Working memory
The small mental workspace that holds and manipulates information in the moment.
What it means
Working memory is the limited-capacity system that temporarily holds and actively works with the information needed for reasoning, comprehension, learning, and planning. In the influential Baddeley and Hitch model it is not a single store but a set of components — a central executive that allocates attention, plus a phonological loop for verbal material and a visuospatial sketchpad for visual and spatial material, later joined by an episodic buffer. Its capacity is famously narrow, often summarized as a handful of 'chunks,' and this bottleneck is one of the central constraints on cognition: when the contents to be juggled exceed capacity, performance on reasoning and self-control degrades. The system distinguishes mere short-term storage from the manipulation of held information, which is why it predicts fluid intelligence and academic achievement better than passive memory span. It is taxed by cognitive load and competes with other demands, which is why distraction and stress impair thinking. It matters because almost every complex task runs through this small workspace, so designs that reduce load — chunking, simplification, external aids — directly improve performance.
Examples
Holding a phone number in mind while you hunt for a pen taxes working memory, and one interruption can wipe it out.
A cook holding "stock after the onions brown, salt only at the end" in mind while the pan spits and the phone rings loses the thread entirely.
A signup form that asks for a code shown on the previous screen forces users to carry it in the phonological loop; printing the code beside the field removes the load.
First described in Baddeley & Hitch (1974); Miller's 'magical number seven' (1956).