If the center of this whole thing is Pakistan and its nukes, because we (correctly) believe them to be under threat in its current civil war, then Pakistan is now the main battleground. Afghanistan is now the secondary battleground, and important only to the extent that its collapse makes a Pakistani solution so difficult. As long as we can kill lots of enemies in place, or fleeing pressure from Pakistan, and maintain our ability to be in Afghanistan, that's all we want.

If that's the thesis, there are at least 2 approaches.

One is McChrystal's classic full-scale counter-insurgency approach (see his report). In terms of insurgency orientation, it's about enhanced legitimacy, and denied control. Protect the population, raise both a Afghan army and local armed networks that can make the Taliban's life hell, shore up governance so that there's a choice that's clearly better than the Taliban from an Afghan point of view (note low bar there). If you succeed, you have fashioned an anvil, against which the Pakistani state can strike a hammer into its own problem. Which, if successful, further eases the pressure on Afghanistan. Ideal result? 2 wins.

Another says: "You know what? We don't have an orientation. We don't want legitimacy OR control. We just want to make sure the Taliban are denied the ability to gain either in Afghanistan. And we'll work with anyone, anyone at all, if we think it will help us achieve that goal over the medium term - defined as 5-10 years." Ideal result: a win (in Pakistan) and a draw (in Afghanistan). Preferably a win in Pakistan that does not include the Islamist ISI going right back to supporting and assisting Islamonazis in Afghanistan, as a way of raising a backdoor army for a Pakistani coup.

What approach #2 needs on the Afghan end is a link between ends and means, that makes sense per Afghan situation and culture. It's ironic that the Obama administration may actually be leaning toward the Ralph Peters approach (shared by Brits like Rory Stewart, who has some experience in the area). But to work, there are some logical bugs between "sure, our forces could happily operate in that mode for a long time, even if the government comes apart" and "...this is how."

I'm actually happy to see that the Obama administration, and SecDef Gates, appear to be thinking this one through. They can't dither about it indefinitely, but now's the time.

« ok, I'm done now

I still remember a typically dishonest and projection-ridden piece in The Economist a while back, which covered Chechnya shortly after the Russians won. While it did backflips to avoid open admissions that Putin had in fact won the war, it went on employ the a classic Economist approach of acknowledging the fact of it in backhanded ways. While saying that in the long run, Putin would pay a price, the situation was not as great as it seemed for Russia, yadda, yadda, yadda.

Did Putin pay a price? Not really, no. It was a classic example of a publication that has only way to look at a situation, and can't even openly acknowledge that others exist. Which is definitely The Economist's core stock-in-trade across the board, so perhaps they're an unfair example to pick. But they're not the only ones afflicted with this kind of blindness, and if the organization is a military, the consequences can be dire.

This is not to say that the American/Western legitimacy approach is necessarily ineffective. And the classical control approach can also fail, vid. the Russians' initial experience in Afghanistan. What one can say with greater certainty is that the legitimacy approach is usually far more expensive than the classical alternative, and inherently more fragile. It has worked, most notably in Iraq. In some circumstances, it can even work better than the classical control approach. Then you have the 50/50 options, like Colombia's war against the FARC narco-terrorists and their backers in Venezuela, which may form a category all their own.

So, it all starts with the acknowledgment that there are several approaches which could be effective.

The question worth discussing is, what circumstances favour one approach vs. the other.

That's a question with long-term resonance beyond Afghanistan, but it also has special relevance there. The Taliban's approach relies heavily on control, not so much on legitimacy. Widespread experiences with their own previous regime ensure that. In contrast, the allied forces rely more heavily on legitimacy, and less on control.

Each side has definite weak points in its stance, but that's the broad outline of it. The 3 key questions as I see them are:

• What local and regional conditions (geographic, cultural, political, ideological) make control more likely to be a losing approach for state forces? For guerillas?

As one example, In Iraq an armed populace and a culture of direct justice (tribal revenge in this case) both became a definite problem for al-Qaeda and their allies. Because a large potential backer (US Marines) existed in the immediate area, the tribes always had the option of going postal on the jihadists. Eventually, they had enough of the abuses and murders, and the jihadists' control strategy collapsed catastrophically. Boyd called this approach "swordlessness", using the enemy's own sword against him. Ironically, it's a classic component of most guerilla strategies.

On the state side, meanwhile, the Russian experience in Afghanistan is just one of many instances which showed that safe havens, plus significant foreign support, plus a perception of low gains to be had from the controller's victory (even if that's wrong, as it was here), will make a control strategy untenable - even one that went to extremes like using chemical weapons.

Fostering backlashes to the evil and ruthless won't work everywhere, however. Whether the issue is being viewed from the state or non-state side, arms availability, culture, and receptivity must all be aligned, or the backlash to a control strategy will either never happen, or will lack sufficient force. This is most emphatically not a one-size-fits-all approach.

What other factors influence this question? Which of those requirements exist in Afghanistan? Some do (armed populace, tribal revenge culture), and experiences in Pakistan suggest that fostering friendly tribal lashkars is possible. Given al-Qaeda's current control of western Pakistan, however, armed backlash in Afghanistan could be seen by the locals as taking the losing side of a long-term tribal war. Is it seen that way? What other inhibitors exist? What would it take to change that? Is that change possible or realistic?

• What legitimacy-based approaches are possible if local governments and/or forces are weak?

The most frequent counter-proposal to the current approach of boots on the ground is to leave Afghanistan as a tribal, semi-lawless region with no effective control. It becomes more of a "special forces war," using small forces backed by mobile strike (aircraft, airmobile groups, or what have you) that focus on arming the local populace, striking the Taliban/al-Qaeda when found, and denying the enemy a successful control strategy. Legitimacy has no connection to any national government, just local alliances. Drugs are a non-issue, though trading them with the Taliban can get you shot. Ralph Peters is probably the most visible advocate for this approach: see his Sept 22/09 and Sept 17/09 columns.

It's a classic smart war strategy, and it's different than the approach taken at the moment. The question is whether it will work if the Afghan government collapses, and the Taliban can shift from their remaining bases in Pakistan to Afghanistan. With a nuclear-armed Pakistan as the long term prize, the enemy force just sloshes from one bowl to the other within "the lawless frontier," enabled at every step buy high-ranking Pakistani collaborators and enablers like Hamid Gul et. al.

The flip side of a special forces reliant strategy should also be acknowledged. You're going to take very valuable casualties, including some real disasters and prisoner-takings, because small teams are vulnerable (read Lone Survivor). This can also hamper operations behind enemy lines in places like Afghanistan, if control is strong enough. Not to mention the fact that the air support they need will mean civilian casualties - something that must be accepted, because a special forces strategy doesn't work well with rigid rules of engagement (vid. Porphyrogenitus' recent post). Which brings us back to our earlier question of whether a "swordlessness" strategy really works if allied forces aren't on the ground in southern Afghanistan in some numbers. Or does it just create a safe haven that can be used to destabilize a nuclear-armed state (not a worry in Iraq)?

What are the options for a "legitimacy, but minimized" approach here? What are their drawbacks? Some serious discussion of those points is necessary, in Washington and beyond.

• What control-based approaches are possible if local governments and/or forces are weak?

Latin America suggests some models, but in a different area and a different culture, and from the opposite vantage point, this is the Taliban's problem. If they try to impose full control, they become vulnerable, as they learned to their sorrow. Their question is how much control needs to be imposed, in order to bring the situation to a correlation of forces that favors them.

Busting that requires either disruption of that control, or weakening of the underlying correlation of forces they're relying on. Which is mostly in Pakistan. But doing either of those things requires another step first: the ability to see and understand the situation from within a different insurgency framework.

The fact that there was a presentation on the Russian approach at the Naval War College is a small piece of encouragement there.

I'd feel better if I saw more evidence of that ability to grok the existence and efficacy of another framework in the US and allied militaries. Need I mention that the Amero-European political class is largely a write-off when it comes to any of these issues, something that surely doesn't help. We'll see.

« ok, I'm done now

Is the war in Afghanistan in the interests of the United States and its allies? If so, at what point do the resources we are expending become too high a cost to bear? What are the strategic limitations of U.S. counterinsurgency doctrine and operations? And if the war is not in the interests of the United States and its allies, what are U.S. and allied interests in Central Asia - and how do you propose to secure them?

That unknown colors my own thoughts. I believe, contrary to many of this blog's readership, that Steve Biddle really does get our interests right in Afghanistan:

The United States has two primary national interests in this conflict: that Afghanistan never again become a haven for terrorism against the United States, and that chaos in Afghanistan not destabilize its neighbors, especially Pakistan. Neither interest can be dismissed, but both have limits as casus belli.

I also believe that both interests can be secured, and I believe that neither interest is worth unlimited U.S. and allied blood and treasure.

A resurgence of terrorist violence across Southeast Asia has exposed links between various Islamist terror organizations that have proved resilient despite a yearslong U.S.-funded crackdown by authorities in the region.

The rise of terrorism has come into focus in the wake of a series of attacks in the Philippines, southern Thailand and, most recently, the suicide bombing here July 17 on the JW Marriott and Ritz-Carlton hotels that killed nine people including the two bombers. The terror resurgence comes after years when authorities appeared to be gaining the upper hand.

As the terror groups expand their activities, investigators are uncovering connections that show how the main organizations across Southeast Asia, many of them inspired by al Qaeda, are providing militants, training and shelter to each other. That has increased their effectiveness and made them especially difficult for authorities to crack.

« ok, I'm done now