinson–Shiffrin); 1972 (Tulving)· Theory or model
Well replicatedStrong, repeated evidence.
Memory is several partly separate systems, from a seconds-long short-term store to lasting memories for events, facts and skills.
See it on the Psychology map →What it says
In 1968 Richard Atkinson and Richard Shiffrin proposed a flow: a fleeting sensory memory, a short-term store holding a handful of items for seconds, then long-term memory. In 1972 Endel Tulving split long-term memory into episodic (personal events, like a closing day) and semantic (facts, like what a cap rate is). Procedural memory covers skills such as riding a bike. Patient H.M. showed these can fail separately.
Why it matters
It explains why amnesia can erase new events yet spare skills, and it guides diagnosis and research on dementia and brain injury.
The evidence
Well replicated. Amnesia patients, brain imaging and lab experiments show distinct systems. Details of the 1968 model were revised, notably its single short-term store, replaced by working memory.
Open questions
How cleanly the systems separate is debated; they interact constantly, and the line between episodic and semantic memory is blurry.
Connections (2)
▸Patient H.M. (Henry Molaison) revealed separate Memory systems: short-term, long-term, episodic, proceduralWell replicated
The connection
After surgery removed much of Henry Molaison's hippocampus in 1953, he could no longer form new memories of facts or events. Yet his short-term memory was normal, and he steadily improved at a mirror-drawing task he could not remember practicing. This showed memory is several systems in different brain areas.
Who showed it and when
Hartford, Connecticut, 1953 surgery by William Scoville; studied from the mid-1950s by Brenda Milner of Montreal, later by Suzanne Corkin at MIT until his death in 2008.
The evidence
Scoville and Milner's 1957 paper, decades of tests, and the 2009 slicing of his brain that mapped the damage; other amnesia patients show the same pattern.
Open questions
His damage extended beyond the hippocampus, so exactly which structures cause which memory losses is still being refined.
▸Working memory replaced the short-term store in Memory systems: short-term, long-term, episodic, proceduralWell replicated
The connection
Atkinson and Shiffrin's 1968 model had a single short-term store acting as a waiting room for long-term memory. Baddeley and Hitch found people could hold several digits and still reason fairly well, which one store could not explain, so they replaced it with a multi-part working memory.
Who showed it and when
Atkinson and Shiffrin at Stanford, 1968; Baddeley and Hitch at the University of Stirling, Scotland, 1974; Tim Shallice and Elizabeth Warrington's patient K.F., 1970.
The evidence
Dual-task experiments, and patients like K.F., who had very poor short-term memory yet normal long-term learning, contradicted the single-store model.
Open questions
Parts of the old model, such as rehearsal moving items into long-term memory, are debated rather than fully discarded.