Engaging Glenn Greenwald

I am hereby ending my boycott of Glenn Greenwald. I'm doing it not because his latest post about me refrained from personal attack or adopted the sort of civil tone to which I think the public debate should aspire. It didn't. As such, I am surely, in breaking my embargo, also breaking the commitment I made not to engage with someone who starts from premise that I am evil or on the take. I'm responding to his post, because the honest truth of the matter is that it made me think about some things---and while I could frame that thinking without reference to him, I don't really want to. I have made my point about civility, I hope. And I'm sure he has made his point about my deep corruption. So now, as Ezra Pound once wrote of Walt Whitman, "I make a pact with you, [Glenn Greenwald]. I have detested you long enough. . . . Let there be commerce between us." Greenwald makes several points about a brief post of mine criticizing New York Times Public Editor Margaret Sullivan for a column she wrote on the paper's coverage of civilian casualties in drone strikes. Sullivan's column had---while acknowledging the Times's important contributions to the drones debate---nonetheless taken the paper to task for not having "aggressively challenged the administration’s description of those killed as 'militants'---itself an undefined term." I found this critique bizarre, writing:

The Times is not an advocacy organization whose job it is to “aggressively challenge” the government’s claims of the rates of civilian casualties---except to the extent that those claims are untrue. And while there’s certainly a debate to be had as to the right way to count civilian deaths, it’s not at all clear that the government’s reports of low rates of civilian casualties are off base. Ironically, the Times has actually done a terrific job compared to many other news organizations of sifting through the disparities between different claims regarding the number of civilians killed in drone strikes. See, for example, this excellent story by Scott Shane from a while back.

Greenwald responded that this was a demand for "servile journalism":

It's amazing that someone not only believes - but is willing to say publicly - that it is not the job of a newspaper to "aggressively challenge" government claims on a highly controversial assassination program that is shrouded in secrecy and uncertainty. That, more than anything else, is the core purpose of journalism (at least in theory): the reason "freedom of the press" is protected in the First Amendment. And it's precisely the media's systematic failure - more accurately: its unwillingness - to engage in this function that has produced the last decade's most destructive outcomes. Wittes' caveat that newspapers should "aggressively challenge" government claims only when "those claims are untrue" is circular and nonsensical. The only way to find out whether government claims are untrue is by aggressively challenging them. A failure to do so ensures that even the most dubious and unproven of government assertions go unchecked. Indeed, Wittes himself has previously argued that secrecy surrounding the drone program is excessive. More to the point, the presumption Wittes advocates is exactly backward: government claims are not entitled to a presumption of truth by media outlets unless and until they are proven false. The opposite is true: they ought to be treated with extreme skepticism by media outlets unless and until they are proven true. That is what "adversarial journalism" means.

Greenwald's surface point is easy to address---as it responds to a misapprehension of what I meant, one to which I'm afraid I contributed with an ill-considered phrase that lends itself to the reading Greenwald adopted. So let me clarify: I did not mean to argue and do not believe that the New York Times should simply accept government civilian casualty numbers unless and until someone comes along and proves them wrong. My point, rather, was simply that the New York Times should not take at face value either the insistence by drone critics that the government's estimates are wrong or unreasonable. And Sullivan's column seemed--still seems on rereading, actually---to be faulting the Times for not taking the same view of the matter as do critics of drones strikes. My view, for whatever it's worth, is that the New York Times has two obligations here. One is to report the controversy accurately, which I think it has done admirably. The other is, to the extent possible, to come to its own independent understanding of the rate of civilian casualties and to report that understanding. This it has not done---though in its defense, the investigation might be an unmanageably difficult undertaking. But Greenwald's post also made me think about a deeper point---one which he does not raise directly but which is latent in Sullivan's critique of the Times and in Greenwald's critique of me. How one counts civilian casualties depends pervasively on certain ex ante assumptions one brings to the project of estimating. Those assumptions may account for very large fractions of the disparities in estimates. And I don't really know what I think the right assumptions are. Counting civilian deaths in drone strikes would be very hard even if we all agreed on what we were counting. At a very basic level, there's a data problem. Because reporting from these regions of the world is terribly dangerous, estimates are largely based on reports in anonymously-sourced media stories that rely on Pakistani Army claims that "militants" were killed. Under the best of circumstances, as critics rightly point out, this leaves the estimators at the mercy of the errors, the self-serving accounting, and the agendas of the Pakistani Army sources. But the data problem is compounded by the fact that we don't all agree about what we're counting. It matters a lot what normative, factual, and legal assumptions an estimator brings to the table.  "Militant" is not a term of art, so one's assumptions about who represents a lawful target matter enormously in any effort to count righteous versus collateral kills---and in how one classifies the deaths of people who were at or near the target. From the government's perspective, if a drone strike hits a lawful target---say, a camp linked authoritatively to Al Qaeda and filled with guys with guns---and it kills a bunch of guys with guns, one might very plausibly start with the presumption that these are all combatants unless one has some significant indication otherwise. Conversely, if we start with the assumption that these are not belligerents but civilians whom one can target only to the extent they are directly participating in hostilities, one may count civilians deaths very differently in the absence of some affirmative evidence of direct participation among those targeted. In other words, when we read news stories describing the deaths of "militants," it is perfectly possible---likely even---that the term is describing people who are plausibly described as "militants" using one set of assumptions and civilians using another. Indeed, that was inescapable conclusion of the New York Times's famous "kill list" piece---which led Greenwald to vociferously criticize the media's repetition of the term "militant":

By “militant,” the Obama administration literally means nothing more than: any military-age male whom we kill, even when we know nothing else about them. They have no idea whether the person killed is really a militant: if they’re male and of a certain age they just call them one in order to whitewash their behavior and propagandize the citizenry (unless conclusive evidence somehow later emerges proving their innocence). What kind of self-respecting media outlet would be party to this practice?

While I'm confident that I know what I think the New York Times's role in this debate is. I have to say that I'm not confident I know what the proper set of assumptions are to bring to the conversation about counting civilian casualties. There have been variety of efforts to count civilian deaths---from the New America Foundation, from the Bureau of Investigative Journalism, from the Long War Journal, and from several Pakistani organizations---and they have led to wildly disparate outcomes. There have also been critiques of these efforts---from the Columbia Law School Human Rights Clinic and from researchers at Stanford Law School and NYU School of Law. While I have read a lot of this material, I have not sat down and studied these documents carefully side by side---zero-basing the assumptions people are making and the means of counting they are using. I'm also keen to understand whether and to what extent some of these seemingly disparate estimates can ultimately be reconciled with one another. In other words, can we isolate the issues that are generating the disparities---and are they ultimately empirical issues or definitional ones? To put the matter simply, Greenwald's post has convinced me that I need to formulate more completely than I have to date my own views about how to count civilian casualties---indeed, about whether it's even possible. It's an issue to which, in the coming months, Ritika and I plan to devote some energy. And in all sincerity, I'd like to thank Greenwald for forcing me to face it more squarely.