This experiment hasn't been done, to my knowledge, but how do you think milgram would have gone if they had begun the experiment by explaining the difference between ethical and unethical orders, and telling people they could be prosecuted for following the unethical ones?

Most normal people haven't thought that through, and are unprepared to make that kind of decision. But soldiers have. Like a doctor professionally worries about malpractice, a soldier spends his entire career thinking that through.

What Milgram measured was how inexperienced people react to a new ethical situation. If he had tested people who were familiar with the context--fellow scientists--I very much suspect the results would be different. That's true in many fields. You might be able to intimidate a junior engineer into including a substandard part in a safety-critical area, but a senior engineer will refuse. And then he'll quit. It's not that he has more character; he has more experience.

Soldiers have a lot of experience. If you order them to hurt people immorally, their reactions are more like the senior engineer than the junior one: measures of character, not measures of how quickly they can think about a new situation.

There have always been soldiers who said, at the eventual trial, that they were "just following orders". I think they're generally inexperienced, but I'm sure there are cowards who think that way at the end of a career, too. I'm sure the prevalence depends on culture and training, but that does show that the effect Milgram measured is present.

I do have some evidence that it isn't dominant, though. Historically speaking, conquering armies have behaved in conquered countries pretty much as they pleased--sometimes very badly. But to get a soldier to kill his own countrymen, it has generally been necessary to threaten him or his family with death -- mere orders have not been enough. The soldier has a limit on what he'll do for authority, and he'll quit sooner than Milgram's subjects did.

It takes courage to make a high stakes moral decision correctly; I won't claim that everyone has it. I don't know what the rate is, and I expect it varies widely by culture and training. But I do think that in order to measure it, you need to test professionals. Extrapolating from amateurs isn't good enough. Perhaps a repeat Milgram would be a start: test all the same subjects again a week later, after they've had plenty of time to think about the experience.

Milgram proves too much. It should prove those soldiers callous as well, and they are not. Heck, it should prove any soldier callous, and they certainly are not. As I explained above, soldiers have training the subjects in milgram didn't.