My response here is also informed by *too brief* reading of his article "Relativism, Reflective Quilibrium, and Justice" - which, happily, in one way or another raises questions about all of the points that we're discussing right now. So my rejoinder is tailored to this article, with consideration of the concerns aforementioned.
The problem, as Justin puts it, is whether relativism can be avoided without denying the reality of interest-based dissension over principles of justice. His suggestion is that to a great extent it can, although not in a way that vindicates liberal neutrality by producing consensual principles of justice. This formulation leads to the next one: Can choice of principles of justice be nonarbitrary, and their justification objective, without their being agree on, actually or hypothetically, by all rational persons?
A Habermasian response is affirmative, although with cautions.
But let me reconstruct the question. Can principles of justice be motivated (nonarbitrary and partial) and impartial (legitimate), without an actual agreement having taken place?
To put it delicately, yes and no.
Yes, in the sense that in order for principles of justice to be impartial, and therefore "just" or "justified" - there must be an actual agreement. This cannot take a hypothetical form.
No, in the sense that we can reconstruct the conditions under which all rational agents *could* agree to such principles hypothetically. So we can spell out the conditions under which principles could be justified.
It is, in practice, impossible to organize a global consensus. The expectation of an actual agreement is completely unrealistic. But we can create conditions, more or less, whereby we can anticipate consensus on certain issues. Participants in this limited discussion, then, must be conscious of the *fallible* nature of any conclusions reached. In principle and in fact, the 'conversation' *must* remain opened ended. There can be no "final" or "absolute" justification of any norms.
Impartiality: impartiality is not the same as neutrality. Impartiality has to do with a generalized expression of a common will ('collective will formation') and not with an objective justification. Habermas rejects the idea of an objective justification of either facts or norms - which is why he couches his point in terms of validity. This is a minor point.
Justification: the procedure of justification has to do with the process of coming to agreement about the validity of norms, not their 'application' in the real world. Discourses regarding justification must be separated from discourses of application. So, we are only talking here about the justification of moral principles, as abstract guiding principles.
We can keep in mind here the separation of three things: the legislature, the judiciary, and the executive. The legislature is only interested in the validity of legal norms. It may turn out that these norms are impossible 'in practice' but this would be a different question which entails feedback from the judiciary and the executive.
Habermas asks: what kind of conditions are required in order for a norm to be considered justified? His answer is simple: a norm can be considered just if, and only if, all participants agree to its validity.
The problem: how do we get to those conditions?
His answer: a norm can only be justified through an impartial assessment.
The problem: what is an impartial assessment?
His answer: an assessment is impartial only if all the interests of the participants in a dialogue have been taken into consideration and all participants actually agree on the validity of one particular norm.
The problem: this is a mess. Given the plurality of interests, how can we reasonably expect that these interests are not mutually exclusive?
His answer: there are different 'levels' of interest: universal and particular. Universal interests pertain to interests that all could agree to, particular interests pertain to interests that are non-generalizable (ie. partial and impartial, common and particular).
The problem: how do we sort these out without siding with one set of particular interests as a mechanism of sorting out these interests. In other words: what guarantees the impartiality of the procedure itself.
His answer: the impartiality of the procedure takes place through reciprocal role-taking. This depends on the capacity of the participants to say 'yes' or 'no' to representations or formulations of their perspective from a second person point of view (there is no 'third person perspective' here to speak of, which is why it is necessary to think about all of this in terms of first and second person - there is no 'objectivity' of perspectives).
The problem: do we have any guarantee that reciprocal role-taking is possible?
Habermas's answer: yes. Studies in cognitive development indicate the reciprocal role taking is possible, but not at a preconventional or conventional level. Only at the level of post-conventional reflection do we have the capacity to 'distance' ourselves from egoistic interests. This question can be decided on the level of empirical analysis.
The problem: how do we move from particular interests to general interests.
Habermas's response: reason giving and taking. We construct arguments for our positions, yes, this is a good norm because it suits my interests, no this is a bad norm because it doesn't suit my interests.
The problem: wouldn't this just amount to a back and forth of irreconcilable differences.
Habermas's response: not necessarily. Some interests are in fact in the interest of all - two such interests can be empirically verified: an interest in instrumental control over the environment and an interest in mutual understanding. In any event - we cannot decide beforehand whether or not there are unreconcilable differences. This point trumps the relativism position, which adheres to the idea of irreconcilable differences before acquire knowledge about differences from a first and second person perspective.
...
I'll stop there for now... ken