Skip to content
Coordinates of ThoughtHome of Philosophy
Join
Discussion

“But for the defendant’s act”: thirty-one years of counterfactuals from the bench

Khalid Mahmood Awan Fellow
· 5 replies · 192 views
LawAnalyticDavid LewisH. L. A. HartRobert Stalnaker#but-for-test#counterfactuals#legal-causation#overdetermination
Follow

For three decades I wrote judgments that turned on sentences of the form "had the accused not done X, the deceased would not have died". I never asked what such a sentence is. It cannot be checked: the accused did do X. Yet we convicted and acquitted on it, and I believe mostly rightly.

Since retiring I have read that philosophers explain these sentences with possible worlds: the claim is true if, in the nearest world where X did not happen, the death did not happen either. I find this both illuminating and alarming. Illuminating, because "nearest" is exactly what a judge does: hold everything fixed except the act. Alarming, because who decides what is nearest? In court that was me, guided by precedent and common sense. I am not sure common sense is a theory of similarity between worlds.

Two cases troubled me in practice and I wonder what philosophers make of them:

  1. Two men independently stab the victim, each wound fatal on its own. "But for" either act, the victim still dies. Did neither cause the death?
  2. A man fails to repair his brakes; the brakes fail; but the road was blocked by a landslide anyway and he would have crashed. Is the omission a cause?

The law has its patches. I want to know whether the philosophy has better ones, or whether it has the same patches with more Greek letters.

LikeReply
Follow5 replies

Members reply here. Reading is always free.

Sign in to reply

5 replies

  1. Hamza Qureshi

    Fellow

    Sir, your two cases are exactly where the "but for" test breaks, and the law knows it. In the first, courts simply hold both liable, which is the right answer and the wrong test. Hart and Honoré's Causation in the Law was in part an attempt to explain how lawyers' common-sense causal judgments go beyond "but for", and one later refinement asks whether the act was a necessary element of a set of conditions sufficient for the harm. Each stab passes that test, though neither passes "but for".

    Helpful · 3
  2. Oskar Nyberg

    Fellow

    On the philosophy side: Lewis's original counterfactual analysis of causation had exactly your problem with the double stabbing, which philosophers call overdetermination (its cousin, pre-emption, gave him just as much trouble). He revised the theory more than once. Many people think he never fully fixed it.

    On "who decides what is nearest": Lewis admitted that similarity between worlds is vague and context-dependent, and he said that this vagueness matches the vagueness of counterfactuals themselves. That is either an honest feature or an admission that the semantics shifts the hard part back onto judgment, which is where you found it on the bench.

  3. Tomasz Wójcik

    Fellow

    One technical difference may interest you. Stalnaker assumed there is always a single nearest world; Lewis allowed ties. In your brakes case, ties seem real: is the nearest no-omission world one with or without the landslide? If the landslide happened anyway, any reasonable similarity keeps it, so the omission makes no difference and is not a "but for" cause. The law and the semantics agree there, I think.

    Helpful · 1
  4. Sana Mir

    Fellow

    As a court reporter I have watched these sentences get written. What strikes me is that the counterfactual is often decided by what the judge finds blameworthy rather than the other way around. A careless driver's omission gets counted as a cause; a pedestrian's ordinary walking does not, though both were "but for" conditions. Is that a corruption of causal reasoning or how causal talk actually works?

    Helpful · 2
  5. Khalid Mahmood Awan

    Fellow

    Sana, it is how it works, and Hart and Honoré said as much: out of all the conditions, the law selects as "the cause" a voluntary human act or an abnormal event. I used to think that was the law's peculiarity. Reading Oskar's account of Lewis and his context-dependent similarity, I now think it may be how everyone reasons, and judges merely do it on paper.

    So my answer to my own question: philosophy has the same patches, with fewer appeals to precedent and more honesty about the vagueness. That is not nothing.

    Helpful · 3