Causes and Explanations: A Structural-Model Approach. Part II: Explanations

By Joseph Y. Halpern, Judea Pearl. 2001

323 citations recorded by the source corpus.

We propose new definitions of (causal) explanation, using structural equations to model counterfactuals. The definition is based on the notion of actual cause, as defined and motivated in a companion article. Essentially, an explanation is a fact that is not known for certain but, if found to be true, would constitute an actual cause of the fact to be explained, regardless of the agent's initial uncertainty. We show that the definition handles well a number of problematic examples from the literature. 1. Introduction2. Causal models and the definition of actual causality: a review 2.1Causal models 2.2Syntax and semantics 2.3The definition of cause3. Explanation: the basic definition4. Partial explanations and explanatory power5. The general definition6. Discussion Introduction Causal models and the definition of actual causality: a review 2.1Causal models 2.2Syntax and semantics 2.3The definition of cause 2.1Causal models 2.2Syntax and semantics 2.3The definition of cause Explanation: the basic definition Partial explanations and explanatory power The general definition Discussion Appendix: the formal definition of causality

Subfields: Artificial Intelligence, Statistics and Probability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Statistics and Probability

Citations in the corpus

Explanation in Human-AI Systems: A Literature Meta-Review, Synopsis of Key Ideas and Publications, and Bibliography for Explainable AI (2019, background): “…to, and perhaps even the same as causal reasoning, since explanations often refer to causation or causal mechanisms, and causal analyses are believed to have explanatory value (Cohen, 2008; Halpern and Pearl, 2005a,b; Krull and Anderson, 1997; Pearl, 1988; Pearl and Mackenzie, 2018; Trout, 2002).”
Explaining Classifiers with Causal Concept Effect (CaCE) (2019, background): “[6] give a formal causal theory of what constitutes an explanation, in terms of what is known as “actual causality”.”
Varieties of Causal Intervention (2004, methodology): “While such a simple model of interaction may be useful in untangling some of the mysteries of causation (e.g., it may have been useful in guiding intuitions in Halpern and Pearl’s study of token causation, [6, 7 ]), it clearly will not do for a general analysis.”
Intuitive Theories of Mind: A Rational Approach to False Belief (2005, unknown): “(See Halpern and Pearl (2001) for a related approach.)”
Sequences of Mechanisms for Causal Reasoning in Artificial Intelligence (2013, unknown): “For example, the method of [Halpern and Pearl(2005)] can generate explanations, but only once a “complete” causal graph enumerating all possible causal ancestors of an event is speciﬁed.”
Explainable Reinforcement Learning Through a Causal Lens (2019, methodology): “For a more complete review of SCM’s we direct the reader to (Halpern and Pearl 2005).”
A Rational Analysis of Rule-Based Concept Learning (2008, background): “(A naive interpretation of the implication terms as material implications would lack causal meaning, but an intervention-based semantics can be given following Halpern and Pearl (2001).)”
Causal Explanation Under Indeterminism: A Sampling Approach (2016, methodology): “Halpern and Pearl (2005a; 2005b) introduced a method (henceforth “HP”) for explanation that links structural equation models (SEMs) and counterfactuals.”