Reasoning about properties: A computational theory

Area of Science:

• Cognitive Psychology
• Logic
• Artificial Intelligence

Background:

• Human reasoning about properties has been studied since antiquity.
• Previous psychological theories inadequately explain complex inferences beyond basic syllogisms.

Purpose of the Study:

• To present a novel theory of property reasoning based on mental models.
• To explain how individuals construct and manipulate models to draw conclusions.
• To account for individual differences in reasoning strategies.

Main Methods:

• Postulating that assertions establish relations between properties.
• Representing these relations using mental models of sets of entities.
• Developing heuristics for scanning models to derive conclusions.
• Implementing the theory in a computer program with System 1 and System 2 components.

Main Results:

• The theory explains the generation, necessity, possibility, probability, and consistency of conclusions.
• A computer program successfully simulated human reasoning across over 200 inference types.
• The model accounts for individual variations in intuitive and deliberative reasoning.

Conclusions:

• The presented theory offers a comprehensive account of property reasoning.
• It extends to complex inferences involving quantifiers, disjunctions, relations, and properties of properties.
• The mental model approach provides a robust framework for understanding human deductive reasoning.