February 25, 2013

Is the "War on Terror" Lawful?

The "Authorization to Use Military Force" serves as the primary legal foundation for the ongoing conflict, but it is now obsolete. What should replace it?

Since September 18, 2001, a joint resolution of Congress known as the Authorization to Use Military Force (AUMF) has served as the primary legal foundation for the “war on terror.” In this essay we explain why the AUMF is increasingly obsolete, why the nation will probably need a new legal foundation for next-generation terrorist threats, what the options are for this new legal foundation, and which option we think is best.   

The AUMF authorizes the president to “use all necessary and appropriate force against those nations, organizations, or persons he determines planned, authorized, committed, or aided the terrorist attacks that occurred on September 11, 2001, or harbored such organizations or persons, . . . .” The authorization of “force” in the AUMF is the main legal basis for the president’s power to detain and target members of al Qaeda and The Taliban. In addition, since September 11, Congress, two presidential administrations, and the lower federal courts have interpreted the “force” authorized by the AUMF to extend to members or substantial supporters of the Taliban and al Qaeda, and associated forces.

  war on terror by jack goldsmith, benjamin wittes, robert chesney, matthew waxman  
  Photo credit: Wikipedia Commons

The main reason the AUMF is becoming obsolete is that the conflict it describes – which on its face is one against the perpetrators of the September 11 attacks and those who harbor them – is growing less salient as U.S. and allied actions degrade the core of Al Qaeda and the U.S. military draws down its forces fighting the Taliban in Afghanistan. At the same time that the original objects of the AUMF are dying off, newer terrorist groups that threaten the United States and its interests are emerging around the globe. Some of the terrorist groups have substantial ties to al Qaeda and thus can be brought within the AUMF by interpretation.

For example, the President has been able to use force against al Qaeda in the Arabian Peninsula (“AQAP”), a terrorist organization in Yemen, because it is a supporter or associated force of al Qaeda. But this interpretive move is increasingly difficult as newer threatening groups emerge with dimmer ties, if any, to al Qaeda. As a result, we are reaching the end point of statutory authority for the President to meet terrorist threats.

We should emphasize at the outset that we do not claim that the increasingly obsolete AUMF demands immediate amendment or alteration. We do not make this claim because we lack access to classified information that would indicate the full nature of the terrorist threats the nation faces, or their connection to al Qaeda, or the nation’s ability to meet the threat given current legal authorities.

We also recognize that any new force authorizations carry significant strategic and political consequences beyond their immediate operational consequences. We nonetheless believe strongly – based on public materials and conversations with government officials – that the AUMF’s usefulness is running out, and that this trend will continue and will demand attention, in the medium term if not in the short term. Our aim is to contribute to the conversation the nation will one day have about a renewed AUMF by explaining why we think one will be necessary and the possible shape it might take.

Part I of this paper explains in more detail why the AUMF is becoming obsolete and argues that the nation needs a new legal foundation for next-generation terrorist threats. Part II then describes the basic options for this new legal foundation, ranging from the President’s Article II powers alone to a variety of statutory approaches, and discusses the pros and cons of each option, and the one we prefer. Part III analyzes additional factors Congress should consider in any such framework.

I. The Growing Problem of Extra-AUMF Threats and the Need for a New Statutory Framework

In this Part we explain why the AUMF is growing obsolete and why a combination of law enforcement and Article II authorities, standing alone, is not an adequate substitute.

1.  The Growing Obsolescence of the AUMF

The September 2001 AUMF provides for the use of force against the entity responsible for the 9/11 attacks, as well as those harboring that entity. It has been clear from the beginning that the AUMF encompasses al Qaeda and the Afghan Taliban, respectively. This was the right focus in late 2001, and for a considerable period thereafter. But for three reasons, this focus is increasingly mismatched to the threat environment facing the United States. 

First, the original al Qaeda network has been substantially degraded by the success of the United States and its allies in killing or capturing the network’s leaders and key personnel. That is not to say that al Qaeda no longer poses a significant threat to the United States, of course. The information available in the public record suggests that it does, and thus nothing we say below should be read to suggest that force is no longer needed to address the threat al Qaeda poses. Our point is simply that the original al Qaeda network is no longer the preeminent operational threat to the homeland that it once was.

Second, the Afghan Taliban are growing increasingly marginal to the AUMF. As noted above, the AUMF extended to the Taliban because of the safe harbor they provided to al Qaeda. That rationale makes far less sense a dozen years later, with the remnants of al Qaeda long-since relocated to Pakistan’s FATA region. This issue has gone largely unremarked in the interim because U.S. and coalition forces all along have been locked in hostilities with the Afghan Taliban, and thus no occasion to reassess the AUMF nexus has ever arisen.

Such an occasion may well loom on the horizon, however, as the United States draws down in Afghanistan with increasing rapidity. To be sure, the United State no doubt will continue to support the Afghan government in its efforts to tamp down insurgency, and it also will likely continue to mount counterterrorism operations within Afghanistan. It may even be the case that at some future point, the Taliban will again provide safe harbor to what remains of al Qaeda, thereby at least arguably reviving their AUMF nexus. But for the time being, the days of direct combat engagement with the Afghan Taliban appear to be numbered.  

 If the decline of the original al Qaeda network and the decline of U.S. interest in the Afghan Taliban were the only considerations, one might applaud rather than fret over the declining relevance of the AUMF. There is, however, a third consideration: significant new threats are emerging, ones that are not easily shoehorned into the current AUMF framework.

To a considerable extent, the new threats stem from the fragmentation of al Qaeda itself. In this sense, the problem with the original AUMF is not so much that its primary focus is on al Qaeda, but rather that it is increasingly difficult to determine with clarity which groups and individuals in al Qaeda’s orbit are sufficiently tied to the core so as to fall within the AUMF. And given the gravity of the threat that some of these groups and individuals may pose on an independent basis, it also is increasingly odd to premise the legal framework for using force against them on a chain of reasoning that requires a detour through the original, core al Qaeda organization.

The fragmentation process has several elements. First, entities that at least arguably originated as mere regional cells of the core network have established a substantial degree of organizational and operational independence, even while maintaining some degree of correspondence with al Qaeda’s leaders. Al Qaeda in the Arabian Peninsula (AQAP) is a good example. Al Qaeda in Iraq (AQI) arguably fits this description as well, though in that case one might point to a substantial degree of strategic independence as well.

Second, entities that originated as independent, indigenous organizations have to varying degrees established formal ties to al Qaeda, often rebranding themselves in the process. Al Qaeda in the Islamic Maghreb (AQIM), formerly known as the Salafist Group for Call and Combat, illustrates this path. Al Shabaab in Somalia arguably does as well. And then there are circumstances (such as the ones currently unfolding in Mali, Libya, and Syria) in which it is not entirely clear where the organizational lines lie among (i) armed groups that work in concert with or even at the direction of one of the aforementioned al Qaeda affiliates; (ii) armed groups that are sympathetic and in communication with al Qaeda; and (iii) armed groups that are wholly-independent of al Qaeda yet also stem from the same larger milieu of Salafist extremists.

This situation – which one of us has described as the emergence of “extra-AUMF” threats – poses a significant problem insofar as counterterrorism policy rests on the AUMF for its legal justification. In some circumstances it remains easy to make the case for a nexus to the original al Qaeda network and hence to the AUMF. But in a growing number of circumstances, drawing the requisite connection to the AUMF requires an increasinglycomplex daisychain of associations – a task that is likely to be very difficult (and hence subject to debate) in some cases, and downright impossible in others.

The emergence of this problem should come as no surprise. It has been nearly a dozen years since the AUMF’s passage, and circumstances have evolved considerably since then. It was inevitable that threats would emerge that might not fit easily or at all within its scope. The question is whether Congress should do anything about this situation, and if so precisely what.

2. The Inadequacy of Law Enforcement and Article II

 Consider first the option of Congress doing nothing. This is, at bottom, a choice to address extra-AUMF threats through a combination of increasingly strained executive branch interpretations of the AUMF, law enforcement and intelligence measures, and whatever supplemental military force the President can and will assert based on his Article II authorities. It is our contention that at some point even strained interpretations of the AUMF will not be possible, and that even before we reach that point, the strained interpretations will call into question the legitimacy of congressional and democratic backing for the president’s uses of force. That leaves law enforcement measures and Article II powers, which in combination are far from ideal.

To be very clear, we do not claim that all terrorism-related threats can or should be dealt with militarily. Law enforcement and intelligence tools can have tremendous effect, and we strongly endorse the view that the President’s authority to use them should not be unduly constrained out of a misguided sense that most or all terrorism scenarios require a military solution. But law enforcement and intelligence tools are not a panacea. In some circumstances – such as the late 1990s in Afghanistan and today in certain areas of Pakistan, Yemen, Somalia, and the Sahel region – these options simply do not provide sufficient capacity to capture individuals or to otherwise disrupt their activities. And in some circumstances, these tools are equally inadequate to the task of long-term incapacitation.

Meanwhile, local governments are sometimes either incapable of or unwilling to address terrorism threats; in some cases, for various reasons, we would not want to entrust them with these responsibilities. Whether this is the case with respect to any given extra-AUMF threat at any given point in time is exceedingly difficult to say, particularly for those (including us) who are outside government and lack access to the relevant intelligence. We proceed on the assumption, however, that some such circumstances do exist or will arise.

Bearing this in mind, the next issue is whether the President’s inherent powers under Article II are adequate to address any gap that may emerge between what defense of the nation demands and what law enforcement and intelligence options can provide in extra-AUMF scenarios. We are skeptical, for three reasons.

First, it is worth bearing in mind that some administrations are more comfortable resorting to claims of Article II authority than others. The Obama administration, for example, has consciously distanced itself from the Bush administration on this dimension, at least in the counterterrorism setting (as opposed to the operation in support of the revolution in Libya, which relied on a surprisingly bold stand-alone Article II argument). In a situation where a military response is appropriate but officials are reluctant to act without statutory cover, a serious problem arises unless there is time to seek and receive legislative support.

Second, presidential action based on statutory authority has more political and legal legitimacy than action based on Article II alone. Article II actions leave the president without overt political support of Congress, which can later snipe at his decisions, or take actions to undermine them. We saw this happen, for example, in response to many of the Bush administration’s unilateral assertions of authority, and also to some degree in response to President Obama’s unilateral assertion of authority in Libya. This is a problem that grows with reliance on Article II over time. Also, of course, any subsequent judicial review of the president’s use of force is more likely to be upheld if supported by Congress.

Third, the president faces significant legal hurdles to detaining dangerous terrorism suspects over the longer term under Article II, and at a minimum would encounter substantial political and legal opposition if he attempted it. The Obama administration has shown no proclivity to detain terrorism suspects (outside the criminal justice system or in a combat zone like Afghanistan) other than those captured and detained under the previous administration.

But a future administration might regard such detention as necessary in some circumstances, and would have a harder time doing so beyond short periods without statutory authorization. Relatedly, an exclusive reliance on Article II would make targeting politically and legally safer than detention, an outcome that would run contrary both to the security interests of the United States (by eliminating the possibility of useful intelligence through lawful interrogation) and the interests of the individual in question (for obvious reasons).

3. The General Drawbacks of a New AUMF

While we believe there will be a need for a new AUMF, and while we discuss options for such a new statute in Parts II and III, we first pause to note the general downsides of a new AUMF.  As the discussion of inherent presidential power implies, a new statutory framework for presidential uses of force against newly developing terrorist threats might diminish presidential flexibility and discretion at the margins. At the same time, of course, it enhances the legitimacy of presidential action in domestic courts and with domestic public opinion.

This constraint-legitimacy tradeoff is commonplace. And to the extent that the constraint achieves legitimacy it promotes sustainable counterterrorism policy, politically and legally, over the long term.  A strong statutory basis makes it less likely that Congress or Courts will intervene later with constraints that dangerously hamper the President’s agility to respond to threats.

Some will also view a new AUMF as very costly to the degree that it extends, and makes indefinite, presidential military powers against terrorist threats. We have already explained why we believe that police authorities will not suffice against the threat, and we think that a realistic assessment of our national security needs likely makes the powers outlined here necessary and not optional. A related potential cost is a version of the hammer-nail problem. Giving the President new tools might lead him to use them even if they are strategically unwise. This is a fair concern and should inform both the timing of the statute (i.e. the statute should not be enacted until truly necessary) and especially its content (which we discuss in the next Parts).

Third are the international costs of a renewed AUMF. This is a complex issue. As a general matter a renewed and clarified AUMF – especially one that (as we propose below) articulates the U.S. view of international law – would contribute to the development of opinio juris under customary international law. So too would the reaction to the new AUMF. That reaction depends on the details of the legislation. To the extent that the legislation is seen as constraining the president in meaningful ways and in hewing to accepted international law, it would be viewed in a positive light internationally. To the extent that it is seen as making permanent an indefinite and geographically limitless war or in stretching international law, it would be viewed in a negative light internationally among NGOs and allied governments. And of course both reactions are likely to some degree.

The attempt to mitigate a negative reception abroad (and, in some quarters, at home), is one reason why we recommend below that any statutory reform in this area should emphasize compliance with jus in bello and jus ad bellum as well as the limited rather than unlimited nature of the authorization (conceptually and temporally). We recognize that the United States’ interpretation of some international law self-defense and law-of-war authorities is broader than our allies’ interpretation; legislating such limitations thus will not end debate. Nevertheless, acknowledging clearly that U.S. operations are to be conducted within, rather than beyond, traditional legal frameworks is an important step in mitigating friction with our allies, and prudent use of these legal authorities will be important in persuading allies that the U.S. position is a reasonable one.

II. Statutory Options to Address Extra-AUMF Threats  

We believe that a better and more stable alternative to reliance on the current combination of an aging AUMF, law enforcement authorities, and Article II is a new congressional authorization for emerging threats. As suggested in the Introduction, we do not maintain that the statute need be enacted immediately. But we do think that current trends likely point to the eventual need for a new statute.

A central challenge in designing such a statute is to provide sufficient flexibility to meet the changing threat environment while at the same time cabining discretion to use force and subjecting it to the sort of serious constraints that confer legitimacy and ensure sound strategic deliberation. There are many possible approaches, and they can be combined in various ways. We here outline three basic options and offer a preliminary recommendation.

1. Tie the Authorization to Article II or International Law.

Congress could authorize the President to use force that is consistent with his extant powers under Article II of the Constitution to protect the nation. That is, Congress could by legislation authorize the President to use force that he would otherwise possess pursuant to his inherent constitutional authority.

It might seem that a mere legislative re-affirmation of the President’s Article II powers against terrorist threats would have little significance. Recall, however, that the Bush administration sought from Congress and was denied similar authorization in September 2001. A congressional authorization of force that is co-extensive with the president’s Article II powers of self-defense would shift the exercise of those powers from Justice Jackson’s Category 2 to his Category 1, with all of the political and legal support entailed by that shift. When the president acts with Congress’s backing as opposed to on the basis of Article II alone, “his authority is at its maximum” and “the burden of persuasion would rest heavily upon any who might attack it.” In short, a self-defense regime is politically and legally more stable when backed by Congress.

There are several major problems with this approach, however. First, it would be entirely unclear against whom Congress was authorizing force, and thus Congress would be adding its political and legal weight to open-ended war of the sort that it was unwilling to support even in the early days after 9/11. Additionally, because members of Congress and the President often disagree about the precise scope of Article II powers, such a statute would be especially susceptible to later inter-branch disagreements. It is also not clear whether the President’s Article II authority includes detention powers. That problem could be fixed in the statute if Congress expressly authorized some powers of detention beyond existing criminal and civil authorities, but this would then reopen questions of against whom these authorities could be used.

A variant on the “authorized self-defense” approach would be to tie the President’s authority not to his Article II powers but to international law. That is, Congress could authorize the President to use force that is consistent with the international law of self-defense. The substantive scope of these alternatives would be similar in practice, and they share many of the same benefits and costs just mentioned.

2. Specify Terrorist Groups or Geographies

Congress could instead authorize the President to use force against specified terrorist groups and/or in specified countries or geographic areas. This would resemble the more traditional approach by which Congress authorizes force against state adversaries or for particular operations within foreign countries. Recent news reports have suggested that some in the administration and the military are deliberating about whether to ask Congress for just such a statute to address Islamist terrorist threats in some North African countries. This “retail” approach – in contrast to the “wholesale” approach laid out in the previous section – is the one that, among our three options, most restricts presidential discretion.

In theory, the retail approach is advantageous because Congress would specifically define the enemy (recognizing, however, the difficulties associated with the AUMF in drawing clear boundaries around transnational terrorism groups). Congress must under this approach stay engaged politically and legally as threats evolve and emerge; it must debate and approve any significant expansions of the conflict.

A downside of the retail approach is that Congress probably cannot or will not, on a continuing basis, authorize force quickly or robustly enough to meet the threat, which is ever-morphing in terms of group identity and in terms of geographic locale. The emerging array of terrorist groups across North Africa, with varying types and degrees of links, and posing potentially different (and again, changing) threats to the United States, illustrates the difficulties of crafting force authorizations that are neither too narrow nor too broad

3. General Criteria Plus Listing.

Based on current trends and the lessons from the past decade, we recommend a third approach: Congress sets forth general statutory criteria for presidential uses of force against new terrorist threats but requires the executive branch, through a robust administrative process, to identify particular groups that are covered by that authorization of force. One model to draw on, with modifications, is the State Department’s Foreign Terrorist Organization designation process.

Under this process, the Secretary of State – pursuant to specific statutory standards, in consultation with other Departments, and following a notification period to Congress – designates particular groups as terrorist organizations and thereby triggers statutory consequences for those groups and their members. We believe that a listing system modeled on this approach best cabins presidential power while at the same time giving the President the flexibility he needs to address emerging threats. Such a listing scheme will also render more transparent and regularized the now very murky process by which organizations and their members are deemed to fall within the September 2001 AUMF.  

The listing is approach is not without significant challenges, however. Some will claim that such a delegation to the president to identify the entities against whom force can be deployed would be unconstitutional. However, Congress has often authorized the president to use force in ways that leave the president significant discretion in determining the precise enemy. In light of this history, the waning of the non-delegation doctrine in other contexts, the congressional specification of the general criteria for the use of force, and the administrative, reporting, and timing limitations on the listing process described below, the constitutional objections can be overcome.

A more serious challenge is that the listing approach will appear to codify permanent war, and to diminish the degree of congressional involvement and inter-branch deliberation compared to the second approach. These concerns can be mitigated in several ways. First, the substantive statutory criteria governing this listing process should be as specific as possible. For example, a new AUMF might authorize force against “an organization with sufficient capability and planning that it presents an imminent threat to the United States.” Or it might authorize force against “any group or person that has committed a belligerent act against the U.S. or imminently threatens to do so.”

In setting out such criteria, Congress could make clear precisely what it means by key terms such as “imminent” and “belligerent act.” The criteria should, moreover, be expressly linked to international self-defense law. Compliance with that law is an obligation of the United States. And from a diplomatic and international legal-policy standpoint it is important that the United States government as a whole make clear that this is not an open-ended “global war on terror” but a cabined application of traditional self-defense to the new realities of non-state threats.

Second, at the front-end of the listing process, the administrative, consultative, and notification procedures should be sufficiently robust to ensure careful deliberation and strong accountability. At the same time, the statute should provide for emergency exercises of Article II power (which the Constitution arguably compels in any event), followed by a process for retroactive listing, to deal with rapidly-moving crises while providing strong incentive for the President to fold his actions into the statutory scheme.

Finally, a listing scheme should include thorough ex post reporting and auditing as well. At a minimum the President should have a duty to report to Congress in detail on the intelligence and other factual bases that led to the inclusion of particular groups on the list. The President should also have a duty to file detailed reports with Congress -- in a more robust form than the usually conclusory War Powers Resolution reports -- about how the statutory authorization of force is being implemented.

As has become typical in the exercise of its oversight of modern national security delegations, Congress would also likely deploy inspectors generals to perform audits on elements of the listing process. Finally, once a group is listed, there will be tremendous political incentive not to de-list it. So to ensure continual reassessment of the need for authorized force against particular groups, all listing should be subject to a review and renewal process (say, every two years) with an automatic sunset if not affirmatively renewed. (We discuss the role of sunset provisions as a general feature of all three proposed authorizations in the next Part.)  

III. Additional Issues

No matter which of the three broad statutory approaches Congress chooses, members will have to address several additional issues that arise almost inevitably in any authorization of modern asymmetric conflict against a shifting set of enemy non-state actors over time.

1. Special Targeted Killing Criteria.

Congress might consider statutory targeting guidance beyond mere authorization of force. The United States does not and will not conduct strikes against everyone whom it deems “part of” enemy forces. Rather, as administration officials have explained publicly, it uses far narrower criteria that take into account a target’s potential threat and importance within an enemy group. Particularly outside of hot battlefield zones, it may be worthwhile to write some of these judgments into law to mitigate anxiety that the United States claims an unrestricted right to wage war worldwide.

Congress might also state clearly as part of this legislation that the authorization for force applies only overseas. Moreover, the administration has stated that it applies heightened screening criteria for the targeting of U.S. nationals, and both in a major speech and in a leaked Justice Department White Paper, it has laid out circumstances in which it regards the targeting of an American overseas as consistent with due process. Congress could write such standards into law – or, to the extent it disagrees, write different ones.

The general point here is that a use of force authorization does not need to be unqualified, especially with respect to locations in world in which U.S. troops are not in a day-to-day sense deployed and confronting the enemy. Congress could, rather, authorize force but calibrate the authorization for a variety of anticipated specialized circumstances that call for heightened review.

2. Detention v. Targeting.

Congress might also consider distinguishing detention and targeting. Broadly speaking, it appears that the two take place under similar standards, namely, whether the individual or individuals are “part of” or “substantially supporting” enemy forces. Yet there is a reasonable argument, in the context of this sort of micro-targeted conflict off the hot battlefields, for treating targeting and detention as legally distinct.

More specifically, one can argue that detention authority should sweep more broadly than targeting authority. Outside of regular zones of armed conflict, after all, the United States is highly unlikely to target mere supporters of enemy forces, so narrower criteria might be workable for lethal military force. Having a somewhat broader criteria (though no broader than the current member/supporter test), at least for short-term detentions, would permit the holding of incidental captures until they can be identified and processed and, if need be, transferred to law enforcement or some foreign authority with a legal basis for longer-term action. 

3. Accountability and Review Processes.

In overseas counterterrorism operations that are inherently secretive by their nature, accountability is critical to ensuring that executive conduct is lawful and prudent and thus to maintaining the legitimacy of the conflict over time. In any future authorization, Congress should build accountability into the authorization itself. It should require public reporting of matters that can be discussed openly, such as the number of strikes and operations, their geographic sweep, and estimates of civilian casualties.

It should demand maximum feasible openness about the procedural elements of listing groups as covered entities and about the legal opinions that underlie the American legal framework. And it should require detailed classified reporting and auditing from relevant department and agency inspectors general as to both the vitality of internal processes and the integrity of the intelligence underlying the listings and claims about civilian and enemy deaths. Consideration might also be given to robust internal administrative processes for certain non-battlefield target selection, and for the creation of formal compensation mechanisms for victims of errors in non-battlefield settings. All of these restrictions may sound like a significant erosion of traditional presidential authority to conduct war. But the conflict Congress is authorizing here is a novel hybrid conflict, and the regime itself – including the accountability mechanisms for that regime – need to reflect that reality.

4. Sunset

Finally, to head off the possibility – or the perception – that Congress is authorizing a perpetual war against an undefined series of groups, the entire regime should sunset after a statutorily defined period of years. Once force is authorized, there is tremendous political incentive not to revoke that authorization – a step that has implications for ongoing operations, as well as for the nation’s sense of itself as at war. To ensure continual reassessment of the need for authorized force, the authorization should be subject to legislative review and renewal (say, every two years), with a default sunset if Congress does not affirmatively renew the granted authority. In the event of a sunset, the legal basis for detentions pursuant to the conflict might evaporate, so the statute would have to address whether and how existing detentions could be extended following a sunset.

IV. Conclusion

As the AUMF becomes obsolete, we are approaching the end point of statutory authority for the President to meet emergent terrorist threats. There are significant downside risks to any successor legal regime to the AUMF and to the option of doing nothing. We have argued that the approach of statutory criteria and listing described above is the least bad option for the country to meet emerging terrorist threats. But at a broader level what is most important is that any legal reform should take account of lessons from the past twelve years’ experience operating under the AUMF. Our broadest aim has been to provide a menu of options for incorporating those lessons into any new legislation that emerges.

To read the full essay, please click here


Robert Chesney is a professor at the University of Texas School of Law, a nonresident senior fellow of the Brookings Institution, and a distinguished scholar at the Robert S. Strauss Center for International Security and Law. He is a cofounder and contributor to the Lawfare blog and writes frequently on topics relating to US counterterrorism policy and law.


Jack Goldsmith is the Henry L. Shattuck Professor of Law at Harvard University and the author, most recently, of Power and Constraint: The Accountable Presidency after 9/11 (W.W. Norton, 2012) and many other books and articles related to terrorism, national security, and international law. Before coming to Harvard, Goldsmith served in 2003–4 as assistant attorney general, Office of Legal Counsel, and in 2002–3 as special counsel to the general counsel to the Department of Defense. Goldsmith holds a JD from Yale Law School, a BA and an MA from Oxford University, and a BA from Washington and Lee University. He clerked for Supreme Court justice Anthony M. Kennedy, Court of Appeals judge J. Harvie Wilkinson, and Judge George Aldrich on the Iran-U.S. Claims Tribunal. Goldsmith is a fellow of the American Academy of Arts and Sciences.


Matthew Waxman is a professor of law at Columbia Law School and an adjunct senior fellow at the Council on Foreign Relations. He previously served as principal deputy director of policy planning (2005–7) and acting director of policy planning (2007) at the US Department of State. He also served as deputy assistant secretary of defense for detainee affairs (2004–5), director for contingency planning and international justice at the National Security Council (2002–3), and special assistant to National Security Adviser Condoleezza Rice (2001–2). He is a graduate of Yale College and Yale Law School. He served as law clerk to Supreme Court justice David H. Souter and US Court of Appeals judge Joel M. Flaum. His publications include The Dynamics of Coercion: American Foreign Policy and the Limits of Military Might (Cambridge University Press, 2002).


Benjamin Wittes is a senior fellow in governance studies at the Brookings Institution and codirector of the Harvard Law School–Brookings Project on Law and Security. His most recent publication is Speaking the Law (Hoover Institution Press 2013), cowritten with Kenneth Anderson. He is the cofounder of the Lawfare blog.


Letters to the editor may be sent to definingideas@stanford.edu. Editors reserve the right to reject or publish (and edit) letters.