David Aaronovitch opens some awkward philosophical questions here:
The man who really caused the death of Jean Charles de Menezes was
not the policeman who put the bullets in the poor Brazilian’s head but an
Ethiopian called Hussein Osman. It was Osman who, one day earlier, had tried
to blow up a train full of passengers at Shepherd’s Bush.... had Osman not existed, or else been content with
allowing his fellow citizens to exist in peace, Jean Charles de Menezes
would still be alive.
As a description of causality, this seems roughly tenable. The counterfactual theory says that if we want to know whether x causes y, one way to find out is to remove x and see whether y still happens. Remove Osman, and de Menezes would still be alive.
Equally, if the two officers who shot de Menezes were removed - say, they happened to be off duty that day - de Menezes would still be dead as some other officers would have shot him. In this sense, they did not cause his death.
So far, so good. But here's a problem. Consider a Palestinian suicide bomber in Israel. He could say:
If I weren't blowing up the bus,
someone else would; as Alan Kreuger tells us, there is an
excess supply of willing suicide bombers. In this sense, I don't cause the bus to blow up.
The true cause of the explosion is the existence of the Israeli state.
Remove Israel and there'd be no bomb.
This argument has the same logical structure as David's: the suicide bomber stands in for the policemen, and Israel for Osman.
And yet David, and his readers, would recoil from this.
One reason for this is that David omits something - responsibility. It would be stretching a point to say Osman was responsible for de Menezes' death. Sure, he meant to kill lots of people, but not in so circuitous a way as that. Equally, although the state of Israel might be a cause of Palestinian terrorism, it's not responsible for it - you can't be responsible for tenuous unforeseeable effects, can you?.
And both the suicide bomber and the officers who shot de Menezes are responsible, in the sense that they could have chosen not to detonate, not to fire.
Responsibility and causation are not the same.
For example, if I push you off the pavement into the road, you cause a cyclist to fall off his bike; had you not been there, he would have stayed on. But you're not responsible for his fall.
And it's possible, I suspect, for there to be responsibility without causation. In traditional firing squads, only one member of the squad was given a live bullet, and the others given blanks, without them knowing who got what. We couldn't, then, say who caused the object of their firing to die. But we could, reasonably, hold a volunteer for firing squad duty responsible for the man's death.
My only inference here is that causality and responsibility are different ideas. But I suspect these are not issues where careful distinctions are made.