I have defined minimal justice as the conception of justice that will be chosen or endorsed under conditions of impartiality, and have given a very broad definition of impartiality, one that appears to be the common denominator of liberal conceptions of justice. In this chapter, I shall defend a more precise and concise notion of impartiality.
How should we conceive of impartiality, given the impossibility of ethical absolutism? It cannot be 'anything goes'; under such circumstances there would no longer be moral or immoral ideas-only evaluative or normative fancies. Allowing each and every evaluative view leaves us with no point of view at all, not even one from which we can account for the 'anything goes' rule itself. Because anything goes, nothing goes: both the idea that every moral or immoral or amoral code is allowed, and the idea that none can be allowed as none can be justified, would be equally valid or-since anything goes-equally invalid.
I shall argue instead that impartiality is a negative although not a nihilistic notion. We are morally and intellectually situated, like it or not. We all have an evaluative point of view, even Camus's étranger. The intention behind introducing impartiality in theories of justice is to step away, as far as possible, from these particular points of view, without excluding their existence or without denying partiality and contingency. 22
One way not to do this is by simply asking real human beings to be impartial and to judge principles of justice. We might trust them to be impartial, but in this respect it is more secure to control them. In this chapter I shall discuss the side-constraints that should be posed on a justification procedure to ensure its impartiality, and the form best fit for representing such a procedure. As to the first part, this comes down to defining impartiality as, among other things, the point of view (the archpoint) at which one postpones judgement on the most