1. Definition
Mental models in Johnson‑Laird’s framework are internal structural representations that people construct to simulate states of the world. Reasoning occurs not by applying formal rules but by manipulating these models to test possibilities, draw inferences, and detect contradictions.
2. Core Principles
Model construction — individuals build mental models from linguistic input, perception, or prior knowledge.
Possibility representation — each model encodes one possible state of affairs; reasoning requires exploring alternative models.
Minimal models — people tend to construct the simplest model consistent with the information, leading to systematic reasoning errors.
Model manipulation — inference arises from inspecting, transforming, or comparing models.
3. How Reasoning Works in the Model
linguistic or perceptual input → construction of an initial model
generation of alternative models (often incomplete or omitted)
evaluation of consequences within each model
selection of conclusions consistent across all viable models
Errors occur when alternative models are not generated or when the initial model is overly simplified.
4. Typical Phenomena Explained
deductive reasoning errors (failure to consider all possibilities)
syllogistic reasoning patterns
conditional reasoning (“if–then” tasks)
spatial and relational reasoning
illusory inferences caused by incomplete model sets
Johnson‑Laird’s theory explains why people reason well in some contexts and systematically fail in others.
5. Distinctions
Rule‑based theories — emphasize formal logical rules; mental models emphasize simulation.
Probabilistic models — focus on uncertainty; mental models focus on structural possibilities.
Schema theories — long‑term knowledge structures; mental models are temporary, task‑specific constructions.
6. Example
Given “The circle is to the left of the square. The triangle is to the right of the square,” people construct a spatial model and infer the relative positions. Errors arise if they fail to consider mirrored or alternative configurations.
7. Why It Matters
Johnson‑Laird’s theory reframes reasoning as model‑based simulation, not rule application. It connects language, perception, and inference under a single representational mechanism and explains both human strengths and systematic biases in reasoning.