Along similar lines, the legal system ends up dealing with this as well. Because a crime is often defined both in terms of actions taken (e.g. "Person X stabbe

Along similar lines, the legal system ends up dealing with this as well. Because a crime is often defined both in terms of actions taken (e.g. "Person X stabbed person Y".) and in terms of outcomes (e.g. "Person Y died."), the legal system needs to connect the cause and the outcome. So, if X is a surgeon, and stabbing was part of an unsuccessful surgery trying to save Y's life, that wouldn't be murder. * X stabs Y, and Y immediately dies. Simplest case, X is responsible for Y's death. * X stabs Y as part of a surgery. Y was predicted to have 2 months to live, but instead dies on the operating table. X did not murder Y, even though they were the cause of Y's death occurring on that specific day, rather than 2 months later. * X stabs Y as part of a mugging, and Y immediately dies. Coincidentally, Y was predicted to have 2 months to live, but instead dies in the mugging. X did murder Y, even though it was the same difference in total lifespan as the previous example than 2 months later, because the intent matters. * X stabs Y as part of a mugging, but it only requires some light stitches. While Y is driving to the emergency room, they are killed in a car accident. X would be liable for causing injury, but wouldn't be liable for the death, because it was an unrelated coincidence. * X stabs Y, and Y later dies of blood loss at the hospital after refusing treatment for religious reasons. The defense argues that this should be counted as manslaughter instead of murder, because Y's refusal of treatment breaks the chain of causality between the action and the eventual outcome. The judge disagrees, as "defendants must take their victims as they find them." These are all hypotheticals based on my layman's understanding of the law, except for the last one, which occurred in the case of a Jehovah's Witness who refused blood transfusions for religious reasons [0]. This became more of a ramble than I intended, but I definitely agree that causation is tricky, and "cause" could be any of "action with a causal link", "immediately preceding cause", or "deviation from a just world that initiated a causal chain", and would lead to vastly different statistics depending on the definition. [0] https://www.bailii.org/ew/cases/EWCA/Crim/1975/3.html https://www.bailii.org/ew/cases/EWCA/Crim/1975/3.html