How Learning Works | Learning Theories: Behaviorism, Cognitivism, Constructivism, Social Learning and Connectivism

Learning theories are lenses, not religions

Behaviorism, cognitivism, constructivism, social learning and connectivism notice different parts of the same difficult event: a human being becoming able to understand, remember, decide or do something more reliably than before.

Search for learning theories, theories of learning, behaviorism, cognitivism or constructivism and the same traditions recur because they provide useful ways to organise questions about learning. Berkeley’s overview foregrounds behaviorist, cognitive-constructivist and social-constructivist traditions; contemporary reviews commonly begin with behaviorism, cognitivism and constructivism before extending the map through social, humanistic, experiential, reflective and network perspectives.

The useful question is not “Which theory wins?” It is “What does this theory make easier to notice, what decision does it improve, and what disappears if we use it alone?”

Return to How Learning Works →

Behaviorism: learning becomes visible through changed behaviour

Behaviorism directs attention to observable relationships among conditions, responses and consequences. It supports clear targets, repeated opportunities, informative consequences and measurement of performance. This is powerful when the target itself is observable: accurate recall, fluent calculation, a safe procedure, pronunciation or a motor sequence.

Its limitation is equally important. The same correct response can be produced by understanding, imitation, prompting or luck. Behaviour is evidence about learning, not a complete explanation of learning.

Cognitivism: what happens inside the learner matters

Cognitivism focuses on internal information processing: attention, perception, memory, organisation, schema, retrieval and problem solving. The learner interprets information through an existing cognitive system rather than simply absorbing a stimulus.

Beatrice’s fraction problem illustrates the difference. She can calculate one quarter. Her difficulty is identifying which quantity is now the whole. More repetition of arithmetic does not automatically repair the representation.

Continue through working memory, prior knowledge and retrieval practice.

Constructivism: learners do not receive meaning unchanged

Constructivist traditions emphasise that learners actively construct knowledge in relation to what they already know and experience. New information is interpreted, connected, reorganised and sometimes resisted by existing models.

Faith can hear the definition of average speed and still apply the arithmetic mean because her existing model of “average” is powerful. Learning requires more than exposure to the correct sentence. The old model has to lose explanatory control.

Constructivism does not mean leaving novices alone. Construction happens during explicit teaching too. Guidance should respond to state: modelling and worked examples early, increasing explanation, inquiry, problem solving and transfer as knowledge grows.

Social constructivism and social learning

Learning occurs in social worlds. Language, norms, examples, tools and standards are inherited from communities. Discussion can expose assumptions; explanation to another person can reorganise one’s own knowledge; teachers and peers can temporarily support tasks that are not yet independently manageable.

Denise’s quiet first draft before discussion is a social-learning design decision. The group matters, but giving her a private construction phase prevents the loudest voice from becoming the group’s thinking.

Modelling also matters. A teacher thinking aloud models decision making. A peer who calmly repairs an error models recovery. A parent who says “I don’t know; let’s verify it” models responsible uncertainty. But modelling must expose the relevant process; polished expert performance can hide the decisions a novice needs.

Humanism: learning belongs to a person with purposes

Humanistic traditions foreground autonomy, meaning, growth and the learner’s experience. They remind education that a technically efficient learning system can fail if the learner has no agency, trust or reason to participate.

Respect for agency can coexist with standards, expertise and demanding work. The practical question is whether the learner can increasingly participate in decisions about goals, evidence, help and return.

Experiential and situated learning

Doing something can create rich material for learning, but experience alone does not guarantee that the right lesson is extracted. Experience becomes educational when action connects to observation, interpretation, feedback, conceptualisation and another attempt.

Situated-learning perspectives add that knowledge develops and operates in contexts of activity. Emily may solve equations beneath a textbook heading and hesitate when the same structure appears inside a word problem. Transfer therefore deserves deliberate design.

See How Experiential Learning Works and How Learning Transfer Works.

Connectivism and distributed knowledge

Connectivism emerged in response to learning in networked digital environments. Its useful provocation is that capability can include forming, navigating and updating connections among people, information sources and tools.

No modern professional memorises an entire field. A doctor, engineer, lawyer or teacher must know enough internally to reason while also knowing how to locate, judge and update external knowledge. The danger is confusing access with knowledge: finding an answer is not the same as being able to evaluate it.

Cognitive load, metacognition and motivation extend the map

Cognitive load theory focuses on the limitations of working memory and the instructional consequences of those limits. Metacognition adds monitoring and control: can learners judge what they know and choose a better strategy? Motivation theories ask why effort starts, persists and returns.

These perspectives are not rivals that require one winner. They answer different questions. A lesson can simultaneously manage cognitive load, use retrieval, provide social modelling, protect agency and employ feedback that reinforces a productive behaviour.

Continue through cognitive load, metacognition and motivation.

Alicia, Beatrice, Ciara, Denise, Emily and Faith: six learners, many lenses

Alicia’s evidence-selection problem can be viewed cognitively as a relevance judgement, constructively as a model of what counts as evidence, socially through discussion of competing interpretations and behaviourally through repeated practice selecting defensible evidence.

Beatrice’s timed fraction problem contains a cognitive representation issue, an examination-control issue and a motivational-emotional layer. Ciara’s fast errors can involve reinforcement history, attention and metacognitive checking. Denise’s participation changes with social conditions and task representation. Emily’s dependence on worked examples concerns cognitive load, scaffolding and transfer. Faith’s overgeneralised shortcut concerns schema, construction and conceptual boundaries.

No learner is “a behaviorist learner” or “a constructivist learner.” The theories describe mechanisms and perspectives, not permanent student types.

The synthesis: choose the lens by the decision

Use a behavioural lens when the question concerns observable performance, practice conditions or consequences. Use a cognitive lens when attention, memory, representation or retrieval is central. Use a constructivist lens when prior models and meaning-making matter. Use social perspectives when modelling, interaction, language and participation change the task. Use humanistic and motivational perspectives when agency, value and willingness to engage are central. Use situated and connectivist perspectives when knowledge must travel through contexts, communities, tools and networks.

Then test the explanation against the learner’s next attempt.

What learning theories cannot do

A theory does not diagnose a learner by itself. It does not guarantee an instructional method. It does not make a fashionable activity evidence-based merely because theoretical vocabulary is attached. It does not remove the need for subject knowledge, assessment evidence or professional judgement.

Educational theory is most useful when it sharpens questions. What changed? What information was processed? What prior model was active? What did the environment reinforce? Who performed the thinking? What support remains? What does the learner do when the context changes?

Research route

• UC Berkeley — Overview of Learning Theories
• NCBI Bookshelf — Learning Theories
• The Science of Learning: Why Learning Theories Matter in Graduate Medical Education
• Some Learning Theories for Medical Educators