Given declining poll numbers and rising casualty figures, it is no surprise that the chattering classes are starting to bail out on a war in Afghanistan that was launched with their enthusiastic support. From Sen. Russ Feingold on the left to columnist George Will on the right, these born-again doves seem to be chastened by the fact that the Taliban won’t simply stop fighting. Rather than rise to the challenge, they propose that we stick to what Mr. Will says "can be done from offshore, using intelligence, drones, cruise missiles, airstrikes and small, potent Special Forces units, concentrating on the porous 1,500-mile border with Pakistan, a nation that actually matters."

If only we could. No one wants to see troops risking injury and death in ground combat. It would be nice if it weren't necessary. But it is. We tried the offshore strategy in the 1990s when Afghanistan became a stronghold of al Qaeda. Even after 9/11 we still stuck to a minimalist approach. Recall the failure to capture Osama bin Laden at Tora Bora because we wouldn't commit enough American troops. As recently as 2008 there were only two U.S. Brigade Combat Teams in the entire country (a brigade has roughly 4,000 soldiers), compared to 20 in Iraq at the height of the surge.

There are now five brigades engaged in combat in Afghanistan. For most of the Bush administration, we relied on unmanned Predator drones and Special Forces to keep the enemy at bay. Afghan Security Forces were too small and ineffective to pick up the slack. Even today there are only 173,000 Afghan soldiers and police compared to 600,000 in Iraq. The result: The Taliban, which had been routed in 2001, staged a disheartening resurgence.

However much advocates of downsizing might want to disguise the fact, there is no alternative to doing the kind of intensive counterinsurgency work on the ground that has paid off in numerous conflicts from Malaya to Iraq. If we don't make a substantial commitment—one that will require raising our troop strength beyond the 68,000 to which the administration is already committed—we are likely to lose.

Losing wars is a bad thing. It is especially bad if you are a superpower that depends on an aura of invincibility to keep rogue elements at bay. That should go without saying, but those calling for a scuttle from Afghanistan seem to have forgotten this elementary lesson. They might cast their minds back to the 1970s when we were reeling from defeat in Vietnam and our enemies were on the march from Nicaragua to Iran. Or back to the 1990s when, following the U.S. pullout from Lebanon and Somalia, Osama bin Laden labeled us a weak horse that could be attacked with impunity.

A U.S. drawdown in Afghanistan would lead to defeat with consequences at least as serious. The Taliban would expand their control, probably seizing Kandahar, the principal city of the south. Then they would besiege Herat, Kabul and other urban centers. No doubt the central government could hold out for some time, and the Taliban would be unlikely to ever capture all of northern Afghanistan—territory they did not control even on Sept. 10, 2001. But they could certainly impose their diktat over substantial territories where narco-traffickers and terrorists would have free run.

The impact on Pakistan—"a nation that actually matters," in Mr. Will's words—is particularly sobering. To the extent that we have been able to stage successful attacks on al Qaeda strongholds in Pakistan, it is because we have secure bases in Afghanistan. To the extent that we have not been more successful in getting the government of Pakistan to eliminate the militants on its own, it is because we have not convinced all of the relevant decision-makers (particularly in the military and intelligence services) that we will be in the region for the long-term. Many Pakistanis still regard the U.S. as a fickle superpower—here today, gone tomorrow. That impression took hold after we left Afghanistan and Pakistan in the lurch in the 1990s after having made a substantial commitment to fight Soviet invaders in the 1980s.

If there is any wavering in our commitment to Afghanistan, officials in Pakistan will take that as confirmation that their old strategy of cutting deals with Islamic militants is more necessary than ever. That means that the Taliban and related groups, which have been on the defensive lately following a Pakistani army offensive, will be more secure than ever in their sanctuaries. They will then use these bases not only to try to topple the governments in Kabul and Islamabad but also to stage international acts of terrorism. It would be the biggest victory for the jihadists since the Red Army marched out of Afghanistan and the biggest defeat for the U.S. since Vietnam.

Such an outcome is by no means inevitable. It is true that winning in Afghanistan—meaning creating sufficient stability for the democratically elected government to secure its own territory without a substantial foreign troop presence—will not be quick or easy. But nor will it be as difficult as in Iraq where we faced not only an insurgency but an incipient civil war. The good news in Afghanistan is that notwithstanding tensions between Hazaras, Pashtuns, Tajiks, and other groups, they are not fighting one another the way that Shiite and Sunni militias were in Iraq.

The security problems are largely confined to the Pashtun belt comprising roughly half the country's population of 33 million. And, although violence continues to rise, it is far below the levels seen in Iraq in 2006-2007. Even recently, when fighting has abated in Iraq, more civilians have been dying most months in Iraq than in Afghanistan. (In June, 340 civilians were killed in Iraq as compared to 198 in Afghanistan.)

The Taliban and related groups are tough, tenacious foes but they are hardly invincible. Their Achilles heel is lack of popular support. An International Republican Institute poll of 2,400 Afghans in July found that only 19% have a favorable view of the Taliban compared to 62% who have a positive impression of the U.S. and 82% who view the Afghan National Army favorably. A poll taken earlier this year by the BBC and ABC found that only 4% of Afghans want the Taliban to return to power. U.S. forces are not going to replay the experience of the Red Army—popular legitimacy is on our side in a way that it never was for the Russians, despite doubts that are emerging about the integrity of the presidential election. The only reason the Taliban have made gains is because of a governance and security vacuum that they have filled with fear and intimidation.

Until now international forces and their Afghan partners have lacked the will and resources to implement a classic counterinsurgency plan designed to secure the populace. But that is precisely what Gen. Stanley McChrystal will undertake—assuming he gets the resources he needs from Washington. To pull the plug on our operations now, when our troops are only beginning to fight in earnest, would be even more foolish than it would have been to short-circuit the surge in Iraq in 2007—as so many who are freely offering advice on Afghanistan today once advocated.

Mr. Boot is a senior fellow at the Council on Foreign Relations and author, most recently, of "War Made New: Technology, Warfare, and the Course of History, 1500 to Today" (Gotham, 2006).