It turns out that in virtually every Torts case once the standard of liability has been established, you have to ask whether the conduct of the defendant is sufficiently connected up with the loss of the plaintiff, that it's appropriate to charge him with losses. There are two ways to do this. The modern way starts with very remote causes and then tries to narrow the situation down. The traditional way starts with immediate causes and then tries to move the system backwards. The modern intellectual approach to the subject is everything about proximate causation cannot be reduced to principles, it's all ad-hoc and it's a matter of social policies to whom would be rather held responsible to whom. Loss redistribution, deciding ways to milk somebody with a deep pocket and the like. And if you switch to that type of situation, essentially the Tort Law becomes unraveled because people are now held responsible for what they did not do, which means that they will have to invest money to prevent injuries when somebody else in the chain of activities is in a much better position to prevent the losses in question. If you're actually trying to try a case before a jury, using the traditional methods it's much clearer, and if you use the modern methods, what happens is the jury simply does not know what it means to talk about necessary and sufficient causes of harm. So starting tight and going broad is the way it's done. And once you do that, what you're looking for, it's not whether injuries are quote, unquote "reasonably foreseeable" you're trying to ask whether this some event which intervenes between what the defendant has done and what has happened to the plaintiff. And if in fact you have clear conceptions of causation, you cannot sort manufacturing, liability and remote partes like those people who paved roads and so forth for the harms that are caused by others.