• relationship where one event (cause) produces another (effect)
  • requires temporal ordering: cause precedes effect
  • requires a mechanism or pathway linking cause to effect
  • distinct from correlation: co-occurrence alone does not establish causation
  • Hume’s criteria: contiguity, succession, constant conjunction
  • Pearl’s do-calculus and directed acyclic graphs formalize causal inference
  • counterfactual test: if the cause had been absent, the effect would not have occurred
  • central to science, law, medicine, and engineering: identifying causes enables intervention
  • confounding variables, reverse causation, and selection bias obscure causal relationships