But the details matter with respect to the meaning of the word "active". Pushing someone off a bridge and watching them die is very different from flipping a switch from the "five people die" to the "one person dies" position.

You are seated at a console. In front of you is a button. If you don't push the button, five randomly selected people will die. If you do push the button, only one randomly selected person will die. That is all you know. Do you push the button?

Or if that still feels like a moral dilemma to you, try this:

You are seated at a console. In front of you is a button. If you push the button you will save the lives of four randomly selected people who would otherwise die. Do you push the button?

Instead, of "you must flip the switch" you get a speeding train. Instead of the switch you get fat people and bridges. Are the five people children and does the fat person have heart trouble? Is the fat person fat because he's greedy, or does he have a glandular problem. Will anyone see me pushing the fat man? And again, what guarantee do I have that a fat man's body will stop a train? How old is the fat man? How young are the kids? Are there suspicious looking people around who, if the train fails to run over the five people, will just back the train up and try again?

But you're right, four versus none is not the same as five versus one. And five versus one in the abstract is not the same as these five versus that one. The details matter. That is the point I was trying to make.

If that's the question you want to ask, then the problem is badly under-specified.

> "out of bounds"

Including the prior probability that the situation can even happen is important information; if a specific value for this probability was desired, the problem should have included it.

> dismisses the spirit of the problem

Quite a few people in this thread seem are dismissing the Bayesian solution even though it does resolve an ethical dilemma.

Is it better to take an unethical action to prevent a greater tragedy? If the situation is very unusual and easily-confused human senses are a more probable explanation, then the ethical response is to recognize your own limitations.

Leaving out these details results in a model that won't actually match anything in the real world.

Yes. It's not supposed to.

It doesn't attempt to model the world accurately, it attempts to make concrete understanding from abstract principles. Unlike the sciences, the experiments are not designed to conclusive, they're designed to be instructive.

One could take issue with philosophy as a discipline for working this way, but "them's the rules."

"You are suddenly transported into a magical portal. You arrive in the home of a powerful magical creature. It explains to you what has happened, and that you are not crazy. It gives you as long as you like to investigate this and convince yourself that it's right. It tells you that it has a one time chance to transport you to a parallel universe for a few minutes. It has never happened before, and it will never happen again, and you will never be able to prove it happened. Nobody in the parallel universe except those directly involved will believe it either. You are then transported to the parallel universe, and the trolley problem/prisoners dilemma happens. Then you are immediately returned home and the magical creature vanishes, never to interact with our universe again."

Under those conditions, I would take the utilitarian solution. In the real word we have unintended consequences, and utilitarian solutions are uncomputable. Deontology is at least fair and predictable.

Ignoring is a mild way to put it, actually, because all of their experiments involve doing multiple additions and subtractions and substitutions. For them, morality is definitely a real number.

I think the complexity of these hypotheticals is entirely a result of a need for varied responses. The real answer is that there's not enough information given to make a determination. In the case of your hypothetical, there's clearly enough information (because one of the premises is that you have all possible information.) The question for the subject becomes easier, and the tough question becomes "Why is this horrible magical creature doing this, and how can we make it stop?"

That may be what it is designed to do. But maybe the design is flawed, and it doesn't actually test that? Maybe when you ask random people on the street (or undergrads or whatever) this question, all those details that you don't want to matter, turn out to have a big effect.