If I understand Bernard Harcourt’s post correctly, he believes that torture should be legal when it is “necessary” (that is, when the necessity defense is satisfied), whereas he attributes to me and a coauthor the view that if torture is to be legal, the legal regime should consist of ex ante rules (for example, authorizing certain practices and banning others) rather than an ex post standard like the necessity defense. He then criticizes us for advocating the “legalize and regulate” approach. However, Harcourt himself advocates the “legalize and regulate” approach since the necessity defense is just a way to “legalize and regulate,” as we pointed out in the article, repeating a point made by Sandy Levinson several years ago. Harcourt is seriously confused; he also misses the point of the article.
The purpose of the article was not to make this argument, but to criticize what we called the “outlaw-and-forgive” approach to torture. Our survey of the law and philosophy literatures revealed that many, perhaps most, scholars, while condemning torture very strongly, argue that it is morally justified in extreme circumstances. These writers try to reconcile their opposition to torture and their belief that it may sometimes be justified, by advocating a legal ban accompanied by political or legal forgiveness when justified—in the form of exercises of prosecutorial discretion, pardons, suspensions of sentences, and the like. (A more common example is euthanasia, which appears not to be prosecuted in many places where it is illegal.) We argued that there is little sense in this proposal: torture should either be legally banned or regulated. For the argument, see here.
The source of Harcourt’s confusion is, I believe, our statement that the necessity defense is a form of outlaw-and-forgive. That made it sound as though we disapproved of the necessity defense. However, we described the necessity defense as “outlaw-and-forgive” because the scholars we criticized did, but our main argument in this setting was that the necessity defense is in fact not an example of outlaw-and-forgive, since it simply legalizes torture when it is necessary, as I noted above. The contrary view rested on a confusion between a legal defense and extra-legal forgiveness (like a pardon). We later turned to the question whether, if torture is to be legal, the necessity defense or a system of rules would be superior. In that discussion, we did not take a strong position, instead pointing out that the answer to this question depends on various empirical factors of which we were ignorant.
The release of the torture memos has no bearing on our critique of outlaw-and-forgive. It could have a bearing on the subsidiary rules-versus-standards question that interests Harcourt. But I believe that his conclusion is premature. If the Obama administration chose to prosecute CIA agents, for example, it could turn out to be easier to convict people who violated the clear, rule-like guidelines, than people who had received a vague order to torture “if necessary.” Alternatively, it could be the case that agents given the standard would be more likely to act appropriately. It is impossible to know on the basis of the memos alone; we first need more information as to what happened.