Can conditionals explain explanations? A modus ponens model of B because A

Area of Science:

• Cognitive Science
• Psychology
• Philosophy of Science

Background:

• Evaluating explanations is crucial for scientific understanding and everyday reasoning.
• Existing models often lack a robust probabilistic framework for assessing causal or explanatory relationships.
• Understanding how people reason about explanations under uncertainty is key to cognitive science.

Purpose of the Study:

• To propose and validate a normative, probabilistic model for evaluating explanations of the form 'B because A'.
• To compare the model's predictions with empirical data on human judgments of explanations.
• To investigate the role of inferential relevance in the evaluation of conditionals and explanations.

Main Methods:

• Developed a normative model based on probabilistic conditional reasoning, specifically the modus ponens model.
• Conducted two studies with participants (N=80 and N=376) judging subjective probabilities of statements.
• Varied inferential relevance of 'A for B' in Study 2 to test boundary conditions of the model.

Main Results:

• Participant judgments of 'B because A' systematically followed the model's predictions in both studies.
• A significant proportion of belief sets exhibited incoherence, deviating from the model's normative standards.
• Study 2 results suggest inferential relevance is important for evaluating both conditionals and explanations.

Conclusions:

• The proposed probabilistic model provides a valid framework for assessing explanations.
• While human reasoning aligns with the model, deviations indicate limitations in normative rationality.
• Inferential relevance emerges as a critical factor in how people evaluate explanations and conditionals.