Rawls first discusses justice in relation to fairness in his paper "Justice As Fairness" dating from 1958 and published in his Collected Papers. In this paper Rawls describes fairness as the "fundamental idea in the concept of justice" and uses this to argue first for a correction in the idea of utilitarianism and secondly to argue for a certain problem in the general sense of utilitarianism. But in this posting I am going to leave aside the discussion of utilitarianism entirely to just concentrate on the sense of "fairness" here and how it is connected to justice.
Firstly, following the earlier paper on rules, Rawls describes justice as a "virtue of social institutions" or of practices. So justice cannot be rendered according to what he earlier described as a "summary" view of rules. Justice relates to a set of practices concerned with positions and offices and relates to these powers and liabilities, rights and duties. This first step is important in following on the work of the earlier paper and defining thus the types of rules that can count as rules of justice.
The second point is the development of two principles that define just practices, namely that those engaging in them have equal right to the most extensive liberty compatible with a like liberty to all (so a principle of equal liberty) and that inequalities are disallowed unless it is reasonable to suppose that they will work for everyone's advantage. We can see these two principles as ancestors of the eventual principles of A Theory of Justice. (A proviso is also added to the first principle to the effect that nations, firms, churches etc could, in addition to human individuals, be counted as "persons".)
The first principle has a ceteris paribus qualification added to it. There has to be a justification for departure from the basic given of equal liberty and similar cases, as defined by the practice, should be treated in a similar way. The view of inequality expressed in the second principle is attached to the "benefits and burdens" attaching either directly or indirectly to practices and it is required that inequalities be to the advantage of all parties. This principle thus moves away from utilitarian considerations.
After the statement of the two principles Rawls also now gives a first resort to a scenario like that specified later in the "original position" and which indicates, as that would, that people placed in the appropriate situation would select the two principles already specified. It is after the discussion of this device that Rawls begins to turn to the understanding of "fairness". The first point is the following one:
The question of fairness arises when free persons, who have no authority over one another, are engaging in a joint activity and among themselves settling or acknowledging the rules which define it and which determine the respective shares in its benefits and burdens. A practice will strike the parties as fair if none feels that, by participating in it, they or any of the others are taken advantage of, or forced to give in to claims which they do not regard as legitimate.
There are quite a few points in these couple of sentences. The first point to note that is that questions of fairness are here understood to be connected to the relations between free persons in which there is no established hierarchy required. So if I have no special authority over you and vice versa we encounter each other roughly in a position of equality and then we first speak to each other about what will count as "fair". So fairness does not, at least originally, have to concern situations of hierarchy. It instead concerns arrangements between persons who are presumptively equal to each other.
This presumptive equality is only the first part of the situation of fairness. There is also a need to see participants as engaged with each other and as being involved in practices in common. It is with these practices in common that the question of what "benefits and burdens" arise comes out and it is in connection with the share of these amongst those engaged commonly together that we speak of "fairness". If the practice in which we are engaged leads some to be taken advantage of by others they will declare it "unfair" so the notion of "fairness" then is bound up with the sense of the distribution of these benefits and burdens. So "fairness" is a matter of public distribution of "benefits and burdens" in a shared practice. What follows from this is a general sense is at issue of what claims are legitimate within the context of this practice. These legitimate claims require to be understood as ones that others could reasonably be expected to accept. Rawls sums up:
It is this notion of the possibility of mutual acknowledgement of principles by free persons who have no authority over one another which makes the concept of fairness fundamental to justice. Only if such acknowledgement is possible can there be true community between persons in their common practices; otherwise their relations will appear to them as founded to some extent on force.
So "fairness" requires this "mutual acknowledgement of principles", or, in other words, it requires publicity. It is necessary, for a practice to meet the standard of "fairness", that we each can freely acknowledge the basis of the principles underpinning it. If we cannot freely acknowledge them then we are being coerced.
Rawls goes on to mention the distinction, taken from "ordinary speech" of the difference between institutions that manifest "fairness" and those to which we think questions of "justice" arise. So business competition includes rules of fair trading and breaking these rules is to engage in certain kinds of crime because, as Adam Smith once put, we thereby engage in conspiracy against the public. By contrast, practices in which people are compelled in certain ways in a direct sense (as e.g. in apartheid regimes) we speak of as being in situations of injustice and we have a notion of justice to hand that requires arriving at practices that avoid such direct compulsion being applied. What this shows is that fairness is generally understood as a criteria within a practice whilst justice is often a notion applied to the ways certain types of practices run in general.
The rules of the practice in general set the standards for what is "fair" and given acceptance of these rules there exists a prima facie duty to act in accord with the rules and Rawls terms this the duty of "fair play". Evidently failure to accord with this duty has a range of ways of being manifested. There can be "cheating" which involves deliberate infractions of the rules but there can also be exploitation of loopholes, making use of special circumstances to avoid the rules, making exceptions that suit one's own case and generally acting, as we put it, in violation of the "spirit" of the practice. This is the basis for Rawls' use of the term "virtue" in relation to justice since the notion of this "fair play" involves a sense of acknowledgement of the constraint of fairness on one's conduct, an acknowledgement that we can term its internalisation and which leads us to speak of someone as being fair (or fair-minded). Those who lack this sense of the practice are understood as "free riders" and they pose certain special problems. Chief amongst these problems is the basic sense that accepting the constraint involved in fairness is part of what we understand engaging with others as persons to involve (so "free riders" are engaging in conduct of a kind that crosses the boundary of acceptance of them as behaving in a way befitting a person).
One of the benefits of this discussion is the sense it gives to thinking of justice as part of a view of morality though it is less obvious how it maintains the understanding of law as still something different from a moral rule. However, in leading to a conception of "fairness" as part of a sense of recognition of personhood it does embody a view of institutions that whilst tending to the ideal still gives a standard for participation in them which has critical edge.