Rationales for maintaining the counterinsurgency in Afghanistan are varied and complex, but they all center on one key tenet: that Afghanistan must not be allowed to again become a haven for terrorist groups, especially al-Qaeda. The debate about Afghanistan has raised reasons to question that tenet, one of which is that the top al-Qaeda leadership is not even in Afghanistan, having decamped to Pakistan years ago. Another is that terrorists intent on establishing a haven can choose among several unstable countries, and U.S. forces cannot secure them all.

But the debate has largely overlooked a more basic question: How important to terrorist groups is any physical haven? More to the point: How much does a haven affect the danger of terrorist attacks against U.S. interests, especially the U.S. homeland?

The answer to the second question is: not nearly as much as unstated assumptions underlying the current debate seem to suppose. When a group has a haven, it will use it for such purposes as basic training of recruits. But the operations most important to future terrorist attacks do not need such a home, and few recruits are required for even very deadly terrorism.

Consider: The preparations most important to the Sept. 11, 2001, attacks took place not in training camps in Afghanistan, but, rather, in apartments in Germany, hotel rooms in Spain, and flight schools in the United States.

In the past couple of decades, international terrorist groups have thrived by exploiting globalization and information technology, which has lessened their dependence on physical havens. By using networks such as the Internet, terrorist organizations have become more network-like, not beholden to any one headquarters.

A significant jihadist terrorist threat to the United States persists, but that does not mean it will consist of attacks instigated and commanded from a South Asian haven, or that it will require a haven at all. Al-Qaeda's role in that threat is now less one of commander than of ideological lodestar, and for that role a haven is almost meaningless.

These trends have been familiar to counterterrorist cognoscenti for years but have gone mostly unmentioned in the discussion on Afghanistan. This is probably because the intervention there in late 2001 was unquestionably a response to Sept. 11 - the "good war," in contrast with the misguided expedition to Iraq, where the only connection to the 2001 attacks was in the Bush administration's contorted selling of that invasion. The U.S. entry into the Afghan civil war succeeded in ousting the Taliban from power and rousting its al-Qaeda allies, and the intervention would have occurred regardless of whether the occupant of the White House was named Bush or Gore.

The issue today does not concern what was worth disrupting eight years ago. And it is not whether a haven in Afghanistan would be of any use to a terrorist group, which it would.

Instead, the issue is whether preventing such a haven would reduce the terrorist threat to the United States enough to offset the required expenditure of blood and treasure and the barriers to success in Afghanistan, including an ineffective regime and sagging support from the population. Thwarting the creation of a physical haven also would have to offset any boost to anti-U.S. terrorism stemming from perceptions that the United States had become an occupier rather than a defender of Afghanistan.

Among the many parallels between Afghanistan and Vietnam, one of the most disturbing concerns inadequate examination of core assumptions. The Johnson administration was just as meticulous as the Obama administration in examining counterinsurgent strategies and the forces required to execute them. But most American discourse about Vietnam in the early and mid-1960s took for granted the key - and flawed - assumptions underlying the whole effort: that a loss of Vietnam would mean other Asian countries would fall like dominoes to communism, and that a retreat from the commitment to Vietnam would gravely harm U.S. credibility.

The Obama administration and other participants in the debate about expanding the counterinsurgency effort in Afghanistan can still avoid comparable error. But this would require not merely invoking Sept. 11 and taking for granted that a haven in Afghanistan would mean the difference between repeating and not repeating that horror. It would require presenting a convincing case about how such a haven would significantly increase the terrorist danger to the United States. That case has not yet been made.

Paul R. Pillar was deputy chief of the counterterrorist center at the CIA from 1997 to 1999 and is director of graduate studies at Georgetown University's Security Studies Program. This appeared in the Washington Post.