Action, outcome, and value: a dual-system framework for morality.

Area of Science:

• Cognitive Psychology
• Computational Neuroscience
• Moral Psychology

Background:

• Dual-system approaches are prevalent in explaining human judgment and decision-making.
• Existing models, particularly in moral psychology, face challenges in characterizing distinct cognitive systems.
• Reinforcement learning in computational neuroscience has identified two core algorithms for learning and choice.

Purpose of the Study:

• To propose a novel dual-system framework for understanding moral psychology.
• To bridge the gap between psychological dual-system theories and computational neuroscience models.
• To offer a more robust characterization of the two systems underlying human judgment and behavior.

Main Methods:

• Reviewing existing literature on dual-system approaches in psychology and neuroscience.
• Analyzing formal models of reinforcement learning to identify distinct value representation algorithms.
• Applying the action-based versus outcome-based value representation framework to moral psychology.

Main Results:

• Identified two distinct algorithms in reinforcement learning: one based on past experience (action-based) and another on causal models (outcome-based).
• Proposed that this action- and outcome-based value representation provides a strong foundation for a dual-system theory in moral psychology.
• Suggested this framework can overcome limitations of previous emotional vs. rational/cognitive distinctions.

Conclusions:

• The division between action-based and outcome-based value representation offers an ideal framework for dual-system theories in moral psychology.
• This integration provides a more precise and computationally grounded understanding of moral judgment and decision-making.
• Future research should explore the neural and behavioral correlates of these two value representation systems in moral contexts.