Credibility, Credence, and Coercion
Summary
This post establishes precise terminological distinctions between “credible” and “credence” within Axio’s coercion framework. A threat is credible when the targeted agent assigns high credence (subjective probability) to the realization of threatened harm. This makes credibility inherently agent-relative and subjective—dependent on the victim’s epistemic state rather than objective features of the threat itself. The toy gun example demonstrates that actual capacity to harm is unnecessary for coercion; only the rational expectation of harm matters. This clarifies why coercion operates through belief manipulation rather than requiring realized violence.
Key Concepts
- Credence – Subjective probability or degree of belief an agent assigns to a proposition/event; quantifies confidence or uncertainty.
- Credible threat – A threat to which the targeted agent rationally assigns sufficiently high credence to influence decision-making.
- Agent-relative credibility – Credibility as subjective property determined by victim’s perceptions, not objective threat characteristics.
- Coercion without harm – Effective coercion requires only rational expectation (high credence) of harm, not actual harm delivery.
- Epistemic state dependency – Whether coercion occurs depends on the victim’s beliefs and information, not external facts alone.
Evolution Notes
- Refines the technical definition of coercion introduced in “What Counts as Coercion.”
- Connects credence (from Bayesian epistemology) to ethical analysis of coercion and harm.
- Demonstrates Axio’s methodological commitment to precise terminology as load-bearing for moral reasoning.
- The toy gun example becomes a recurring illustration of belief-dependent harm in later work.
- Lays groundwork for discussions of deception, information asymmetry, and epistemic coercion.
Tags
Cross-References
Open Questions
- How should we handle cases where an agent’s credence is systematically miscalibrated (e.g., paranoia or naivety)?
- Does this framework imply that manipulating someone’s beliefs to lower their credence in a real threat is itself a form of coercion?
- What role does shared knowledge or common knowledge play in establishing credibility in multi-agent scenarios?
- Can credibility be objectively measured, or is it irreducibly subjective?