As he announced on Tuesday, President Obama is betting that sending 30,000 more troops to Afghanistan will rapidly change the balance of power in the field, erode local support for the Taliban and bolster their local opponents, give breathing space to the Kabul government to clean up its act, allow humanitarian aid and development to reach the countryside, and possibly bring some war-weary Taliban to the negotiating table.
Al Qaeda would thus be deprived of any sanctuary, and the US mission there would be accomplished.
In essence, the American president announced a short-term military surge in Afghanistan to lay the ground for implementing a long-term political agenda, first put in place by the Bush administration in 2002, that focuses on good governance, fighting corruption, training a professional police force, and promoting economic and social development.
Since the political project has failed over the past eight years, the logic goes, only military action can revive the conditions for it. So everything depends on military progress in counterinsurgency.
It is true that, at a time when the Taliban are on the move and the Kabul government embodies more than ever a failed state, nothing can be done without a military surge. The Taliban smell victory and have no interest in negotiating. The only alternative is to leave or to escalate the fighting.
But can the new counterinsurgency work?
The idea seems to be to use tactics that worked in northern Iraq: playing traditional tribal leaders against radical Taliban, offering them incentives, and hoping that the large strata of the population who don't share the Taliban's agenda will turn against them.
In this perspective, the corrupt and distrusted Kabul government is more a liability than an asset, which means that the American and NATO troops would have to be politically involved at the local levels instead of handing over the keys to Kabul once the field has been cleared.
For such a policy to work, the Taliban insurrection must be correctly understood and Pakistan's influence in Afghanistan must be neutralized at least, and turned in a good direction.
The Taliban insurrection is both an ethnic and a social movement. The Taliban embody both a Pashtun irredentism and a shift in the traditional tribal system. The insurgency is limited to Pashtun-populated areas or pockets: the south; and, in the north, Baghlan, Kunduz, Balkh and Badghis, often delivered by the Hizb-e-Islami of Gulbuddin Hekmatyar. In Pakistan, too, the "liberated Islamic areas" are all Pashtun. Non-Pashtun Islamic militants choose other ways to act.
The issue of Pashtun frustration at being shut out of power has not been ignored by the Western powers. They supported the dismantling of the ethnically non-Pashtun Northern Alliance forces that took Kabul in November 2001: a rather easy task after the assassination of their charismatic leader, Ahmed Shah Massoud.
But now the non-Pashtuns in Afghanistan have no more military means to protect themselves from a bloody Taliban comeback, and they cannot rely on an Afghan national army. Thus the quandary is how to placate the Pashtuns without weakening further the other ethnic groups whose fears of a Taliban comeback make them the best allies of the NATO troops.
President Hamid Karzai was appointed largely because he could embody a traditional Pashtun identity. When he came to power, he brought many Pashtuns from the diaspora with him. He appointed Pashtun governors and has played on Pashtun traditions. Yet, this has been to no avail because the tribal aristocracy he represents has lost its roots in the tribal areas.
In northern Iraq, traditional tribal leaders happily answered Gen. David Petraeus's opening to get rid of the threat of non-Iraqi Al Qaeda fighters who ignored or even tried to suppress them. But in Afghanistan as well as in Pakistan, traditional leaders of this kind have almost disappeared. They have been replaced by a new elite of young madrasa-educated Taliban, more connected to Pakistan and the Gulf than to the West.
What of the role of Pakistan?
If they find a shelter in Pakistan, the Taliban could easily escape the brunt of the two coming years of a military surge. They can reasonably expect that the United States will be unable to bolster a counterpower in the Afghan tribal belt or strengthen the Kabul government. So, they just have to wait.
Pressure from outside on Pakistan will yield very little: the arrest or the killing of some Taliban leaders or Al Qaeda cadres. Until now the Pakistani Army has used both Taliban and Islamist militants as a proxy tool of its regional policy of "strategic depth" vis-à-vis India. It still wants a Pashtun Islamist government in Kabul.
This complex and dangerous cooperation between the Army and the Taliban was based on a deal: The Taliban, Afghan or Pakistani, might push their agenda in Afghanistan or in the northwest territories in Pakistan but should not contest the leadership of the Pakistani Army. Islamabad is off limits.
The Taliban broke this deal when they made a foray from their Swat stronghold through Buner in the direction of Islamabad. The Army had no choice than to counterattack. But the objective of the Pakistani Army is not to destroy the Taliban. It is to bring them back into the fold after a red line has been crossed.
As long as the Pakistani Army does not consider its campaign against the Taliban as a matter of life and death for itself, it will not help in any serious way with the American and NATO agenda in Afghanistan. Pakistan has been fighting through proxies in Afghanistan for more than 30 years. It can wait for American and NATO troops to leave the region.
Only finding a way to alleviate Pashtun frustration in Afghanistan and getting Pakistan to give up its decades-old policy of supporting Islamists in power there will change anything fundamental.
Unless a broader and more coherent policy is defined that includes these elements, 30,000 additional US troops plus more from NATO are not going to make a difference.
Olivier Roy is author of "Globalized Islam: The Search for a New Ummah."