One possibility is to have degrees of legitimacy. Compliance with the LPL would be a condition of full legitimacy, although other kinds of regimes can be more or less legitimate. So Rawls could think that my religious democracy is not fully legitimate, since it violates the LPL, but it is more legitimate than Kazanistan, since that is not democratic, and Kazanistan is more legitimate than a military dictatorship, which may have none at all, or some small measure if it is relatively benign. In this case, the moral obligations of the Kazanistan citizenry arise from the partial or imperfect legitimacy of the system.

Alternatively, a Rawlsian could think of legitimacy as a discrete notion, so that the Kazanistan regime has no right to rule as such, but nevertheless has sufficient other moral qualities to give rise to the moral obligations of the citizens. So the fact that it does implement a common good conception of justice, as opposed to a system of naked exploitation and oppression, may warrant compliance, or at least respectful and orderly disobedience.

I don’t know how to square Mandle’s take with the LPL. The relativism I worry about is that since one group of societies has a certain kind of history and cultural baggage, the LPL is appropriate there, whereas other kinds of societies are subject to different principles of legitimacy. So I agree that it either applies everywhere or nowhere (and I think nowhere). But I really can’t see how an illiberal and undemocratic regime could satisfy the LPL. In principle, there is a small gap between the LPL and secular democracy, but I think that gap is closed when we ask whether citizens could reasonably be expected to accept a system in which they are disenfranchised (albeit consulted) and governed in accordance with a religion they may reject. Perhaps in some exceptional circumstances, the answer would be yes, but surely not often enough to treat as a general rule. And then I don’t think there is much room to regard the objections of Kazanistan citizens to their regime as less reasonably compelling than the objections of the citizens of liberal democracy to the possibility of comprehensive moral legislation. Rawls would really have a lot of explaining to do if the LPL required the excision of Kantian or Millian morality from the exercise of basic political power in one part of the world, but did not require a similar excision of Islamic reasoning in another part of the world. That would have been a significant issue in the exposition of political liberalism. I don’t know how it could be supported without reintroducing relativism, this time about reasonableness.

So I think that Rawls advocates a single master principle of internal or domestic political legitimacy in the liberal principle. In effect, this liberal principle requires a secular democracy to be satisfied. But on his view that fact does not determine the standard of external or recognitional legitimacy. If an undemocratic or illiberal regime meets the standards of decency, then it has enough moral quality in whatever form to entitle it to recognition from the international community. This seems a plausible view to me. I just think it is too demanding in the domestic sphere and too weak in the international sphere, and that anyway we need to premise recognitional legitimacy on domestic legitimacy.

“(b) The second part is that a decent people’s system of law must be such as to impose bona fide moral duties and obligations (distinct from human rights) on all persons within the people’s territory…” (Rawls LP 65-66)

That doesn’t say only that the people ought to believe they are obligated. It says that in decent societies, people are obligated.

Mandle proposes an interpretation on which there is a non-relativistic principle behind this, according to which law is legitimate when it represents the collective decision of the society. This is an objective constraint that law in many societies would not meet. But he argues that on Rawls’s view societies count as decent partly because they do meet it. Democratic societies are more just, and Rawls allows that in practice democracy might be required to prevent abuse of human rights, but societies can, in principle, be legitimate without being democratic. And this is tied to their having a right to non-intervention.

I agree that it is very unclear how this relates to the “liberal principle of legitimacy.” That principle does not directly require liberal or democratic institutions, so that’s not the problem. It’s just not clear whether Rawls thinks merely decent societies could meet that principle, or whether that principle doesn’t apply to them. That would be an odd view. If a society doesn’t have to be liberal or democratic to be legitimate, it’s hard to see how even liberal democratic societies have to be liberal or democratic to be legitimate. So it wouldn’t apply to anyone if it doesn’t apply to everyone. So the other alternative is to say that it applies everywhere, but (in principle even if not in practice) a society could meet it even without being liberal or democratic. That, interestingly, would apply the concept of “reasonable comprehensive conceptions” in all societies, since that’s integral to the principle of legitimacy. It’s still somewhat puzzling, but this looks like the more plausible interpretation.

The Mandle paper is in this volume:
The political philosophy of cosmopolitanism By Gillian Brock, Harry Brighouse, Cambridge UP 2005

The link you included didn’t work. It said it couldn’t install a cookie, but my browser is set to allow it. It also looks like it might require e subscription or something.

I do think that legitimate governments have the right to rule in ways that are morally wrong, and that even violate the rights of their citizens. The example I use here is of a democracy (”Apostolica”) with an established religion that prohibits elective abortion and any form of assisted suicide, which I take to be inconsistent with what justice requires. Nevertheless, the government’s rule is legitimate because it is consistent with a basic democratic principle of political equality. The democratic legitimacy of its rule creates a right to recognition and to non-interference on the part of foreign governments with its practices. It would be wrong for a secular liberal state to impose sanctions on Apostolica on account of its abortion and euthanasia policies.

But I also think that a regime cannot have a right to recognition and non-interference on the part of foreign governments unless it is a legitimate democratic regime.* So I would want to run domestic and recognitional legitimacy together in a way that Rawls wants to keep apart. One of the reasons why he wants to keep them apart, I think, is that he has a pretty demanding conception of domestic legitimacy in the liberal principle. But I think he goes too far in the opposite direction when it comes to recognising foreign regimes as the representatives of decent peoples, as with Kazanistan. I don’t think we can settle questions about when regimes are entitled to recognition and non-interference by appeal to generic considerations of toleration. I think we ultimately have to appeal to a conception of when they actually do have the right to rule within the society. Otherwise I think we’ll forever be at sea with the hard question about where to draw the limits.

*Where a right to non-interference would be only one of the many moral considerations that bear on the question whether a regime ought to be recognised.

It is not too clear from your story, but I do worry about the bias of equating a state religion with the treatment of women as second class citizens or as effectively non-persons before the law. Few argue that the aforementioned treatment of women could warrant international sanctions but I have a hard time seeing how a society that does not treat women as genuinely persons differs from one that does not genuinely see blacks as persons such as apartheid South Africa. Organized collective non-cooperation enjoyed wide support in the South African case and I am not sure what kind of argument could show that the case is different in a morally relevant way when women are not treated as persons.

The neighbors would do better to just admit that it is worse to coerce the Kazan’s than it is to let the family to continue as it is. It is of course unfair that the Kazan kids suffer from their parents’ behavior, but the harms are just not serious enough to warrant the kinds of individual sanctions being proposed. In fact, the neighbors should also admit to themselves that the only reason they even considered fines, shaming, and non-cooperation is because it was a simple response that places blame squarely on someone specific bad person, Stan Kazan. The response is a way for the neighbors to avoid reflecting over reform of larger social conditions that could actually help but would also implicate the entire neighborhood. If the neighbors could admit this to themselves they might see the connection between the neighborhood’s culture of status consumption, long working hours, tax aversiveness and the Kazan’s division of labor. After all if the labor market was organized in such a way that employers expected their employees to be the primary care givers for children and if the neighborhood financed parental leave and daycare in such a way that encouraged both men and women to do the work of caring for children while giving men and women more equal opportunities to pursue work outside the home the community could actually make some progress on the issue of equal labor division in the family.

ps. I approve (I think) of the Dutch response. Thirteen is light years away from 16; the Dutch government is correct not to treat is as a grey zone.