Attributiecriteria

Attributiecriteria are the standards or rules used to determine whether an event, action or property can be attributed to a particular agent, cause or source. The term is used across disciplines such as philosophy, psychology, law and computer science, and the criteria are often context-dependent and debated.

In philosophy and ethics, attribution criteria relate to responsibility and intention. They address questions such as whether a person acted with agency, whether they could foresee outcomes, and whether they were in control of the relevant action. Moral and causal attribution may diverge: one can be causally connected to an outcome without bearing moral responsibility.

In psychology, attribution theory studies how people infer causes for behavior. Classic criteria include consistency (is the behavior repeated in similar situations), consensus (do others behave similarly), and distinctiveness (is the behavior tied to specific contexts). Later models add factors such as controllability and stability, influencing judgments of character or blame.

In law, attribution criteria concern liability and agency. Courts assess whether a party owed a duty, breached it, and caused damage, often requiring a causal link and foreseeability, sometimes also considering intent or negligence.

In data science and artificial intelligence, attribution criteria describe how decisions can be traced to input features or data sources. This includes methods to quantify the contribution of features to a prediction, aiding transparency, accountability and fairness.

Because attribution depends on goals, theory, and domain conventions, different fields may disagree about the appropriate criteria. Critics highlight biases in attribution, and emphasize the need for operational definitions, reliability, and context-aware interpretation.