I've recently been reading this collection of Michael Walzer's essays. I'd not really read anything by him other than Just and Unjust Wars before, and it's quite fun: it's not as pithy or as astute when interested in something - there's nothing as good as Williams' reading of the torture scene in in 1984 in Truth and Truthfulness, for example - but it tends towards a Bernard Williams-esque valorization of the political, except with a kind of quite intriguing and probably distinctly American civic republican spin; if Williams had wanted to defend a principle of non-intervention, then one feels it would have been on the negative grounds of the tendency towards of disaster of interventions, rather than on the positive grounds of self-determination Walzer appeals to (in this paper).
Walzer appeals to an example where we could, by putting some chemical in the water supply, costlessly and permanently turn the post-independence Algerian government into Swedish-style social democrats. Presumably we feel there's something undesirable about this; at least I share Walzer's thought that something's gone wrong here, and plausibly something about self-determination. Williams, I think though - and this is probably a pure rhetorical device here: who really knows what Williams would have thought - would just ask what the moral evaluation of the acts of gods are supposed to tell us about what we, who are not gods, should do here and now. We can't costlessly lift a whole nation from the place history has brought it to so as to carefully put it down somewhere else instead; the point of politics is that we can't go round doing things like that. There are costs, and we have to work out whether they're ones we can bear, and ones we can impose. What should be done when there are no costs at all is a question for someone else entirely. It may be worth noting here that Walzer talked about a ticking bomb case in a paper in the early seventies.
Anyway, so Walzer wants non-intervention, including neutrality in civil wars, on grounds of self-determination. Even if we can partly set aside cases involving the total change of political, economic and cultural institutions and attitudes simply through an act of will, we cannot totally set them aside: even if we think that part of what has gone wrong in the case he uses is that gods shouldn't treat mere mortals like that, or that our moral reactions aren't supposed to be callibrated for those kinds of cases, it's still seems like a wrong against self-determination. So it looks like the pro-interventionist still has a case to answer: alright, this society is by any plausible standard less than fully just, but you can't make it fully just; properly, it is for its own people to do that. Walzer doesn't want to push that as far as it might go: he allows that in cases of genocide, ethnic cleansing, or mass enslavement, intervention is allowed. Other than those, though, he is prepared to require the citizens - better the inhabitants perhaps - of an unjust state to suck it up: it would be wrong for outsiders to help them make their state more just.
I think this is probably wrong, but I don't have to think it's wrong because self-determination doesn't matter: ecumenically, I might think it's wrong - and this is probably something that's been said before, but because of my unfamiliarity with the literature, I just don't know it's been said or by whom - because self-determination is not a zero-sum game. For Walzer to be right that self-determination rules out active interference, it has to be the case that active interference would always in some way infringe on self-determination: maybe not self-determination across all time, but at least self-determination in the roughly here and now. That, though, doesn't look plausible about other cases of self-determination, and is likely to be even less plausible about self-determination for groups.
Consider an individual who is in a domineering relationship. Although there might be a dispute about what counts as active interference, it seems to me we could fairly actively interfere with them and yet increase their self-determination, under certain conditions at least: encourage them and provide them with resources to leave or otherwise restructure that relationship, for example; if not that, at least give them the option of doing so. That looks like active interference, and certainly like the sorts of things Walzer wants to analogously rule out for societies: helping one side in a civil war by providing it with certain resources unavailable to the other; here, not encouraging the individual to stay in the domineering relationship. Yet that could certainly increase their self-determination in the roughly here and now: an abusive partner doesn't have to be beating you to within an inch of your life - presumably the analogue of Walzer's 'no intervention without genocide, ethnic cleansing, or mass enslavement' condition - for you to become more autonomous by leaving them.
More, note that individuals are singular, whereas societies are plural, and so the dispute about what counts as self-determination for a particular individual within that individual is unlikely to be as radical as the dispute about what counts as self-determination for a particular society within that society. Nor are parts of individuals able to silence other parts of individuals in ways that parts of societies are able to silence other parts of societies. A society where everyone is involved in the self-determination is more self-determining than a society in which only some are so involved. Then, though, there are trade-offs to be made: we might be able to raise the voices of some without lowering the voices of others quite as much, or even really at all. If self-determination's not a zero-sum game, then interference doesn't have to limit it, and Walzer's argument against interference fails, but it doesn't seem that self-determination is a zero-sum game, even for individuals and surely not for societies.