The post-9/11 era is over, some say, but many hard questions of national security law and policy remain. Some now contend that the Obama administration not only should shutter Guantánamo, but also should foreswear reliance on preventive detention more generally when it comes to non-state actors such as al Qaeda.

David Cole is not so sure. Well-known for his criticisms of the war on terrorism, Cole has significant concerns regarding the use of preventive detention. Nonetheless, he endorses a very narrowly tailored policy that would preserve the military detention option subject to certain substantive and procedural restraints.

I have much sympathy for the general thrust of Cole’s approach, which seeks to address criticisms of the post-9/11 military detention system while recognizing that some form of detention remains appropriate at least so long as the military remains engaged in actual combat operations against al Qaeda and Taliban forces. And I wholeheartedly agree that this situation calls for legislative intervention. The traditional rules associated with determining eligibility for military detention do not map onto the non-state actor scenario terribly well; legislation not only can produce a more carefully calibrated system, but also can address the objection that no current source of law adequately justifies the use of detention in the first place. Whether Cole has identified an appropriate blend of substantive detention criteria and procedural safeguards, however, is less clear.

Consider Cole’s proposal for redefining the scope of detention authority. The Bush administration’s standard for detention, employed in the Combatant Status Review Tribunal (“CSRT”) system, encompasses:

An individual who was part of or supporting Taliban or al Qaeda forces, or associated forces that are engaged in hostilities against the United States or its coalition partners. This includes any person who has committed a belligerent act or has directly supported hostilities in aid of enemy armed forces.

Cole would narrow this definition in two respects. First, the mere fact of membership in or provision of support to al Qaeda or the Taliban would no longer justify detention; the government would also have to show that the person’s conduct contributed in a relatively direct manner to armed hostilities against the United States. Whether this is genuinely different from the CSRT standard depends, however, on how “direct” the link to hostilities must be in Cole’s model. Cole makes clear that his test would reach not just those who personally participate in violence, after all, but also those who plot or command violence, those who train for violence, and presumably also those who provide support to violence such as constructing explosive devices. Might it also encompass financiers, recruiters, or those who provide various forms of logistical support, at least when the government can show such persons knew or should have known that their actions would contribute to violence? If so, there may be only a marginal difference between Cole’s model and the status quo. And if not, we might then ask whether it really makes sense to draw the line in this way. Why shouldn’t an al Qaeda member be subject to detention if he or she provides financial support knowing or intending that al Qaeda will use the money for violence?

Cole’s model also differs from the CSRT approach in that his applies only to persons linked to al Qaeda or the Taliban, while the CSRT’s extends to action on behalf of any entity “engaged in hostilities against the United States or its coalition partners.” This move responds to concern that the government has detained persons at Guantánamo based on ties to groups that may be ideologically compatible with al Qaeda yet are distinct from it. But the proposed solution goes too far in attempting to address this issue, particularly if Cole’s model is meant to apply to detentions occurring overseas in connection with continuing armed conflict in the Afghan-Pakistan border region. Al Qaeda and the Taliban are not the only groups engaged in armed hostilities against the United States and the Afghan and Pakistani government today. A dozen or more groups with varying degrees of interconnectedness participate in the insurgency. Indeed, the very notion of group affiliation can be tenuous and ephemeral in this setting, making it difficult to implement a test employing strict associational distinctions. A requirement that the military discern the associational status of its detainees and hold only the subset of them shown to be associated with al Qaeda and the Taliban seems unwise and potentially unenforceable. Better to maintain the CSRT definition while actually adhering to its limits.

Whatever criteria are used, the next task is to identify the procedural safeguards that should be employed in determining whether a detainee in fact satisfies them. Many believe that the government has held individuals in custody at Guantánamo and elsewhere based on little or no evidence. Responding to such concerns, Cole advocates a screening process in which the government must provide clear and convincing evidence (presumably to a judge), with the detainee represented by counsel and with the government barred from using “ex parte” evidence (that is, evidence shown only to the judge, but not the detainee or his counsel).

This raises two sets of issues. As an initial matter, the Supreme Court this past summer held in Boumediene v. Bush that federal judges can and must exercise judicial review over CSRT detention decisions. Do we need further reform? We certainly do in the sense that it is the responsibility of Congress to flesh out and make uniform, within constitutional bounds, the procedural safeguards to be employed during habeas review. But it does not follow that Congress should follow all of Cole’s particular prescriptions. Lacking legislative direction, for example, the district judges carrying out habeas review in recent weeks have settled on a “preponderance of the evidence” standard. Cole calls for raising that standard to the “clear and convincing evidence” level, but it is not clear that it would be wise to impose a stricter evidentiary burden. Applying the “mere” preponderance standard, the first federal judge to render a decision on a habeas petition since the Supreme Court’s ruling in Boumediene—a Bush appointee, notably—recently held that the government lacked sufficient evidence to justify the detention of five out of six detainees whose status was at issue (including Mr. Boumediene himself). This suggests that the current standard has ample teeth.

Of course, not all military detainees are held at Guantánamo. Today, the majority are held in Afghanistan and Iraq, having been captured in those zones. No federal judge has ever asserted habeas jurisdiction over those detainees, and, though some such attempts are pending at the moment, it is not at all clear that they will succeed. Cole’s model would produce a sea change if applied in that setting. Would such a change be desirable as a policy matter, even if not legally compelled? In accordance with counterinsurgency doctrine, some such steps might be in order; the United States would be wise to consider any measures that would enhance the perceived fairness of its detention practices without unduly compromising safety. Whether such measures should involve review by a federal judge, however, let alone the degree of judicial review contemplated in Cole’s model, is not obvious.