Some of my readers may be familiar with G. A. Cohen's paper, Facts and Principles. In it Cohen argues that facts can only serve as part of the justification for some action-guiding principle in light of some further principle in virtue of which they support the first principle. For example, if you ask why you should keep your promises, and I say, because it allows people to form and pursue plans, then I seem to need a further principle that we ought to do things which allow people to form and pursue plans in order to explain why that fact should play a role in the justification of the first principle, that promises should be kept. If however, the way in which facts play a role in the justification of action-guiding principles is dependent on there being some further principle which explains how the consideration the fact points to is relevant, then facts cannot by themselves justify any action-guiding principles; they must be supplemented by some principle, which, likewise, if it has a justification in which a fact plays a role, will therefore also have a justification in which a further principle explains the role that that fact plays. A regress is thereby generated, and facts cannot be the last items in a chain of justification, since their role in any justification itself requires justification in light of some principle.
This argument matters in political and moral philosophy because if it is correct, it may show (some versions of) a particular kind of political and moral theory to be incorrect, since that theory relies on certain facts to generate principles. It only may, because it seems possible that there is a principle that says that all further principles are to be generated in light of facts. The argument may of course be incorrect: it seems to appeal to a remarkably rigorous version of the fact-value distinction - is 'he was rude' a fact, and if it is, is it really neccessary for justifying censuring him afterwards to appeal to a principle 'censure rude people'? - and relies on a distinctly foundationalist notion of justification - why, when an appeal to some facts and principles is made in order to justify some other principle, should justification proceed in transitive, one-to-one relations, and not some kind of web, where the kind of regress and eventual reliance on bloody-minded insistence on self-evident principles that Cohen's model of justification must lead to can be avoided? What interests me right now, though, is a question about why it should be facts, and not falsehoods, that play this role in the justification of principles, although if I am right, it bears both of the just-mentioned issues about the correctness of Cohen's argument.
Clearly if keeping promises frustrated people's ability to form and pursue projects, then saying that appealing to the claim that keeping promises enabled people to form and pursue projects could not be justificatory of the principle that we ought to keep our promises. Why is that though? Presumably it is a fact that it is falsehoods cannot play a role in our justifications of principles. Then the question becomes what principle explains that fact. Any plausible candidate for such a principle looks like it just restates the fact, that falsehoods cannot play a role in our justifications of principles, because there seems to be nothing else to say about why it is crucial that premises in that kind of justification are true. If that is the case, though, that any putatively explanatory principle merely restates the fact that they are supposed to be explaining, then we have both an example of the breakdown of the fact-value distinction - the fact simply is normative - and a kind of example of justificatory holism - whatever it is that is important about facts and their role in justification, that cannot be captured by asking for one-to-one, directly transitive justifications. Of course, it is a while since I actually read Cohen's article, so it may be that he has an answer to this.