A counterfactual asks, after an outcome is known, whether the result would have changed if one condition had been different. It is not an unconstrained invention of another life, but a comparison disciplined by the facts.
Suppose a meeting is delayed because the papers arrive late. To test whether the late delivery was a cause, we should not imagine that everything changed. We hold other conditions as steady as possible and compare the case in which the papers arrived on time. If a power failure would have delayed the meeting anyway, the original causal judgement becomes weaker.
A counterfactual differs from a prediction: prediction concerns a future that has not happened, while a counterfactual reconstructs an alternative condition within an event that did happen. Nor is it the same as regret. Regret carries emotion and evaluation; counterfactual thinking can be purely analytical.
Its value is that it turns ‘what could have happened’ into a question we can examine. Yet the comparison always depends on what we keep fixed and what we alter. Counterfactuals can clarify causation, but they do not automatically produce one uniquely correct answer.
Discover more from Geoffrey Chen
Subscribe to get the latest posts sent to your email.