Speaking on Monday to the Veterans of Foreign Wars in Phoenix, President Obama could not have been more definitive. "We must never forget," he said of the conflict in Afghanistan. "This is not a war of choice. This is a war of necessity."

The president did not break new ground so much as reinforce existing policy. Earlier this year, he decided to send an additional 17,000 combat soldiers and 4,000 trainers to Afghanistan, raising American force levels there to more than 60,000. And in March he articulated a broader mission: The United States would now "take the fight to the Taliban in the south and the east," in effect making the United States a full party to Afghanistan's civil war.

As a result, American soldiers are fighting the Taliban, partly to provide time and space while Afghan forces are better trained and partly to persuade some Taliban that resistance does not pay. Call it armed state-building.

But is Afghanistan a war of necessity? And if not — if in fact it is a war of choice — so what?

Wars of necessity must meet two tests. They involve, first, vital national interests and, second, a lack of viable alternatives to the use of military force to protect those interests. World War II was a war of necessity, as were the Korean War and the Persian Gulf war.

In the wake of 9/11, invading Afghanistan was a war of necessity. The United States needed to act in self-defense to oust the Taliban. There was no viable alternative.

Now, however, with a friendly government in Kabul, is our military presence still a necessity?

Of course, our interests in Afghanistan include making it difficult for al-Qaida to mount operations from that country and limiting Taliban use of Afghan territory to destabilize neighboring Pakistan. Minimizing the chance of a terrorist attack on American citizens is vital, as is making sure that Pakistan's nuclear arsenal doesn't fall into the wrong hands.

But even if the United States were to succeed in Afghanistan — with "success" defined as bringing into existence an Afghan government strong enough to control most of its territory — terrorists could still operate from there and would put down roots elsewhere. And Pakistan's future would remain uncertain at best.

Moreover, there are alternatives to current American policy. One would reduce our troops' ground-combat operations and emphasize drone attacks on terrorists, the training of Afghan police officers and soldiers, development aid and diplomacy to fracture the Taliban.

A more radical alternative would withdraw all U.S. military forces from Afghanistan and center on regional and global counterterrorism efforts and homeland security initiatives to protect ourselves from threats that might emanate from Afghanistan. Under this option, our policy toward Afghanistan would resemble the approach toward Somalia and other countries where governments are unable or unwilling to take on terrorists and the United States eschews military intervention.

Afghanistan is thus a war of choice — Obama's war of choice. In this way, Afghanistan is analogous to Vietnam, Bosnia, Kosovo and today's Iraq. Wars of choice are not inherently good or bad. It depends on whether military involvement would probably accomplish more than it would cost and whether employing force is more promising than the alternatives.

Making this assessment in Afghanistan is difficult. The Taliban are resourceful and patient and can use Pakistan as a sanctuary. It is not obvious that Afghans can overcome ethnic and tribal loyalties, corruption and personal rivalries. No matter who is declared the winner, Thursday's election is almost certain to leave the country even more divided.

The risk of ending our military effort in Afghanistan is that Kabul could be overrun and the government might fall. The risk of the current approach (or even one that involves dispatching another 10,000 or 20,000 American soldiers, as the president appears likely to do) is that it might produce the same result in the end, but at a higher human, military and economic cost.

All of which makes Afghanistan not just a war of choice but a tough choice. My judgment is that American interests are sufficiently important, prospects for achieving limited success are sufficiently high and the risks of alternative policies are sufficiently great to proceed, for now, with Obama's measured strategy. But the administration, Congress and the American people — who, recent polls suggest, are turning against the war — must undertake regular, rigorous assessments of whether these efforts are bearing fruit or are likely to. If it appears they are not, the president should roll back the combat role or withdraw militarily.

If Afghanistan were a war of necessity, it would justify any level of effort. It is not and does not. It is not certain that doing more will achieve more.

And no one should forget that doing more in Afghanistan lessens our ability to act elsewhere, including North Korea, Iran and Iraq. There needs to be a limit to what the United States does in Afghanistan and how long it is prepared to do it, lest we find ourselves unable to contend with other wars, of choice or of necessity, if and when they arise.