In his interestingly wrong-headed essay, Paul Miller argues that there are extraterritorial evils so great that they oblige the United States to intervene militarily to preempt them when they threaten, to put a stop to them when they’ve already begun, and to redress their consequences when they’re over.

Afghanistan is the case that motivates Miller’s argument, and the principal purpose of his essay is to advocate a long-haul American military presence there. But what makes the Afghan atrocities different from others? What, for example, about the current large-scale violence in Congo, Mali, Sudan, and Syria? What about the regime-produced body count from starvation and preventable disease in North Korea during the last two generations and continuing today? What about the ongoing mass killing of the unborn in the United States and many other countries? What about the large-scale civilian death produced by the American invasion of Iraq from 2003 through 2011? And so on.

Most of these cases”perhaps all of them”have brought about more non-combatant deaths by violence than anything that happened in Afghanistan during the 1990s, and more than are likely should we withdraw. If body count is among the criteria for distinguishing exceptional from quotidian evils, as surely it should be, it is hard to see what makes Afghanistan distinctive. Miller predicts that civil war like that of the 1990s will follow a precipitate American withdrawal. Perhaps, but compared to the decade of warfare since the American invasion, that might seem to Afghans no bad exchange.

Miller never shows why Afghanistan requires the soothing Samaritan presence of the United States’ armed forces while Mali and Syria and North Korea do not. This troubles him. He sees that his position can easily be understood to require military intervention in and occupation of many places, so as to provide a worldwide Pax Americana as the Romans once imposed their Pax Romana. And because he doesn’t want to arrive at that conclusion, he must, somehow, block it.

Of course, he says, his argument doesn’t mean that “the United States should have a direct combat role against every Islamist insurgency in the world.” No, we should intervene militarily only when an “insurgency threatens to overwhelm local governments or threatens broader regional instability.” But, then, why haven’t we invaded Yemen? Why no occupation and stabilization of Somalia?

He also says that what he argues is “not a call for an American empire.” But he offers, so far as I can see, no reason for this denial, other than the pragmatic claim that we can’t manage it. But that isn’t an argument against empire; it’s an argument that we should be as much of an empire as we can manage.

He then insists that “the United States does not have a moral duty to intervene everywhere there is injustice” and that we have such a duty only when we are faced with exceptional evils. Because he provides no way to distinguish exceptional evils from mere injustices, this is no more than hand waving.

Perhaps there is a modest pragmatic case for extending the occupation of Afghanistan beyond 2014”I don’t myself think so, but that’s a matter for reasonable disagreement. What there certainly isn’t is a high moral case of the sort Miller makes for continuing occupation.

There are deeper difficulties with Miller’s argument, peculiarly American and deeply dangerous ones.

The first is that Miller’s use of the Samaritan analogy is objectionable. The Samaritan’s instruments are those of mercy: water, oil, bandages for the wounded body, provision for continued care. The military’s instruments are, mostly, those of violence: guns, Humvees, drones, missiles, helicopter gunships. Tanks aren’t ambulances; bullets aren’t bandages; military encampments aren’t hospitals. To pretend otherwise is both confused and repellent. The Samaritan analogy works beautifully for Doctors Without Borders; it works not at all for the U.S. Army.

The second has to do with Miller’s axiomatic assumption that the political arrangements of the liberal world order are more just than the alternatives, particularly the Islamic alternatives. He doesn’t argue for this, but it’s not obvious, and probably not true. Capitalist democratic states have their own characteristic forms of injustice, their own peculiar ways of shedding innocent blood, their own idiosyncratic contributions to the malformation of the human. We Americans, for instance, kill by violence within our own borders and extraterritorially at a rate exceeding that of most contemporary Islamic states, and our history is not less bloody than theirs.

Blindness to these facts is necessary for a view like Miller’s to gain traction. But it is an indefensible blindness, and, I think, a very American one. We tend to excuse”or not to see”the blood we shed, because we are dazzled by the ideals in the service of which we take ourselves to be shedding it.

The third difficulty has to do with Miller’s belief that we can easily (or at all) predict the outcome of what we do militarily. He assumes that things will be better for Afghans if the United States continues its occupation than they would be were it to be ended summarily. This is an irrational assumption. A glance at recent attempts to predict the results, short- or long-term, of extraterritorial military adventures shows that, even when some prediction turns out to have been right, so many made at the same time turn out to have been wrong that we can have no confidence in the rightness of our predictions. Miller seems blithely unaware of this. That too is culpable.

Miller’s argument is a moral one, and I respect him for that. But it is vitiated, again in a very American way, by overconfidence, which is the fundamental and constitutive American sin, evident already in the Declaration of Independence (“We hold these truths to be self-evident”). The fundamental question, the one I worry about in the dark of the night, is whether we can resist the temptation to think that the only proper limit on our use of force to remake the world in our image is our capacity to deploy it. Miller has not resisted that temptation. His position is an imperial one, at least by aspiration, and that means blood and violence and darkness, as utopian imperial aspirations always do.

Does justice require that the United States continue, open-endedly, its occupation of Afghanistan? No. The argument Miller makes yields that conclusion only at the unacceptable cost of requiring morally a Pax Americana. The argument is distressingly intimate with a culpable blindness to the deformities of the liberal world order, and it requires an unjustified and unjustifiable optimism about our capacity to predict outcomes.

Can we do better? Perhaps, if we learn to lament our military necessities rather than celebrate them, and if we learn to separate the works of mercy from those of violence and to act as if we believed such a separation important. Then we might indeed become a beacon of hope for the world, rather than the reviled symbol of empire that we too often are.

And as for Afghanistan? We should do what we are doing, which is to withdraw militarily with all due speed. And we should offer the angels of mercy, the true Samaritans”doctors and teachers and engineers and donors, in quantity and without stint”to whichever government eventually emerges. Those efforts too, should they occur, would inevitably fail to some considerable extent. But they would be the right thing to do. And they would be beautiful, acts worthy of America’s own beauty.

Paul J. Griffiths is Warren Chair of Catholic Theology at Duke Divinity School.