Cognitive Psychology

Cognitive psychology is the study of the mental processes that lie between stimulus and response: how people perceive, attend, remember, reason, decide, and use language.

18 frameworks, from 1879 to today · Psychology · Social Sciences

still a working view ended · each leads to its entry below

Overview

Cognitive psychology is the study of the mental processes that lie between stimulus and response: how people perceive, attend, remember, reason, decide, and use language1. Its central wager is that these processes can be studied scientifically even though they are not directly visible—by measuring behaviour precisely, by building models of what must be happening inside, and increasingly by watching the brain while it works. What is at stake is nothing less than whether the mind can be a natural science: whether thought, memory, and meaning can be explained by mechanisms rather than left to philosophy or introspection.

Before the Name: Structuralism, Functionalism, and the Introspective Mind

Scientific psychology began in the late nineteenth century with a dispute about method. Wilhelm Wundt, who founded a laboratory at Leipzig in 1879, and his student Edward Titchener in the United States, built structuralism: the attempt to analyse conscious experience into its elementary components—sensations, images, feelings—through trained introspection, in which observers reported their inner experience under controlled conditions. The method assumed that the mind's contents were directly accessible to careful self-observation.

Functionalism, associated above all with William James and later with John Dewey and James Angell at Chicago, objected that this missed the point. James's Principles of Psychology (1890) treated consciousness as a stream, not a bundle of elements, and asked what mental processes were for—how attention, habit, and emotion helped an organism adapt. Functionalism was less a laboratory method than a stance: minds are tools shaped by evolution and use, and psychology should study their workings in the world.

Both traditions depended on introspection, and that was their vulnerability. Introspection gave different results in different laboratories, could not be checked by anyone else, and seemed unable to reach processes—like the rapid parsing of a sentence—that occur too fast or too unconsciously to report.

Behaviourism and the Exile of the Inner

John B. Watson's 1913 manifesto, "Psychology as the Behaviorist Views It," argued that psychology should discard consciousness altogether and become the study of observable behaviour: stimuli and responses, habits and conditioning. The programme was enormously successful in the laboratory. Ivan Pavlov's work on conditioned reflexes, Thorndike's law of effect, and later B. F. Skinner's operant conditioning gave psychology precise, reproducible methods and powerful practical applications in education and therapy.

Behaviourism's strength was also its limit. It could describe that an organism learned, but not what it had learned internally. Edward Tolman's experiments in the 1930s and 1940s on rats running mazes showed that animals could learn a layout without any reward and then use it when a reward appeared—behaviour that looked like the use of an internal "cognitive map." The behaviourist vocabulary had no place for such a thing, and the gap became harder to ignore.

The Information-Processing Revolution

The decisive break came from outside psychology as much as within it. During and after the Second World War, engineers and mathematicians developed information theory (Claude Shannon, 1948) and the digital computer. If a machine could manipulate symbols according to rules, then mental processes could be conceived as computations: operations on representations. This was the founding idea of symbolic computationalism, and it gave psychology a new vocabulary—input, encoding, storage, retrieval, output.

The landmark demonstration was George Miller's 1956 paper on the "magical number seven, plus or minus two," which showed that immediate memory is sharply limited in capacity and that people can overcome the limit by chunking—recoding information into larger units. The limit was a property of a processing system, not of behaviour, and it could be measured precisely. The same year saw the Dartmouth conference that named artificial intelligence, and in 1960 Jerome Bruner and George Miller founded the Harvard Center for Cognitive Studies. Ulric Neisser's Cognitive Psychology (1967) gave the movement its name and its textbook1.

The information-processing approach treated the mind as a system of stages: a stimulus is registered, attended, encoded, stored, and retrieved, each stage taking time and subject to interference. Reaction-time experiments made these stages measurable. Saul Sternberg's 1966 studies of memory search, for instance, used response times to argue that people scan a short list of items serially rather than all at once—an inference about an internal process drawn purely from behaviour.

Models of Memory

Memory became the proving ground. In 1968 Richard Atkinson and Richard Shiffrin proposed the multistore model: information passes through sensory registers into a short-term store of limited capacity and duration, and from there, through rehearsal, into a long-term store2. The model made testable claims—about the serial position curve, in which early and late items in a list are remembered better than middle ones, and about the effects of rehearsal—and it organized a generation of research.

It was soon challenged from within. Fergus Craik and Robert Lockhart's 1972 levels-of-processing framework argued that what matters is not how long information is rehearsed but how deeply it is processed: a word judged for its meaning is remembered better than one judged for its typeface, even with equal exposure3. The encoding-specificity principle, formulated by Endel Tulving and Donald Thomson in 1973, added that retrieval succeeds to the extent that the cues available at test match those encoded at study—explaining why context and mood can shape what comes back.

The short-term store itself was reconceived. Baddeley and Hitch's working-memory model (1974) replaced a single buffer with a system of components—a phonological loop for verbal material, a visuospatial sketchpad, and a central executive that allocates attention—later joined by an episodic buffer4. Working memory was not a passive shelf but a workspace actively maintained and manipulated, and the model's component structure was supported by dissociations: concurrent tasks that interfere with one component but not another4.

Perception: Ecological and Computational Alternatives

Perception split into rival programmes. David Marr's computational theory (early 1980s) asked what problem vision solves and at what level of description, treating perception as inference from impoverished input—recovering a three-dimensional world from two-dimensional images. The ecological approach of James J. Gibson (1904–1979) rejected that framing. In The Ecological Approach to Visual Perception (1979), Gibson argued that the environment supplies affordances—opportunities for action specified directly in the ambient optic array—and that perception is direct pickup of that information rather than inference from insufficient data. The two camps disagreed about what perception is: computation over representations, or direct attunement to structured light. Both left lasting marks, and the debate over whether perception is inferential continues.

Connectionism and the Return of Learning

In the 1980s a second computational wave arrived. Connectionism replaced explicit rules and symbols with networks of simple units whose weighted connections are adjusted by learning. The pivotal work was Rumelhart, Hinton, and Williams's 1986 backpropagation algorithm, which showed how a multilayered network could learn internal representations by reducing error, and Rumelhart and McClelland's two-volume Parallel Distributed Processing (1986). Networks could generalize, degrade gracefully, and learn patterns without being programmed with rules—properties that seemed more brain-like than the symbol systems of classical AI. The resulting "symbols versus networks" debate reshaped how psychologists thought about language, categorization, and development.

Cognitive Neuroscience: Watching the Mechanism

A parallel development made the inner processes visible. Studies of patients with focal brain damage—above all the amnesic patient H.M., whose hippocampi were removed in 1953—had already shown that memory is not one faculty: H.M. could learn new skills while denying any memory of the sessions, a dissociation that supported the division between explicit and implicit memory. From the 1980s, positron emission tomography and then functional magnetic resonance imaging allowed researchers to measure brain activity during cognitive tasks in healthy people. Cognitive neuroscience joined the computational and behavioural traditions to anatomy, asking which brain systems implement which processes5. Its findings—that different memory types depend on different structures, that attention modulates sensory processing—fed back into psychological theory rather than merely illustrating it.

Dual Processes, Bayesian Minds, and Prediction

Three later currents reshaped the field's picture of thinking. Dual-process theory, developed from the 1970s onward by researchers including Daniel Kahneman and Amos Tversky and later Keith Stanovich and Richard West, proposed two kinds of cognition: fast, automatic, associative processing and slow, deliberate, rule-governed processing. The heuristics-and-biases research programme showed that the fast system produces systematic errors—judging probability by how easily examples come to mind, for instance—and the two-system framing became a standard way to explain why people reason well in some settings and badly in others.

Bayesian cognitive science treated the mind as an inference engine that combines prior expectations with evidence according to probability theory6. Its appeal was that a single formal principle could explain perception, learning, and reasoning, and that it made quantitative predictions; Bayesian models of cue combination in perception, for example, predicted how people weight conflicting sensory signals, and the predictions held.

Predictive processing, emerging from the 1990s and 2000s in work by researchers such as Karl Friston and Andy Clark, pushed this further: the brain is said to constantly generate predictions about its input and to update its models by minimizing prediction error. Perception becomes controlled hallucination corrected by the senses, and action becomes a way of making the world match expectation. The framework is influential and contested, and its precise empirical content is still being worked out.

Embodied Cognition and the Present

Embodied cognition, gaining prominence from the 1990s, challenged the assumption that cognition is computation over abstract representations inside the skull. Its proponents argued that body, sensorimotor capacities, and environment are constitutive parts of thinking—that concepts are grounded in bodily experience and that much intelligent behaviour is offloaded onto the world7. The claim ranges from modest (bodily states influence judgment) to radical (the mind extends beyond the brain), and the evidence is correspondingly debated.

Today cognitive psychology is plural. Information-processing experiments, computational modelling, connectionist and Bayesian approaches, neuroimaging, and embodied and ecological perspectives coexist, often within the same department and sometimes within the same paper. The field's founding questions remain open: how representations are structured, how much of cognition is modular versus distributed, whether Bayesian or predictive formalisms capture what brains actually do, and how the findings of the laboratory scale to thought in the wild. What has not changed since Miller counted seven chunks is the conviction that these questions can be answered by measurement and model rather than by introspection alone.

References

  1. Cognitive psychology (introduction, History), Wikipedia
  2. Atkinson–Shiffrin memory model, Wikipedia
  3. Levels of processing model, Wikipedia
  4. Baddeley's model of working memory, Wikipedia
  5. Cognitive neuroscience, Wikipedia
  6. Bayesian cognitive science, Wikipedia
  7. Embodied cognition, Wikipedia

The frameworks of Cognitive Psychology

Each framework is a way of working the field has used: when it held, and the question it answers.

  1. 1879-1920

    Structuralism

    What elementary parts make up conscious experience?

    Fly to its tower in Nousfarer

  2. 1890-1930

    Functionalism

    What are mental processes for in adapting to the world?

    Fly to its tower in Nousfarer

  3. 1913-1960

    Behaviorism

    How do stimuli and responses become linked by learning?

    Fly to its tower in Nousfarer

  4. 1956-1990

    Computational theory of mind

    Can the mind be understood as rule-governed symbol manipulation?

    Fly to its tower in Nousfarer

  5. 1956-1990

    Information processing

    What stages lie between a stimulus and a response?

    Fly to its tower in Nousfarer

  6. 1968-1985

    Atkinson–Shiffrin memory model

    How does information move through memory stores?

    Fly to its tower in Nousfarer

  7. 1968-Present

    Multistore and working-memory models

    How is information held and manipulated in mind?

    Fly to its tower in Nousfarer

  8. 1972-Present

    Levels of processing model

    What makes some experiences remembered better than others?

    Fly to its tower in Nousfarer

  9. 1973-Present
  10. 1974-Present

    Baddeley's model of working memory

    What components hold and manipulate information at once?

    Fly to its tower in Nousfarer

  11. 1974-Present

    Dual-process theory

    When does thinking run automatically and when deliberately?

    About Dual-process theory
    In cognitive psychology, Dual-Process Theory addresses a fundamental puzzle: why do people sometimes act swiftly and intuitively, yet at other times reason slowly and deliberately? The framework proposes that the mind operates through two distinct kinds of processing—one fast, automatic, and associative (often called System 1), and the other slow, controlled, and rule-based (System 2). This simple contrast has proven remarkably powerful for explaining a wide range of phenomena, from reasoning biases and social judgment to decision-making under uncertainty. What makes Dual-Process Theory distinct from neighboring frameworks such as Schema Theory or the Computational Theory of Mind is its insistence on a functional division within a single cognitive architecture, rather than treating all cognition as either purely symbolic or purely associative. It emerged as a systematic framework in cognitive psychology during the 1990s, consolidating earlier observations about automaticity and control into a coherent research program.
    A first course in Dual-Process Theory typically begins with the foundational distinction between automatic and controlled processes. Automatic processes are unintentional, effortless, and can run in parallel; controlled processes require attention, effort, and operate serially. From this, the curriculum moves to the two-system architecture: System 1 is fast, associative, and often emotionally charged; System 2 is slow, analytical, and deliberative. Students learn to identify the conditions under which each system dominates, such as time pressure, cognitive load, or the availability of motivation and capacity. Key phenomena include cognitive biases—systematic errors in judgment that arise when System 1 governs decisions that, ideally, would benefit from System 2 correction. The heuristics and biases program (associated with Kahneman and Tversky) provides a rich set of examples, such as the representativeness heuristic and the availability heuristic, where intuitive judgments lead to predictable mistakes. The curriculum also covers dual-task interference as a method to dissociate the two systems: if a secondary task burdens System 2, performance on a primary task shifts toward more automatic processing. Later topics include metacognition—how people monitor and regulate the engagement of System 2—and individual differences in need for cognition, the tendency to enjoy effortful thinking. Finally, students explore applications in social cognition, such as stereotyping (automatic activation of stereotypes) and correction (controlled override).
    The intellectual history of Dual-Process Theory traces back to precursors in the mid-20th century. In the 1950s and 1960s, researchers studying attention and memory began distinguishing between automatic and controlled processes; for example, Shiffrin and Schneider's work on automatic versus controlled visual search (1977) provided an early experimental foundation. However, the modern synthesis took shape in the 1990s. A key figure is Daniel Kahneman, whose Nobel Prize-winning research with Amos Tversky on judgment under uncertainty (1970s-1980s) documented systematic violations of rational choice, laying the groundwork for a two-system explanation. Kahneman's popular book Thinking, Fast and Slow (2011) codified the framework for a broad audience, though the core curriculum was already established earlier. Another seminal figure is Jonathan St. B. T. Evans, who explicitly proposed a dual-process account of reasoning in the 1980s and 1990s, distinguishing Type 1 (heuristic) and Type 2 (analytic) processes in deductive reasoning. Keith Stanovich and Richard West further developed the framework, emphasizing the role of individual differences in cognitive ability and thinking dispositions. Stanovich's work on dysrationalia—the failure to think rationally despite adequate intelligence—highlighted that System 2 engagement is not guaranteed. The framework coexists with its peers: it shares ground with Evolutionary Psychology (which sees some automatic processes as evolved heuristics) and with Connectionism (which models associative learning at the neural level), but it does not require the nativist commitments of evolutionary psychology or the neural implementation details of connectionism. In the 2000s, Bayesian Cognitive Science began to offer normative models of inference that sometimes replace or complement dual-process accounts, and Predictive Processing has reframed automatic processes as predictive inference. Dual-Process Theory thus endures as a practical taxonomy for organizing the mixed outcomes of human thought—smart yet fallible, swift yet reflective.
    The 9 concepts its tower teaches
    • Automatic and Controlled Processes
    • Cognitive Biases
    • Dual-Task Interference
    • Heuristics and Biases Program
    • Metacognition
    • Need for Cognition
    • Stereotyping and Correction
    • System 1
    • System 2

    Fly to its tower in Nousfarer

  12. 1979-Present

    Ecological approach to perception

    How do we perceive opportunities for action directly?

    Fly to its tower in Nousfarer

  13. 1980-Present

    Cognitive neuroscience

    Which brain systems implement mental processes?

    Fly to its tower in Nousfarer

  14. 1982-Present
  15. 1985-Present

    Bayesian cognitive science

    How does the mind reason under uncertainty?

    Fly to its tower in Nousfarer

  16. 1986-Present

    Connectionism

    Can learning emerge from networks of simple units?

    Fly to its tower in Nousfarer

  17. 1991-Present

    Embodied cognition

    How does the body shape the way we think?

    Fly to its tower in Nousfarer

  18. 2000-Present

    Predictive coding

    Does the brain constantly predict its sensory input?

    Fly to its tower in Nousfarer

How the ideas connect

Each arrow runs from an earlier idea to the later one it shaped. The list below says what passed between them.

StructuralismFunctionalismBehaviorismComputational theory of mindInformation processingAtkinson–Shiffrin memory modelMultistore and working-memory modelsLevels of processing modelEncoding-specificity principleBaddeley's model of working memoryDual-process theoryEcological approach to perceptionCognitive neuroscienceComputational theory of visionBayesian cognitive scienceConnectionismEmbodied cognitionPredictive coding

Beyond the field

Where Cognitive Psychology's ideas came from in other fields, and what they led to there: the strongest 12 of 93 links.

A city in Nousfarer: a red-carpeted stair climbing terraces to a crowned tower on an acropolis, more towers round it, forest and snowy mountains beyond.
A city in Nousfarer, the game played in the browser.

Cognitive Psychology is a planet in Nousfarer

Its city has a tower for each of these 18 frameworks. Land, walk in, and climb one concept at a time.

Land on Cognitive Psychology

Related fields

Fields its ideas are linked to

More in Psychology

See where Cognitive Psychology sits on the Map of Human Knowledge