It just occurred to me by reading discussions regarding distributive justice and luck-egalitarianism, that a big source of problems for political philosophy is that devising rules for the organization of political societies requires to regard human beings in two different perspectives.
Think about Rawls' argument against equality and for the difference principle (which has been famously attacked by G. Cohen). Rawls argues that differences in income are required to make people willing to work more. They are treated as incentives, and when the effect of the existence of incentives on production goes to the advantage of the worst off class in society (which might happen through a suitably revised redistribution scheme) income differences are justified.
Cohen's famous remark is that people are responsible for their choices, and therefore they don't have to expect an incentive. People should work more for moral reasons.
Now it is striking that Rawls and Cohen place human beings in two radically different perspectives. When Rawls thinks of incentives he treats human psychology as a given. One might express this by saying that his view of people's psychology is (in that respect) instrumental. The incentive argument treats individuals as means not as ends, in so far as they are provided incentives in order to elicit a behavior which will be used for a goal - the good of the worst off class in society - which is not their goal in that behavior. And why can we say that it is not their goal? Because if the good of the worst off class in society were the goal for which they were doing the extra hours of work, they would not need an incentive to do that in the first place. But what is most important is that here Rawls is looking at social realities as formed by individuals regarded in a "third-person" perspective, as if they are natural objects, whose behavior is (partly) explained by certain natural (psychological) laws.
In Cohen's reply, on the contrary, human beings are regarded as free wills. Cohen's expects people to work extra hours for an egalitarian ethos. Cohens thus treats people as free will, as agents which are responsible for the choices they make and the reasons they have.
So maybe this indicates why political philosophy is so difficult and there are so many different schools, methodologies, and ways to understand and explaining them. Because political philosophy wants to address us simultaneously as moral agents and as agents who are victims of their own psychology. It has to do the former if it wants to be taken seriously at the normative level (the normative level requires criticism of existing customs and moral psychology), and it has to do the latter if its prescriptions want to apply to the social world (which is, after all, made by humans in flesh and bones, not by pure rational, noumenal actors).