Is Rawls’ version of "Reflective Equilibrium" a defensible method of theorizing about social justice?

When John rawls publishes his Theory of justice in 1971, he is the first to use the expression "reflective equilibrium" (RE). Indeed, even if this kind of idea had been employed before by Nelson Goodman (Goodman, 1955), rawls employs it as his method to build a complete theory of the justice and sort out this concept which is now subject to many debates. We won't come back on the complete definition of the reflective equilibrium, but we can remember the definition given by rawls himself: "It is an equilibrium because at last our principles and [considered] judgement coincide; and it is reflective since we know to what principles our judgements conform and the premises of their derivation" (rawls, 1971). We won't come back on the problem of the wide and narrow equilibrium either, following Norman Daniels who considers essentially the wide reflective equilibrium (Daniels, 1996, chap 1).
We will weigh up here that the RE is a general and important question, by trying to involve both an inductive and deductive approach as a method to build a theory in social justice, which is a particular subject of the philosophy. Indeed, by speaking about "social justice", we are in the "field" of the political issues and of human being. Considering that the method is a key point in a theory, it becomes crucial to try to validate or not the RE. In other words, how can the RE answer the problem of constructing a moral theory applying to a human society?
For this purpose, we will try here to show how we can justify the RE, the logic that can lead to its use. We are conscious that this way of proceeding can forget some problems that a critical approach would have pointed out more, but we aim here at finding a way to justify a theory of social justice.
We will start by studying the particularity of ethics and especially justice as a political process, to define some characteristics necessary to build a theory in this area. We will then try to find what method could fulfil the criteria we defined first and any possible limits.