Three times before last week's decision to launch airstrikes against the self-styled caliphate, the Islamic State, President Obama was urged to intervene in Middle Eastern conflicts: in Libya in the spring of 2011, in Syria from 2011 onward and in Iraq two short months ago, when Baghdad was threatened by the swift advance of the Islamic State.

In each case, there were good reasons to hesitate. In Libya, we had little to gain strategically from Moammar Gadhafi's fall, and more to fear from the vacuum that might follow. Syria was a more significant theater, but there as in Libya, there was little clarity about what forces (liberals? warlords? jihadis?) we would be empowering and what would follow Bashar Assad's rule.

A similar problem existed for the recent battles outside Baghdad. There was no question that America had an interest in seeing the southward advance of the Islamic State rolled back. But dropping bombs on behalf of Nouri al-Maliki's thuggish, failing government was a possible fool's errand: We would have been essentially serving as "the air force for Shia militias" (to quote David Petraeus, no dove) and by extension for the Islamic Republic of Iran.

All three situations were hard calls, and the fact that intervention in Libya and inaction in Syria produced similar outcomes — rippling chaos and jihadi gains — has allowed both hawks and doves to claim vindication. But in all three debates, the noninterventionist position ultimately had the better of the argument. We were better off sending advisers but not warplanes when the Islamic State threatened Baghdad; we were wise not to funnel arms (or at least not that many, depending on what the CIA's been doing) into Syria's chaos; and Obama would have been wise to heed the cautious Robert Gates on Libya.

This time, however, the case for war is much stronger, and the decision to intervene is almost certainly the right call.

In the earlier debates, the humanitarian case for action was in clear tension with strategic issues on the ground. In northern Iraq right now, the two are much more closely aligned. Alongside a stronger moral obligation to act than we had in Syria or Libya, we have a clear enough military objective, a more tested ally in the Kurds and a plausible long-term strategy that could follow from intervening now.

The stronger moral obligation flows from two realities. First, this humanitarian crisis is one our actions directly helped create: The cleansing of Christians, Yazidis and other religious minorities began in the chaos following our invasion of Iraq, and it has taken a more ruthless turn because the Islamic State profited from the fallout from our too-swift 2011 withdrawal. Second, the Islamic State represents a more distinctive form of evil even than a butcher like Assad.

These arguments still do not compel action absent a clear strategic plan, which is why the president was right to hesitate to take the fight to the Islamic State around Baghdad.

But in this case, such a plan is visible. We do not need to re-invade or restabilize Iraq to deal the Islamic State a blow and help its victims, because Kurdistan is already relatively stable, and the line of conflict is relatively clear. And the Kurds themselves, crucially, are a known quantity with a long-standing relationship to the United States — something that wasn't on offer in Libya or Syria.

So our intervention in northern Iraq has a limited, attainable objective: Push the Islamic State back toward the Sunni heartland, allow its victims to seek refuge in Kurdish territory and increase the Kurds' capacity to go on offense against the caliphate.

But if this president is thinking strategically, instead of just conducting a humanitarian drive-by, this intervention could also set the stage for a broader policy shift. Swiftly or gradually, depending on political developments in Baghdad, an independent, secure, well-armed Kurdistan could replace an unstable, perpetually fragmenting Iraq as the intended locus of U.S. influence in the region.

That influence will be necessarily limited: We are not going to stamp out the Islamic State on our own, or prevent the Middle East's rival coalitions from continuing their brutal proxy wars. There is not going to be a major U.S.-aligned model nation in the Arab world anytime soon, of the sort the Iraq invasion's architects naively hoped to build.

But by protecting a Kurdistan that can extend protection to groups made homeless by the fighting, we can still help save something from the wreckage.

Not a model, but a refuge.

Ross Douthat's column is distributed by the New York Times News Service.