Friday, 3 May 2013

Sam Harris and the Ethics of Torture

Sam Harris has been having a bit of a rough time recently. In addition to being denounced to the point of ridicule as a neo-con apologist in the New Media, the old media has started taking shots. Glenn Greenwald's piece in The Guardian, it seems, was the final straw, and now Harris has revised the section of his website in which he takes the time to address in more comprehensive terms his exact position on his alleged "thought experiments" that have angered the liberal intelligentsia.

As someone who wrote a fairly lengthy critique of Harris a while back, I was all up for this - I've never been of the impression that Harris is a genuinely malicious individual intent on providing moral justification for war crimes, and it's always worth giving people an opportunity to explain themselves. The problem with Harris isn't that he's explicitly bad-natured,  but that his past writings on just about anything besides neuroscience reveal him to be a lazy thinker who fails to properly consider the implications of the things he says in public. Unfortunately, his latest revision of his response to critics smacks of a thinker wanting to eat his cake and have it; in one breath denouncing Abu Ghraib as inexcusable, and the next calling for the moral necessity of torture. 

What follows is ostensibly a critique of Sam Harris' thinking on torture, because it's the aspect of Harris' thinking which I find to be the most troubling and the most morally inexcusable. However, I feel some of the issues with Harris' writings on torture can be extended towards much of his other writings on other subjects. In particular, I think Harris is particularly troublesome when he turns his attention towards foreign policy, the war on terror and ethics. His contempt for researchers in the field of terrorism studies is a separate piece in itself. So whilst this piece will address torture, it should hopefully serve as a partial critique of Harris' writings in general. So without further ado, let's jump into some of the points Harris raises.

Unfortunately Harris' response to critics on torture is largely unchanged, and he's still rehashing old points about time-bomb scenarios and posing problems he claims nobody has addressed head-on. It is true that the majority of the criticisms of Harris use his support of torture as an in-itself indictment of his writing rather than addressing his arguments. It happens that his arguments are extremely weak and rely on straw men and false equivalences, but I suppose this has given Harris the impression that his opponents don't have any ammunition. The real reason support of torture is itself an indictment is because torture has been debated extensively for centuries in philosophical and legal discourse, and due to the hard work of an extremely dedicated human rights legal movement, is now considered a jus cogens crime in international law. That is, it's a crime on which moral debate is presumed to have ended, and anyone committing such a crime can be subject to universal jurisdiction. What that means is, say you're a high-ranking government official responsible for torture in, oh I dunno, Chile - you can be arrested in England with an arrest warrant written by a Spanish judge and tried in an English court and found guilty of your crimes.

Legal positivism is, of course, no argument in and of itself - it won't do to simply say it's illegal and that's that. Not least because, strangely enough - Harris bizarrely argues that torture's (il)legal status should remain, but that it should still be employed in extreme circumstances. Maybe it's a lack of imagination on my part, but I fail to see the point of having reams of treaties, statutes, case-law judgments, etc. converging on torture's illegality if security officials going to on a whim suddenly decide that Scenario A is "extreme" enough to warrant torture. Harris' rebuttal is to compare torture to trespassing - as trespassing on someone's property may be illegal, but ethically necessary, so it goes with torture. Yet unless one gets specific with thought experiments it's difficult to engage in them properly; as it stands, any well-drafted law on trespassing will have caveats built into it to not incriminate a person performing an ethically necessary act, and a common-law system will give a fair amount of discretion to a judge on deciding whether a strict application of the law is appropriate for a trespasser who was, say, using the land as a shortcut to take a dying person to the hospital.

Anyway the comparison between torture and trespassing is spurious since it's quite plain that there are different justifications for outlawing trespass and torture. The former is, amongst other things, to ensure property rights are legally meaningful. Torture, on the other hand, has reams of reasons for prohibition ranging from the consequentialist (torture isn't effective and may be counter-productive) to the deontological (torture is a degradation of humanity regardless of consequence) to virtue ethics (a torturer does not possess morally desirable characteristics) to straightforward old-fashioned international legal rules about reciprocity (if you throw away the rulebook, so will everybody else). Using all moral frameworks except from an extremely crude version of utilitarianism in fact bodes not so well for the pro-torture advocate. As it stands, the prohibition on torture within international law is non-derogable, and is so precisely because exceptions to the rule in "extreme" circumstances quickly become the norm. If you think  this applies only to other countries and not nice western democracies, then check out the Constitution Project's report on US detainee mistreatment and then realise that Bush-era torture policies extended to far more than the 9/11 hijackers. What was justified in exception became the norm.

Harris is, however, right to say that the fact that torture was abused in the past does not logically entail that it will always be abused and a 'perfect' torture policy (ahem) is conceivable in a conceptual sense. That is - one that is only used in the most extreme ticking-bomb scenario and that always yields the information that it is aimed at extracting. Yet even if you allow all the premises (which I wouldn't, but let's bite) that this New Torture Regime will ensure that it its use only ever occurs in a ticking bomb scenario and that its use would yield useful information that would lead to the disarming of the bomb, it's still far from a foregone conclusion that torture is moral because you are faced with an important question, and one which I think any would-be torturer would feel extremely uncomfortable about considering its implications. That question is:

How much suffering will you permit?

The argument for Perfect Torture in a Time-Bomb Scenario relies on a fairly crude utilitarian calculus which states that the suffering of one person (the terrorist) is outweighed by the suffering that would be visited upon the inhabitants of City X, which has a bomb rigged to go off in its center. So, the argument goes, if torture is the only means of getting the terrorist to talk, we should use torture to extract that information. That is, in a nutshell, the only reasoning by which the torture-justifier supports his or her argument, and its difficulties can be highlighted by another very simple question:

What if the terrorist manages to withstand the torture, but will only talk if his innocent 6-year old daughter is tortured?

The ethically consistent pro-torture advocate, of course, would be forced to conclude that the torture of the terrorist's family and/or friends is also acceptable, providing the population of City X is sufficiently high to outweigh the suffering of those being tortured. It's not enough to say that the terrorist is the one who planted the bomb and is therefore responsible for his own torture, since the torture is being justified on the basis of lives that would be saved, not on the terrorist's responsibility for his own treatment. The time-bomb scenario is also inadmissible for other reasons relating to realism of the scenario or the fact that the terrorist could simply lie in order to send the authorities on a goose chase, etc. but as a conceptual moral argument it fails because it allows for the torture of not just the individual terrorist, but also for the torture of innocents if that'll get the prisoner talking. Oh, and yeah - this kind of thing does happen, so we're no longer arguing about simple hypotheticals. If it is morally possible to reach a point where time-bomb torture justified on the utilitarian principle of "City X would be saved" would also not permit the torture of innocents, please email me a reference.

Collateral damage and torture

The other major point that Harris presents as a kind of trump-card in his few discussions of torture is that anti-torture advocates don't seem quite so bothered about collateral damage as they do about torture. Now, right off the bat, I'm going to object, because most of the people I've seen arguing against states' torture policies are also totally not OK with illegal invasions, cluster bombs and the like, either. When you're an activist, you really can be in two places at once. It is slightly frustrating, however, to see that Harris, having argued this since 2004 in The End of Faith, either hasn't been directed to or has willfully ignored the literature on torture to the point where he still thinks this is an interesting argument.

Harris' argument, like so many of his propositions, relies on a lot of assumptions. That's permissible if one is genuinely doing hypothetical moral philosophy (If A is...then B), but it's slightly irritating to see Harris use direct examples of individuals in the war on terror like Osama Bin Laden, and talk about actual real-world conflicts and then claim he's only posing hypotheticals when he's criticised. Again, eating the cake and having it. So anyway, the argument Harris is putting forward is basically this: wars will always involve the deaths of innocent civilians, but may still be sometimes justifiable on other grounds. If one therefore justifies the deaths of civilians as a means to an end, one must also justify the torture of some captives in extreme circumstances as a means to an end. That's basically his argument, I'll quote it so readers don't think I'm being disingenuous:
My argument for the limited use of coercive interrogation (“torture” by another name) is essentially this: If you think it is ever justifiable to drop bombs in an attempt to kill a man like Osama bin Laden (and thereby risk killing and maiming innocent men, women, and children), you should think it may sometimes be justifiable to water-board a man like Osama bin Laden (and risk abusing someone who just happens to look like him).
Again, which is worse: water-boarding a terrorist or killing/maiming him? Which is worse, water-boarding an innocent person or killing/maiming him?  There are journalists who have volunteered to be water-boarded. Where are the journalists who have volunteered to have a 5000-pound bomb dropped on their homes with their families inside? 
This is, quite frankly, silly. Nobody is arguing that torture is morally worse than murder. Yet simply saying that some ethically permissible acts in war are worse than torture is not a ethical argument in favour of torture. For instance, wars involve the mobilisation of large armies, and many of the men and women in those armies are likely to be put in harm's way and a lot of them will die. That is a tragedy of war, and in a hypothetical just war scenario it is unavoidable, but necessary.

It does not, however, follow from "soldiers might/probably will die" that mistreatment of the soldiers is somehow permissible, simply because it is less bad than the worst thing that can happen - namely their deaths. Or that, say, failing to provide them with adequate food is ok on the grounds that, whilst unethical, it's still not as tragic as their deaths. I just don't see any interesting point in conflating the worst and yet inevitable aspects of just war - namely civilian deaths, as a framework for a discussion of the ethics of detainee treatment.

Harris' protestation to those who consider the comparison between torture and collateral damage to be irrelevant is to pose a scenario in which one had to choose between torturing a terrorist or bombing civilians.  "Torturing the terrorist should seem like the more ethical option", he says, but there is also the possibility of, you know, doing neither. Once you've gone out of the ticking bomb scenario, with its assumptions of Perfect Torture and Perfect Results, into a scenario in which a General has to choose between bombing a city or torturing a terrorist, you've defeated the central argument the time-bomb scenario is based on - that we cannot disarm the bomb without torturing the terrorist. If we're choosing between shooting a missile or torturing a terrorist in order to get valuable information, the only qualitative difference between the two scenarios is that with the missile, we can actually choose whether or not the bomb goes off. The ethically sensible answer is therefore to not shoot the missile and not torture the terrorist. Unfortunately, though, it's difficult to go into too much more detail on this because Harris doesn't give any actual examples of how a military officer might actually be in the position where he or she would be forced to choose between torture of a terrorist or bombing a city, and it really is such a ludicrous scenario that it warrants a little bit more than postulating that it "could happen." The example he does provide of a drone strike against a Taliban leader in which 12 innocent people were killed says absolutely nothing about the ethics of torture at all. Harris just simply says that had his wife been tortured in order to obtain "relevant intelligence" rather than killed by a missile, everything would have been OK. It seems to me that in this case the authorities actually had all the relevant intelligence, and proceeded with a drone strike that killed a bunch of innocent people, anyway - how would waterboarding the guy's wife have changed the practical realities of the military operation - would she be able to tell the authorities when all the innocent people would be out shopping? 

I'll just make this brief, practical, point - if a military officer is in a position whereby the only information that can inform his decision whether or not to bomb a target is extracted through torture - it's probably a good idea to get better intelligence.

The worst versus the still-awful-but-not-quite-the-worst

Besides, I would be interested to know how far Harris and other pro-torture advocated would take their moral reasoning regarding collateral damage as "the worst" and torture as "less than the worst" as a reason to permit torture. Consider another horrendous, yet common practise in war - that of the rape of civilian women in an occupied territory. This is, of course, a morally reprehensible act and it's rightly treated by international courts as a war crime, but according to the Harris logic, raping the women in a village is less bad than carpet bombing the village into oblivion, therefore any supporter of just war theory also has to permit rape in extreme circumstances.

You could, of course, argue that the torture is intended to serve a purpose of extracting vital information whereas rape is simply barbaric acquisition of the spoils of war, but you would be wrong. Rape is utilised by unscrupulous war criminal commanders precisely because it is an effective means of humiliating the enemy into submission and preventing an all-out conflict. None of the horrors of rape in wartime matter to the logic of the pro-torture advocate - all that matters is that a full-blown conflict has been averted and nobody was killed, since collateral damage is "the worst", whereas rape/torture is "the still-awful-but-not-quite-the-worst". Or consider our ticking bomb thought experiment above. Let's pose another question:

Suppose that the terrorist is a woman - is it morally permissible to rape her to get her to talk?

These are undoubtedly disgusting things to have to consider - which is precisely why the pro-torture argument is so useless as an ethical position. Once we follow it to its logical conclusion, we are forced to permit the most fundamentally barbaric acts of depravity simply because they appear to be "less worse" than the deaths of civilians. It is, conversely, far more ethical to treat torture with a kind of hands-off absolutism which permits it in no circumstances whatsoever, regardless of consequence or spurious comparisons to other seemingly less desirable (though separate) outcomes.

Torture is hardly the only instance in which we permit this kind of absolutism, either. Harris is a neuroscientist, and is surely aware of the benefits that would be brought to his specific field of research if him and his colleagues were permitted to experiment on live human subjects. It's entirely possible that the field could be advanced by leaps and bounds if even a small number of live human subjects (say, terrorists) were permitted to be analysed in this way. It could, in fact, lead to such advances in the field that thousands, if not millions of lives are saved in years to come if even a small number of humans were able to be used as lab rats for scientific experiments. We don't, however, permit this. Why? The reason is obvious, and there's no need to discuss it in detail, because it relies on the same arguments that render torture a morally reprehensible act.

Some other points

So I think I've addressed the bulk of what makes Harris' argument so weak as an ethical defense of torture, but here's a cheat sheet of the major points in case you need a cocktail-party rebuttal:

1) Harris tries to have it both ways by arguing for torture's utilisation in the war on terror whilst simultaneously arguing that he is merely posing hypotheticals about torture's use "in principle."

2) To the pro-torture advocate, it remains to be seen how they would permit some forms of torture (waterboarding) but disallow others (rape) since the only argument for torture is based on numbers of people dying, rather than the degree of suffering of the tortured.

3) Allowing torture, therefore, allows for the torture of many more people than merely the terrorist, and could even include the terrorist's children.

4) The collateral damage argument is spurious, because simply saying that worse things happen in war does not permit a "gloves are off" scenario (even in extreme circumstances) with regards to other aspects of war such as detainee treatment.

5) Harris' argument is notably weak because it is attempting to argue in the realm of ethical and theoretical abstraction for a policy that begins to show its true colours the more and more one applies it to a practical scenario.

There are a few other points that I'd like to address Harris on that are separate to the main arguments, but still worth addressing because they come up quite regularly in popular discussion of torture. Here they are.

1) Journalists have submitted themselves to waterboarding, therefore it is not torture

Harris doesn't quite make this exact point, but says something very similar to it:
There are journalists who have volunteered to be water-boarded. Where are the journalists who have volunteered to have a 5000-pound bomb dropped on their homes with their families inside?
These journalists generally regret their choice and nearly unanimously come out after withstanding only a few seconds of waterboarding to insist that it's a form of torture. Take this video of right-wing radio host Mancow Muller being waterboarded. His intention was to prove that it wasn't a form of torture, yet all it took was a few seconds to convince this die-hard right-winger that waterboarding isn't frivolous. It's probably more accurate, therefore, to say that journalists ill-advisedly opt to be waterboarded and regret their decision later on. That leaves us where, exactly? If suddenly a bunch of journalists opt to be placed in an iron maiden on the erroneous belief that the spikes won't penetrate the flesh so much, does that say anything about whether or not it's an acceptable way to treat prisoners in wartime?

2) Keep it illegal, but acknowledge that people will break the law in a time-bomb scenario

The problem with this argument is that it turns the legal prohibition on torture into a discretionary opt-out, thereby rendering the illegality meaningless. It preferences the whims of interrogators, who are far more likely to view their scenario as especially important, over the consensus of the international legal and philosophical community, who have worked hard to make torture a non-derogable strict liability crime for wider ethical reasons.

One could make a comparison with other forms of strict liability - take drunk driving. It's probably quite unlikely that even the majority of people over the legal alcohol limit will put someone in danger, much less kill someone. Yet if we take the same approach that drunk driving should not be a strict liability crime and that drunk drivers should feel free to break the law if they deem the circumstances extreme enough to warrant hopping in the car, it renders the whole principle of strict liability meaningless and the prohibition in reality doesn't exist at all.

3) Sam Harris supports torturing innocent people, too

Yes, he does:
If you think it is ever justifiable to drop bombs in an attempt to kill a man like Osama bin Laden (and thereby risk killing and maiming innocent men, women, and children), you should think it may sometimes be justifiable to water-board a man like Osama bin Laden (and risk abusing someone who just happens to look like him).
My italics. At the very least, this is a frank admission of where a pro-torture policy will end up taking you. Inevitably you will every so often get the wrong guy and end up submitting an innocent person to tremendous acts of barbarism and cruelty. Torture is wrong regardless of the guilt or innocence of the person being tortured, but its reality is made all the more horrifying when considered in the context of the authorities having got the wrong guy. Instead of seeing this statistical certainty as a reason to prohibit torture in all circumstances (better to let a guilty man go unpunished than punish an innocent), Harris runs head-on into it and admits that he's OK with torturing someone who 'happens to look like' the suspect you're after as long as your intention is to obtain intelligence for the war on terror.

There really is no other way to put it than to simply say that this is a sentence unworthy of an individual claiming to have something interesting to say about morality. It's regressive and barbaric, despite the scholarly detachment with which it's stated and poorly-thought out, morally irrational drivel.

Concluding remarks

I hope that I've addressed most of the points Harris brings up when discussion torture. Part of what makes this such a frustrating exercise is that Harris often seems to portray himself as a long dissident voice in moral reasoning speaking truths to absent-minded intellectuals who haven't fully considered that torture could be moral. My advice to Harris or anyone else with a sympathetic ear for his views would be to treat academic philosophy with a little bit less contempt (he calls it "boring" in The Moral Landscape) and in particular to read some legal philosophy and court judgments, and track the development of the prohibition against torture in international law. Then, maybe, he'll realise that its prohibition is not simply relegated to absent-minded squeamishness on the part of a few dewey-eyed liberals. Unfortunately, much of Harris' work seems to suggest that he seriously thinks this to be the case.

Update 1: On Nuclear Terrorism

I've recently become aware of another one of Harris' articles addressing torture. He says he'd rather not talk about it, but since he's the one who brought it up in the first place, it's simply reaping what you sow if you take flack for saying irresponsible things in public. The thing that I found most interesting, though, is that Harris tries to limit his approval of torture to the following circumstance:
We will not torture anyone under any circumstances unless we are certain, beyond all reasonable doubt, that the person in our custody has operational knowledge of an imminent act of nuclear terrorism.
This is actually the only time that Harris tries to limit his defense of torture to only the single most extreme scenario of nuclear terrorism. As noted above, his other writings reveal him to be much more cavalier about the subject, showing him justifying its use even if there's a low chance of its yielding useful information, and willing to take the risk of torturing someone who "looks like" the suspect providing the intention is good.

The above example is a demonstration of a pretty ugly side of Harris' work, revealing an instance of what is a popular term in post-modern academic circles - that of "othering". I'm not a huge fan of the term, but if we're going to use it - this seems like a good example. Harris' arguments about torturing terrorists is always couched in ways of "us" torturing "them", which from the perspective of moral and ethical debate is far more prone to hypocrisy and bias than if we imagine the same things done by "them" to "us".

Now, it turns out that we don't need to use the looming threat of nuclear terrorism as a thought-experiment boogeyman in order to debunk it because, as most readers will know, nuclear bombs have actually been used twice before in wartime by the US against the Japanese. It would be interesting to hear Harris answer the following question:

Had Japan learned of America's intentions to drop a nuclear bomb beforehand, would it therefore be ethically necessary for the Japanese to have captured and tortured high-ranking US officials in order to obtain intelligence necessary to prevent the nuclear bombs being dropped on Hiroshima and Nagasaki, thereby saving hundreds of thousands of civilian's lives?

If Harris would answer "yes" to this question then we can take the discussion further from there, but I'm sure readers will see that this sort of reasoning opens up a can of worms that I doubt Harris and other torture advocates intended.

Put simply - if all your concern in the nuclear terrorism scenario is for the protection of innocents, and your justification for torture proceeds on that basis, you should also acknowledge that those suffering under US foreign policy could somehow be justified in capturing and torturing US officials in order to prevent the destruction of their cities.


  1. Thanks for this article. I linked it to a discussion on Sam Harris, whom I detest, on Google plus.