Rawls desperately wants the flexibility that comes with John Stuart Mill’s utilitarianism, but at the same time wants a moral philosophy that respects persons. Moreover, this philosophy must be neutral and devoid of controversial language, so he adopts part of Kant’s Categorical Imperative that is, treating persons as ends in themselves, rather than as means to some other end. His doctrine is artificial, however, since he views justice as political rather than metaphysical.
An injustice is tolerable to avoid greater injustice. He wants to further extend social contract theory in the tradition of Locke, Rousseau, and Kant with his famous thought experiment. Antony Flew has criticized Rawls for his question-begging assumption that the resources that exist already exist as property of the collective, so essentially Rawls’ argument is circular, since it assumes what it sets out to prove.
This social contract experiment is not a literal one. He wants to establish the actions of free and equal persons and what they would do under particular circumstances. Parties are rational and mutually disinterested, but not egoists.
Why he would want to ascribe a morally charged word like ‘fair’ to this thought experiment is puzzling. I suppose if I wanted to avoid being on the lower rung of society from a purely rational perspective, then there’s a chance I would suggest Rawls’ solution. However, just because it is rational for me to choose X over Y and just because that obtains for everyone else too does not necessarily make it ‘fair’. In fact, Nozick could readily agree to the rationality behind it, but still call the distribution of wealth unfair, since it violates his own theory of justice.