Philosophical thought experiments

When I bring up Objectivism in online discussions sometimes people refer me to an article by Michael Huemer. This article is trash for several reasons. For example, Huemer rarely provides any references so that the reader could check whether he is correctly describing the position he is criticising. Huemer’s article is an obstacle to honest discussion. But I want to discuss something else about the article that seems badly problematic and is typical of the sort of crap many philosophers try to get away with: philosophical “thought experiments”.

This thought experiment is supposedly a criticism of rational selfishness. He describes a society in which you’re allowed to kill homeless people by law and custom. And a homeless person gets in your way and this inconveniences you. If you kill him he will be out of your way and you won’t suffer any adverse consequences. (In reality, killing somebody isn’t as easy as just stepping around him. So this thought experiment is already ridiculous but there is another problem, explained below.) Huemer claims that it is rationally selfish to murder the homeless man. He continues: 

That is, the egoist could argue that for some reason, it was really not in my interests to destroy the homeless person. You never know when a person, presently homeless, might become useful, after all. Some day in the future, for example, he might get a job, and then he might possibly work for my company, or be a client, or otherwise contribute to the economy of my society. Or he might someday be able to be an organ donor, if not for my destroying his body…
But even if the egoist is able to think of some very plausible harm that I would be likely to suffer from killing another person, I willjust modify the example to remove it. In other words, I stipulate thatthe homeless guy is not a potential client of my company, he is notgoing to get a job, he does not have a gang of friends to defend him, the passers-by on the street will not be angry with me, etc. And the question is, then does it seem that it’s right to kill him?
Many Objectivists misunderstand the way hypothetical counter-examples work. The point is to test a general principle: “The only thing that ought to matter to me, is what promotes my own good.” One tries to test this by imagining a specific situation in which an actionpromotes my own good, but it goes against some other thing that is often held to be valuable. The creator of the counter-example gets to stipulate what goes on in the example. So I get to stipulate, by fiat, that, in the hypothetical situation, I do not receive reprisals for my action, et cetera. The only thing that I do not get to stipulate is the verdict on the example, i.e., would the action thus described be right or wrong. That is where the reader or listener is supposed to exercise his own judgement. If the hypothetical action I describe seems to you to be morally right, then my argument has failed. If it seems to you to be morally wrong, however, then it shows that you are not truly an ethical egoist.

This thought experiment postulates something that is impossible. There is no way for you to know the homeless man’s future, even in principle. To do that you would have to be able to predict what knowledge he will create in the future. But he will interact with people who will interact with you, directly or indirectly, so to predict his knowledge growth you would have to be able to predict your own. But you can’t predict your own future knowledge because then you would already have that knowledge. There is no way out of this objection. If you can create new knowledge, then you can’t know the content of your future knowledge, so either you totally reject the growth of knowledge, or you reject this thought experiment. Huemer could have understood this if he actually thought seriously about what he was stipulating, but he didn’t.