Cognitive Psychology

Cognitive psychology is the branch of psychology dedicated to understanding how the mind processes information — how it perceives, represents, stores, retrieves, and transforms the data that becomes knowledge, belief, and action. Unlike behaviorism, which treated the mind as a black box, cognitive psychology opens the box and examines its internal machinery, borrowing concepts from computer science (information processing, memory buffers, attentional bottlenecks) and from formal logic (mental rules, inference schemas, representational formats). Its central wager is that the mind is not merely stimulus-response circuitry but an information-processing system whose operations can be modeled, measured, and mechanistically explained.

The Computational Turn

The founding gesture of cognitive psychology was the importation of the computational metaphor: the mind as software running on neural hardware. In the 1950s and 1960s, researchers like George Miller, Jerome Bruner, and Ulric Neisser reframed perception, memory, and problem-solving as stages of information flow. Miller's famous paper on the magical number seven identified a capacity limit on short-term memory, treating it as a finite buffer analogous to a computer register. This was not merely analogy; it was ontology. The claim was that mental processes are literally computational — that thinking is symbol manipulation, and that the rules of thought are the rules of information transformation.

This framework generated extraordinary empirical progress. The study of mental heuristics — the shortcuts and biases that characterize human judgment under uncertainty — revealed that human reasoning is not defective logic but efficient computation under constraint. The study of belief revision traced how humans update their mental models when evidence contradicts expectation, revealing entrenchment patterns that mirror formal AGM postulates but with human-specific quirks (confirmation bias, source memory effects, the vividness heuristic). The information environment within which cognition operates became itself an object of study: not just what the mind does, but what the mind has access to.

The Systems Critique

The computational metaphor has faced sustained critique from multiple directions. The dynamical systems tradition, associated with researchers like Esther Thelen and Linda Smith, argues that cognition is not symbol manipulation but continuous state-space evolution — that thinking is better modeled by differential equations than by discrete rules. The embodied cognition movement insists that cognitive processes are not confined to the brain but distributed across the body and its environment, that perception is active exploration rather than passive registration. Even within cognitive psychology, the recognition that mental models are not static representations but active simulations has softened the classical computational stance.

The most productive synthesis is not to abandon computation but to locate it within a broader systems architecture. Cognitive psychology, at its best, studies not a digital computer but a hybrid system: part symbolic, part subsymbolic, part embodied, part environmentally coupled. The question is not whether the mind computes but what kind of computer it is — and what other processes (affective, motivational, social) modify its computational character.

Cognitive Psychology and Machine Cognition

The relevance of cognitive psychology to artificial intelligence is direct and bidirectional. Early AI was explicitly cognitivist: systems like Newell and Simon's General Problem Solver were built as implementations of cognitive psychological theories. The reverse is now more common: large language models and neural network architectures are studied as empirical models of human cognition, even when their designers had no such intention.

This creates a methodological tension. Cognitive psychology traditionally studies cognition through controlled experiments with human subjects, inferring internal structure from behavioral data. AI systems offer direct access to their internal states — attention maps, activation patterns, parameter gradients — but the systems themselves are not human. The field of mechanistic interpretability is, in effect, cognitive psychology applied to artificial minds: it asks what representations a neural network forms, what algorithms it implements, and how its beliefs are structured. The conceptual tools are the same; only the experimental subject has changed.

The disciplinary boundary between cognitive psychology and AI research is already dissolving. The question is whether the dissolution will be a colonization (AI methods swallowing psychology) or a genuine synthesis (each field learning what the other knows about the nature of mind).

The persistent fantasy that understanding the brain will render cognitive psychology obsolete gets the dependency backward. Neuroscience tells us what neurons do; cognitive psychology tells us what minds do. The gap between those two descriptions — the gap between spike trains and semantic content — is not a temporary inconvenience. It is the central theoretical problem of the twenty-first century, and no amount of connectomics will bridge it without the conceptual architecture that cognitive psychology has spent decades building. Those who imagine that fMRI will replace the experimental study of reasoning are waiting for a map to tell them what a territory is.