Loading field map
Cognitive psychology is the scientific study of mental processes: how people acquire, represent, transform, store, and use information. Its domain includes perception, attention, memory, language, reasoning, problem solving, decision making, and judgment. The field treats the mind as an information-processing system, and its central project is to explain the internal mechanisms that mediate between environmental input and observable behavior. Unlike behaviorism, which dominated American psychology for much of the first half of the twentieth century and restricted itself to stimulus–response relationships, cognitive psychology insists that unobservable mental states and operations are legitimate, indeed necessary, objects of scientific inquiry.
The field is organized around a cluster of enduring questions. How does the brain transform physical energy—light, sound, pressure—into percepts with meaning? How much of perception is driven by the stimulus itself, and how much is constructed by prior knowledge and expectation? What are the limits of attention, and how do selective and divided attention operate? How is information encoded into memory, why is some of it lost, and how is it retrieved? Is memory a faithful recording or a reconstructive process? How do humans produce and comprehend language, and what does the structure of language reveal about the underlying cognitive architecture? How do people reason deductively and inductively, and why do they systematically deviate from normative logic and probability? What makes a problem hard, and what strategies do people use to solve novel problems? How do people make decisions under uncertainty, and what role do emotion and heuristics play?
The stakes are both theoretical and practical. Theoretically, cognitive psychology aims to characterize the architecture of the human mind—the basic components, their organization, and the principles by which they operate. Practically, its findings inform education, clinical treatment of cognitive disorders, human factors engineering, artificial intelligence, legal testimony, and public policy. The field also carries a philosophical weight: it addresses questions about the nature of knowledge, rationality, and the relationship between mind and brain that have occupied thinkers for millennia, but it addresses them with empirical methods and testable models.
Cognitive psychology emerged as a distinct field in the mid-twentieth century, but its intellectual roots reach back much further. Philosophers such as Plato and Aristotle raised questions about memory, perception, and reasoning. The British empiricists—John Locke, George Berkeley, David Hume—developed associationist accounts of the mind, proposing that complex ideas are built from simple sensations through principles of association. In the nineteenth century, experimental psychology began with Wilhelm Wundt's laboratory in Leipzig, where introspection was used to study the contents of consciousness. Hermann Ebbinghaus's experiments on memory, published in 1885, demonstrated that higher mental processes could be studied quantitatively. William James, in his Principles of Psychology (1890), offered rich theoretical analyses of attention, memory, and habit.
The early twentieth century, however, saw a sharp reaction against the study of mental processes. Behaviorism, associated with John B. Watson and later B. F. Skinner, argued that psychology should be the science of behavior, not of the mind. Introspection was deemed unreliable, and mental entities were considered unscientific. For several decades, particularly in the United States, research on memory, language, and reasoning was marginalized or recast in stimulus–response terms. This was not a universal development; in Europe, figures such as Jean Piaget in Switzerland and Frederic Bartlett in England continued to study cognitive development and memory as constructive processes. Gestalt psychologists in Germany, including Max Wertheimer and Wolfgang Köhler, argued that perception and thought are organized by wholes rather than by simple associations, and their work on insight and problem solving remained influential.
The cognitive revolution of the 1950s and 1960s was a convergence of several streams. In linguistics, Noam Chomsky's 1959 review of Skinner's Verbal Behavior argued that language acquisition and use cannot be explained by reinforcement alone, because children produce and understand novel sentences they have never heard. Chomsky proposed that humans possess an innate, rule-governed system for language. In psychology, George Miller's work on short-term memory capacity and his 1956 paper on the "magic number seven" suggested that information processing has measurable limits. In computer science, the development of digital computers provided a powerful new metaphor: the mind as a symbol-manipulating system. Allen Newell and Herbert Simon built computer programs that solved problems and played chess, demonstrating that complex cognition could be modeled as search through a problem space. Donald Broadbent's work on attention, published in 1958, proposed a filter model in which information is selected early in processing. Ulric Neisser's 1967 book Cognitive Psychology gave the field its name and synthesized its early findings.
The cognitive revolution did not simply replace behaviorism overnight. Rather, it redefined the questions that were considered legitimate and the methods that could be used to answer them. Behaviorist methods—careful control of stimuli, precise measurement of responses, rigorous experimental design—were retained, but the explanatory vocabulary changed. Mental representations and processes became the objects of theory, and behavior became evidence for those internal states rather than the end of inquiry.
Cognitive psychology is not a single monolithic paradigm but a family of approaches that share a commitment to explaining mental processes while differing in their assumptions, methods, and levels of analysis. These approaches have largely coexisted, with periods of dominance and cross-fertilization, rather than replacing one another in a simple sequence.
The information-processing approach, dominant from the 1960s through the 1980s, treats the mind as analogous to a computer. Information flows through a series of stages: sensory registration, pattern recognition, attention, short-term or working memory, long-term memory, and response selection. Each stage transforms the information in some way, and the stages are often depicted as boxes in a flow diagram. This approach was enormously productive. It generated detailed models of memory, such as the modal model of Richard Atkinson and Richard Shiffrin, which distinguished a brief sensory store, a limited-capacity short-term store, and a vast long-term store. It also produced influential theories of attention, including Broadbent's early filter and Anne Treisman's attenuation theory, which proposed that unattended information is not simply blocked but weakened.
The information-processing approach had important strengths. It made mental processes concrete and testable, and it generated precise predictions about reaction times, error patterns, and memory performance. Its limitations became apparent over time. The flow-diagram models were often descriptive rather than mechanistic; they specified stages but not how the stages were implemented. The computer metaphor, taken literally, implied that cognition is serial and rule-based, but much of human cognition is parallel, automatic, and context-sensitive. Moreover, the approach had difficulty accounting for the role of prior knowledge and top-down processing, which became central to later work.
In the 1980s, a rival approach emerged under the banner of connectionism, or parallel distributed processing (PDP). Inspired by the structure of the brain, connectionist models consist of networks of simple, neuron-like units connected by weighted links. Information is represented as patterns of activation across the network, and learning consists of adjusting the weights so that the network produces the correct output for a given input. The most influential learning procedure, backpropagation, was popularized by David Rumelhart, James McClelland, and the PDP Research Group in their 1986 volumes.
Connectionism addressed several limitations of the information-processing approach. It provided a natural account of how knowledge is acquired gradually from examples, how it generalizes to new cases, and how it degrades gracefully when damaged. It also explained phenomena that were awkward for rule-based models, such as the fact that people can learn to pronounce irregular English past tenses and then overgeneralize the regular pattern. However, connectionist models had their own limitations. They were often criticized as black boxes: the networks learned to perform tasks, but the internal representations they developed were difficult to interpret. They also struggled with tasks requiring systematic, rule-governed behavior, such as logical reasoning, where symbolic models seemed more natural. The debate between symbolic and connectionist approaches was vigorous in the 1980s and 1990s, and it has not been fully resolved. Many contemporary researchers use hybrid models that combine symbolic representations with connectionist learning, and the rise of deep learning in artificial intelligence has renewed interest in connectionist principles.
A third tradition, rooted in the work of James J. Gibson, rejects the premise that perception requires internal representations and computations. Gibson's ecological approach, developed from the 1950s through the 1970s, argued that the information available in the ambient light, sound, and other energy arrays is rich enough to specify the environment directly. Perception, on this view, is not a process of constructing a world from impoverished sensory cues but of detecting information that is already there. Gibson introduced the concept of affordances: the opportunities for action that objects and surfaces offer to an organism, such as a chair affording sitting or a handle affording grasping.
The ecological approach differs fundamentally from both information-processing and connectionist approaches. It denies that perception involves inference, representation, or computation, and it emphasizes the role of the active, moving observer rather than a passive receiver of stimuli. Its influence has been strongest in the study of visual perception, particularly in areas such as optic flow, depth perception, and the perception of events. It has also influenced the study of action and motor control. However, it has been less successful in accounting for higher-level cognition, such as language and reasoning, where internal representations seem unavoidable. Most cognitive psychologists accept that perception involves both direct pickup of information and constructive processes, and the debate between ecological and constructivist views remains active.
A fourth approach, which has grown increasingly prominent since the 1990s, treats cognition as a form of probabilistic inference. The Bayesian approach, as it is often called, begins with the observation that the brain must infer the causes of its sensory input, and that this inference is inherently uncertain. The brain is modeled as maintaining a probability distribution over possible states of the world, and perception and cognition are described as updating these distributions according to Bayes' theorem: the posterior probability of a hypothesis given the evidence is proportional to the prior probability of the hypothesis times the likelihood of the evidence under that hypothesis.
This approach has been applied with considerable success to perception, where it explains phenomena such as the integration of multiple sensory cues, the perception of motion, and the effects of prior knowledge on ambiguous stimuli. It has also been applied to reasoning, decision making, and motor control. The Bayesian approach is not a single theory but a framework for building models, and it has been criticized on several grounds. Some critics argue that it is too flexible—that with enough free parameters, a Bayesian model can fit almost any data. Others question whether the brain actually performs the complex computations that Bayesian inference requires, or whether it uses simpler heuristics that approximate Bayesian results. The approach has nonetheless become a dominant framework in cognitive science, and it has fostered a productive relationship between psychology and machine learning.
A more recent tendency, often called the embodied or situated approach, argues that cognition cannot be understood in isolation from the body and the environment. Proponents such as Lawrence Barsalou, Andy Clark, and others contend that cognitive processes are not abstract symbol manipulations but are grounded in sensorimotor experience. Concepts, on this view, are not amodal symbols but are reenactments of perceptual and motor states. Reasoning, problem solving, and even language comprehension draw on simulations of the relevant experiences. The approach also emphasizes that cognition is situated: it occurs in real-world contexts, often with the aid of external tools and the social environment, and it is shaped by the demands of action.
The embodied approach has generated a substantial body of research, particularly on concepts, language, and social cognition. It has also been criticized for being vague and for failing to specify mechanisms as precisely as computational models do. Some of its claims, such as the strong thesis that all cognition is grounded in sensorimotor experience, are disputed. Nevertheless, the approach has had a lasting influence by reminding the field that the mind is not a disembodied computer but belongs to a living organism that acts in a physical and social world.
Cognitive psychology is methodologically pluralistic. The experimental method remains central: researchers manipulate independent variables, control extraneous factors, and measure dependent variables such as reaction time, accuracy, and eye movements. Chronometric studies, which measure the time taken to perform mental operations, have been used since the nineteenth century and remain a staple. The subtractive method, introduced by Franciscus Donders, infers the duration of a mental process by comparing reaction times across tasks that differ by one operation. More sophisticated methods, such as the additive factors method of Saul Sternberg, extend this logic.
Memory research uses a variety of paradigms, including free recall, cued recall, recognition, and priming. Each paradigm taps different aspects of memory, and the patterns of performance across paradigms reveal the structure of the underlying systems. Attention research uses tasks such as dichotic listening, visual search, and the Stroop task, in which participants must name the color of a word that spells a conflicting color name. Language research uses comprehension and production tasks, as well as analyses of speech errors and reading times. Reasoning research presents participants with syllogisms, conditional statements, or statistical problems and examines their judgments and justifications.
In addition to behavioral experiments, cognitive psychologists use computational modeling to instantiate theories as running programs. Models can be symbolic, connectionist, or Bayesian, and they serve several functions: they make theories precise, they generate predictions, and they demonstrate that a proposed mechanism can actually produce the observed behavior. Neuroimaging and electrophysiological methods, such as functional magnetic resonance imaging (fMRI) and event-related potentials (ERPs), have become increasingly important. These methods do not replace behavioral experiments but complement them by providing evidence about the neural substrates of cognitive processes. Cognitive neuropsychology, which studies patients with brain damage, offers a further source of evidence: the patterns of spared and impaired abilities in patients can constrain theories of normal cognition. For example, the finding that a patient can recognize faces but not objects, or can produce language but not comprehend it, suggests that these abilities depend on separable systems.
A recurring issue in cognitive psychology is the relationship between levels of analysis. The field distinguishes between the computational level (what the cognitive system is trying to do and why), the algorithmic level (what representations and processes it uses), and the implementational level (how these are realized in the brain). These levels are related but not reducible to one another. A theory at the algorithmic level, such as a model of working memory, is not invalidated by the fact that the neural implementation is not yet known, and a complete neural account would not by itself explain the cognitive phenomena. The relationship between cognitive psychology and neuroscience is therefore one of mutual constraint rather than reduction.
Contemporary cognitive psychology is characterized by several broad trends. One is the increasing integration with neuroscience, giving rise to cognitive neuroscience as a field that studies the neural basis of cognition. This integration has been productive, but it has also raised questions about the autonomy of cognitive explanation. A second trend is the growing influence of Bayesian and computational approaches, which have provided a unifying framework for modeling perception, learning, and decision making. A third trend is the expansion of the field's scope to include topics that were once considered peripheral, such as emotion, motivation, and social cognition. The relationship between cognition and emotion is now a major research area, and the boundaries between cognitive psychology and social psychology have become porous.
A fourth trend is the increasing attention to individual differences and to the diversity of cognitive processes across populations. Research on cognitive development, cognitive aging, and cross-cultural cognition has shown that many cognitive processes are not universal in their details but are shaped by experience, culture, and development. This has tempered earlier claims about the fixed architecture of the mind and has led to more nuanced accounts of how cognitive abilities emerge and change.
A fifth trend is the growing concern with replicability and methodological rigor. Like many fields in psychology, cognitive psychology has experienced a period of self-examination, with large-scale replication efforts revealing that some published findings do not replicate. This has led to changes in research practices, including preregistration, larger sample sizes, and more careful statistical analysis. The field has also become more open about the distinction between exploratory and confirmatory research.
The field's enduring contribution is its demonstration that the mind can be studied scientifically. The questions that motivated the cognitive revolution—How do we perceive, remember, think, and decide?—remain open, but they are now addressed with a rich array of methods, a sophisticated set of theoretical frameworks, and a body of empirical findings that constrain any account of human nature. The debates among approaches are not signs of weakness but of a healthy field in which competing explanations are tested against evidence. The map of cognitive psychology is not a settled territory but a living landscape, with established regions, contested borders, and new frontiers.