Mental models are internal cognitive representations of real-world or hypothetical situations that people use to reason, predict outcomes, and understand complex systems. Philip Johnson-Laird developed mental model theory as an account of deductive reasoning, but the concept has been broadly applied to understanding how people reason about physical systems, social situations, and complex technologies.
Johnson-Laird's Mental Model Theory
In deductive reasoning, Johnson-Laird proposes that people construct mental models representing the situations described by premises. A conclusion is valid if it holds in all possible models of the premises. Errors occur when people construct only some models and fail to consider alternatives. The difficulty of a reasoning problem is predicted by the number of models required — problems requiring multiple models are harder because they strain working memory capacity.
Mental Models of Systems
People develop mental models of how physical and social systems work — how a thermostat operates, how the economy works, how a computer virus spreads. These models are often incomplete, inconsistent, or incorrect, but they serve the pragmatic function of enabling prediction and control. Gentner and Stevens (1983) collected research on "mental models of naive physics" — people's intuitive but often incorrect theories about physical phenomena (e.g., believing that an object dropped from a moving carrier falls straight down).
In human-computer interaction, the concept of mental models is central to usability. Users develop mental models of how software and devices work, and their effectiveness depends on how well these models correspond to the system's actual operation. Good design creates transparent mappings between user actions and system responses, supporting accurate mental model formation. When the user's mental model diverges from the system model, errors and frustration result.