A while back, reader David Cheifetz assigned the Pithlord a number of articles on the subject of the Supreme Court's "causation" jurisprudence, and why it sucks. I haven't said anything about it since, so I wanted to get my excuses out there.
If you want to sue someone about a car accident, it isn't enough that you are hurt and that they were driving badly. Their car also has to hit you, and cause your injuries. Most times, we get by with the "but for" test: the injury can only be compensable if it wouldn't have happened without the accident. But there is a more exotic form of legal causation, "material contribution." The theme of the articles Cheifetz sent me is that the "but for" test is just fine, and the Court's varying statements about when other forms of causation should be applied is confusing and confused.
I'm not totally sure. Humans are given intuitions about countefactuals by divine providence or natural selection, but they are very hard to get a more scientific grip on. That means that we can be sensible about what would have happened when the events are the kind of things we have pre-scientific intuitions about, but breaks down when the white lab coats get involved. Sure, there are some forensic experts who don't mind opinionating on what would have happened, but they shouldn't necessarily be trusted.
If it is granted that there is a genuine practical problem here, what is the solution? I'd say it isn't a better verbal formula for what causation really means. It is a more concrete sense of what kind of evidence needs to be called to get to liability in various contexts.
Which gets me out of the uncomfortable world of counterfactuals and into my favourite legal process point. People in general are better at deciding what should be done in particular situations than stating coherently the general principles behind them. Judges are people. So we should expect that they get cases right more often than doctrine. It follows that they might have avoided Mr. Cheifetz's well-founded criticisms by being a bit quieter. The only time it is legitimate to talk about how evidence that doesn't support an inference of "but for" causation should result in liability, is when the court decides that evidence that doesn't support an inference of "but for" should result in liability. Then the facts of the case gives guidance as to what the hell the abstract language is supposed to mean. On the other hand, if there are well-understood reasons to reject liability anyway, doctrinal language is just going to confuse everybody.
So, in Resurfice, there was no need to throw out incomprehensible dicta.There was nothing wrong with zamboni, its design didn't increase the risk of the accident, and that's all that had to be said.