Some familiarity with strategic theory might not save the hopelessly confused debate on targeted killings in the "AfPak" region, but it might help. The basics: policy is a condition or behavior. Strategy is an instrumental device that accomplishes the policy. Policy, in turn, is created and continuously shaped by political processes. Even a good strategy can be rendered meaningless by bad policy. Targeted killing is not a "policy," as many have argued, as an action cannot be policy. Policy is a precondition to action. Strategy, in turn, is not entirely about fighting but is mainly based around the threat or use of military force. Hence targeted killing is a strategy of the state, not a tactic. All strategy takes empirical form in tactics but the art of strategy itself is arranging those tactics in a manner that achieves the political object.
Sometimes a strategy will be ineffective and cannot achieve the desired policy ends. The sudden collapse of the Japanese Empire in 1945 rendered irrelevant the Chinese Communist Party's strategy of building a conventional army around northern base areas while simultaneously building guerrilla bases farther to the south. Sometimes the policy itself is delusional and no strategy can save it, as was the case with the Islamic Republic of Iran's dream of conquering 1980s Iraq after it repelled Saddam Hussein's troops. Strategy and policy, however, have to be meaningfully separated and discussed in order to find clarity.
Most discussion about targeted killings in Pakistan revolve around questions of strategy--is a targeted killing strategy preferable, given the alternatives? By far, Joshua Foust is distingushing himself as one of the most thorough commentators on this question. A strategy of targeted killing may do less harm than the traditional Pakistani (and first British) method of exercising control over the regions criscrossed by American combat platforms: "butcher and bolt" punitive combined-arms raiding. It would certainly be preferbable if Pakistan would exercise its sovereignty with a method other than indirect rule and collective punishment. That's unfortunately as unlikely as Pakistan ceasing its support for militant groups engaged in armed aggression against its neighbors.
It also makes little sense to raise the cost of targeted killing if doing so will set back the policy the United States is seeking to achieve. If the policy is presumed to be correct, then the dominant concern should be ensuring that the right targets are selected. If greater transparency incentivizes this process, then we should all be for transparency. Strategy accomplishes the policy by either disarming the opponent or breaking his will to fight on. Killing the wrong people should be avoided, and if less and more precise strikes accomplish this goal then targeting authorizations should be tightened. But it makes little sense--if we feel the policy is correct--to limit the levels of force employed if the alternatives to achieving the policy goal are distinctly suboptimal.
Now, all of this presumes that the policy is correct. That proposition is in fact extremely debatable, and this uncertainty has implications for the implicit strategic and ethic calculations we are debating. First, US policy towards Pakistan has been, as Christine Fair notes, a case study in catastrophe. The United states has attempted to transform Pakistan's domestic politics through various forms of statecraft, all of which have ended in failure. If Bob Woodward is to be believed, the 2009 Afghan surge was built around the implicit idea that stabilizing Afghanistan would help the US better manage Pakistan. And as Greg Scoblete has noted, the intensity of the targeted killing campaign--and its attendant collateral damage--in part issues from the necessity of fighting the Pakistani border side of the war in Afghanistan.
Also embedded in the conversation about targeted killing is the problematic policy assumption that the greater problem is al-Qaeda rather than Pakistan itself. As Scoblete noted: "is the U.S. targetting militants that threaten the Pakistani state, or those sponsored by the Pakistani state. It's rather perverse to argue that drones are critical to protect Pakistan from militant violence when that country's intelligence service believes it is at war with you and uses militants to advance its own interests." This has immense implications for the American theory of victory behind the targeted killing campaign. There is no theory of victory if the policy presumes that we are unable to disarm or coerce the enemy because Pakistan's geostrategic postures make doing so impossible.
If the policy goal itself is flawed, then the targeted killing campaign may not, in fact, be the best way for the United States to defend itself from its declared enemies. Thus the ethical calculations that we presume when weighing air strikes vs. Pakistani air-ground offensives may be more complex than we think. Fair has outlined an alternative set of policies--containment and neglect--and a set of politico-military instruments that may obviate the need for the US to rely on targeted killings altogether. Or at least change the shape of how the United States employs force within Pakistan. A shift in regional posture, such as the United States adopting a different policy towards Afghanistan itself, may also open up different coercive options.
The question of policy is unfortunately rarely raised. Rather, we seem to prefer an circular conversation about the morality of machine warfare that was empirically decided a hundred years ago with the invention of Maxim guns, indirect fire systems, and ground-attack aircraft. Even talking about strategy may not necessarily yield insights. Let's focus on the policy.
UPDATE: James Joyner pointed out in Para 3 I use the phrase "policy" to refer to targeted killings. Mea culpa: the pervasiveness of the use of the public language has infected even my head. I have changed the paragraph.