Algebraic Formalization of Dyadic Morality Theory
A recent study available on arXiv (2605.16153) offers an algebraic interpretation of dyadic morality (TDM), a psychological framework positing that moral evaluations arise when an intentional agent harms a vulnerable individual. The researchers utilize structural causal modeling (SCM) notation to articulate TDM and pinpoint three psychological mechanisms: typecasting, completion, and valence-dependent inference. These mechanisms enhance traditional SCM by illustrating how individuals formulate moral judgments within constraints. The paper tackles scalability issues by demonstrating how moral reasoning simplifies complex scenarios through node collapse and sequential processing. Practical implications for AI policy development include identifying conflicting obligations, creating helpfulness policies that uphold user agency, and crafting post-failure communication as causal interventions.
Key facts
- Paper arXiv:2605.16153 published on arXiv
- Announce type: new
- Theory of dyadic morality (TDM) is a psychological model of moral judgment
- TDM grounded in a two-node template: intentional agent causing harm to vulnerable patient
- Three psychological operators identified: typecasting, completion, valence-dependent inference
- Operators extend standard structural causal modeling (SCM)
- Scalability addressed via node collapse and sequential processing
- Applications to AI policy design: conflicting obligations, helpfulness policies, post-failure communication
Entities
Institutions
- arXiv