Bayesian reasoning with ifs and ands and ors.

Area of Science:

• Cognitive psychology
• Decision making
• Logic and reasoning

Background:

• Traditional logic uses binary consistency, but human reasoning involves uncertainty.
• Studies show probability judgments often lack coherence.
• Bayesian approach offers a framework for reasoning with uncertainty.

Purpose of the Study:

• To investigate human coherence in probability judgments for conjunctions, disjunctions, and conditionals.
• To determine if explicit inference tasks improve coherence.
• To examine interpretations of conditional statements in reasoning.

Main Methods:

• Participants provided confidence judgments on statements or explicit inferences.
• Two groups: statements group and inferences group.
• Responses analyzed for coherence under different interpretations of conditionals.

Main Results:

• Participants generally exhibited above-chance coherence for most inferences, irrespective of explicit tasks.
• Coherence was maintained when interpreting conditionals in a Bayesian manner.
• Incoherence persisted in conjunction fallacy scenarios, even with explicit inference tasks.
• Incoherence was observed when interpreting conditionals as material conditionals.

Conclusions:

• Bayesian principles adequately describe deductive reasoning under uncertainty.
• Human coherence in reasoning is robust but susceptible to specific logical fallacies.
• The interpretation of natural language conditionals significantly impacts reasoning coherence.