I first wanted to discuss the idea that theft is not, by definition, always wrong, contrary to what Jad suggested in his comments. In moral philosophy we draw a distinction between things that are wrong and things that are prima facie wrong (or "wrong at first glace"). Basically the application of the idea is this: As we've already discussed, the concept of theft comes from a moral framework which recognizes that people can have a moral claim to certain property holdings. Within this framework, to steal something is, by the nature of the system we're considering, prima facie wrong: that is, it's wrong at first glance. To put that another way, a system which recognizes rights to property is, I think necessarily, built on the idea that it's wrong to steal capriciously, or under circumstances which could be thought of as more or less "normal."
But a number of perspectives on property rights stop short of saying that it's never permissible to steal. For example (and this example is undoubtedly a complete bastardization of medical fact; I'm sure it couldn't actually happen this way), let's say you felt chest pains while walking down the street, and knew (somehow) that you were starting to have a heart attack. You happened to be passing by the open window of a house where some aspirin was sitting on the table, but no one was home. Now, let's say (probably falsely--I don't know) that if you took the aspirin, it would increase your chances of surviving the heart attack significantly. Would it be wrong to steal it? I don't think that the answer is clearly yes, or that by saying no we would be denying any form of property rights. We would just need to say that property rights aren't absolute like that; they don't tell us what we always must do or must refrain from doing.
If we accept the moral legitimacy of property rights of some form, then we would say that you're normally prohibited from stealing; to do so would be prima facie wrong. But in some circumstances, like where your life is in danger, the normal rules might not apply. That's not necessarily because the rules are bad ones, but rather because the rules were never meant to apply in all situations and contexts.
That's a big part of ethics that Stefan seems to completely ignore, and I noted it as my second objection to his book. He seems to believe that ethical rules need to be rigid and absolute in order to be legitimate. But not only does this not need to be the case; it's generally acknowledged to not be the case. I think that most people, including those in the philosophical community, don't think that you should be obligated to let yourself die in order to avoid taking someone else's asprin. So it seems like we would want to be skeptical of a system which would tell us that property rights commit us to saying that it would be wrong to do that, or otherwise that we must accept the position that it would be morally obligatory to constantly go around looking for aspirin to steal. But that's precisely the sort of thing Stefan needs to say in order to make his system work, and look at all the trouble he had trying to twist himself into plausible positions on "lifeboat scenarios." The reality is: rigid and absolute rules don't usually work, and when they do, it's because they've defined themselves in circles.
But your other question, which I think is important, is how we can decide what kinds of circumstances are morally significant, and which ones are simply irrelevant to determining what would be right or wrong. You suggest that the mere fact that someone has an FBI badge shouldn't matter in determining what's right and wrong in a certain scenario, and I don't want to dispute that outright (though there are important reasons why that might be problematic, and we can discuss those later if you want to). What I do want to point out is that you've now moved into the realm of axiology (that is, the study of value, or more specifically for our purposes, the study of what kinds of things have moral weight and why). Once we enter into this area of thinking, consensus among philosophers breaks down extremely quickly, and so I'm not particularly eager to critique Stefan's views through an axiological alternative of my own (of course, we can discuss my own views on the grounding principles of ethics later if you want). But I will note that Stefan doesn't actually provide a rigorous axiology in his book (this was my final objection in the other post), except to say that we should come to an understanding of ethical principles through the scientific method.
That brings us to Jad's final line of questioning, which concerns whether the scientific method is useful in determining what moral code is the correct one. Jad writes:
...is it not reasonable to begin with a system isomorphic to the scientific method? I find it appealing on every level (I know that's not relevant to its validity). The amount of human suffering that is let pass by ethical systems which weigh intangible concepts (usually concepts that favor the powerful and the aggressive) is staggering. Is it only coincidental that an ethical framework grounded in reason and evidence favors instead voluntary relationships and non-aggression? Is it coincidental that civil society, as a matter of course, follows the single maxim of a framework so grounded? Is it coincidence that the only violators are those that interact with strangers by killing or threatening to imprison/kill them?
But there's a problem with this line of reasoning that I think Jad picked up on a little bit when he noted that "...the validity of the statement isn't determined by the appeal of the outcome." The problem can be understood in two parts: First, there is no such thing as "The scientific method." Science is done through testing hypotheses, through clarifying the meanings of concepts and exploring their entailments, through abstracting from data sets, and probably through a number of other means I'm leaving out. With each approach to science, there is undoubtedly a different idea of what it would mean to try to learn about morality through the scientific method, and evaluating the possibilities available through each method would require that we at least specify which one we're talking about. So simply saying that morality might be aided by the scientific method doesn't mean anything. You need to point to a kind of scientific inquiry you're hoping to use, and look for some way that it could be used to learn about morality.
But the second problem with the line of reasoning is that you need to figure out what sort of thing you're looking for in reality which would lead you to some conclusion about morality. Jad is absolutely right to point out that the examples he cited are outcomes, but further, they're not outcomes of morality, as such, but rather of individuals performing certain actions. So in order to learn anything about morality, he'd need to postulate some connection between morality and real world events which is not already clearly defined.
The idea he seems to be leaning on is that an acceptible moral framework should lead individuals to do things which will produce beneficial outcomes. This idea is not new to moral philosophy, and is most closely tied to utilitarian and consequentialist schools of thought. But it need not be solely those schools which can lay claim to those notions. Most moral philosophers these days acknowledge that there must be some place for consequences in our evaluation of a moral theory; if a moral theory leads to horrible social outcomes, that might very well be seen as evidence for rejecting it, or at least for accepting an alternative view with equal plausibility (see, for example, Schmidtz in Elements of Justice). So if consequences are a factor by which we judge moral theories, and we can predict what consequences we will likely observe if we follow this moral theory or the other one, then we can compare them somehow and perhaps move towards a better understanding of what our position ought to end up being. In that sense, various approaches to science might prove very useful.
But Stefan cuts himself off from this avenue of reasoning when he seems to repeatedly argue that consequences don't have anything to do with what's right. Perhaps things that are right happen to produce good consequences, but part of his position seems to be that if this were not the case, it wouldn't matter in determining the correct moral framework. So hence my confusion about Stefan's invocation of the scientific method as the sole hint to his axiology. Most of the evidence I would expect someone to use in favor of their theory won't work for Stefan, because he's thrown it out. He can't appeal to what people generally think or what consequences would be produced through acceptance of his view. And he's seemingly ruled out the idea that morality is a social or psychological phenomenon by attempting to prove a realist ethics. So I guess I'm just not sure what he wants science to tell us.
Hopefully some of that is helpful. As always, keep the questions coming!