Cognitive Load Theory: The Definition Behind Sweller's Research

Cognitive load theory and the information processing model

Cognitive load theory sits inside cognitivism, the paradigm that treats learning as information processing, breaking with behaviorism and its focus on observable stimuli and responses. Within that paradigm, the theory zeroes in on a specific bottleneck: working memory, the mental function that keeps information active while you use it. A 2018 review on collaboration and cognitive load traces this bottleneck’s limits back to the 1950s: a narrow capacity, established by George Miller in 1956, and a short duration, established by Peterson and Peterson in 1959.

Sweller didn’t discover those limits. He used them as the foundation for a different question: what happens when a task asks working memory to hold more elements than it can manage at once.

The origins: Sweller’s 1988 study on problem solving

In 1988, educational psychologist John Sweller published the paper that gave the theory its name, in Cognitive Science. It didn’t start from a classification of load types — that came later — but from research on how people solve problems. Sweller studied a widespread strategy called means-ends analysis: to use it, a problem solver has to hold in mind, all at once, the current state of the problem, the goal state, the relationship between the two, the operators available, and, when multiple steps are involved, a stack of subgoals. According to Sweller, juggling all of that can absorb so much mental capacity that little is left for the theory’s real target: schema acquisition, the stable knowledge structures that remain once the problem is solved.

That observation produced the theory’s first practical application: the goal-free effect. If a problem has no specific goal to reach, a solver can no longer rely on means-ends analysis, because that strategy needs a precise target to work backward from. Remove the target, and working memory has less to hold at once — Sweller already proposed this, in 1988, as a technique for lowering load during learning.

The three types of cognitive load, and a caveat

The three-way split isn’t in the 1988 paper: later research, through the 1990s and 2000s, formalized it on top of that original work. A 2022 study in Frontiers in Psychology, focused on solving word problems, gives a compact description of the three.

A 2018 review on cognitive load in collaborative settings reports a position, attributed to a 2011 paper, holding that germane load is essentially indistinguishable from intrinsic load: both would arise from the same effort of processing content, and treating them as two separate categories ends up being redundant. The three-way split remains the most common framing in the research that followed Sweller, but the third entry, unlike the first two, isn’t something the field agrees on.

Cutting extraneous load: examples from the research

Since intrinsic load depends on the content itself, it isn’t something instructional design can touch without changing what’s being taught. The room to maneuver is in extraneous load, and a 2020 review in the Journal of Education and Health Promotion lists concrete techniques for lowering it: completion tasks, goal-free problems, worked examples, merging multiple information sources into one material, using different modalities together (text and images, for instance), cutting redundancy, and minimizing interruptions during study.

Among these, worked examples have a particular history. In 1988, Sweller cites them only as a bibliographic reference to another paper on teaching algebra; it’s a 2022 study that credits Sweller and colleagues, in 2011, with formalizing the worked-example effect: studying a problem that’s already solved, step by step, produces better outcomes and a lower load than solving the same problem alone, especially for someone new to a subject. The mechanism, according to a 2018 review, is that each step of the example can be considered on its own, without weighing alternative moves at the same time: the interaction between the problem’s elements, and with it extraneous load, goes down.

Practical example: a student in the first lessons on equations learns faster by reading three equations already solved step by step, then trying one alone, than by jumping straight into five equations with no example at all. In the first case, working memory follows one step at a time; in the second, it has to track every possible move while searching for the solution.

Anyone looking to apply these principles to everyday study, not just course design, will find practical pointers in how to study effectively: many of the techniques that cut mental clutter during study work precisely because of the cognitive load a disorganized material adds, unnecessarily, on top of what the subject already demands.