Why Sam Harris’s argument that it’s ethically OK to kill people for what they believe still doesn’t stack up.
It’s not uncommon for a polemicist’s words to come back to haunt them, as Sam Harris knows all too well. For the past decade, he’s has been trying to exorcise the ghost of a notorious line in his 2004 book The End of Faith: “Some propositions are so dangerous that it may even be ethical to kill people for believing them.”
Recently, someone going by the handle @dan_verg_ Tweeted a picture of Harris with these words plastered across it, along with the judgment that Harris is a “genocidal fascist maniac” (an assessment I’m not going to defend). This was re-Tweeted to hundreds of thousands of people, leading Harris to write another defence of his Infamous Argument, entitled ‘On the mechanics of defamation’. It’s time to revisit the argument.
What did he say?
In The End of Faith, Harris wrote:
The power that belief has over our emotional lives appears to be total. For every emotion that you are capable of feeling, there is surely a belief that could invoke it in a matter of moments. Consider the following proposition:
Your daughter is being slowly tortured in an English jail.
What is it that stands between you and the absolute panic that such a proposition would loose in the mind and body of a person who believed it? Perhaps you do not have a daughter, or you know her to be safely at home, or you believe that English jailors are renowned for their congeniality. Whatever the reason, the door to belief has not yet swung upon its hinges.
The link between belief and behavior raises the stakes considerably. Some propositions are so dangerous that it may even be ethical to kill people for believing them. This may seem an extraordinary claim, but it merely enunciates an ordinary fact about the world in which we live. Certain beliefs place their adherents beyond the reach of every peaceful means of persuasion, while inspiring them to commit acts of extraordinary violence against others. There is, in fact, no talking to some people. If they cannot be captured, and they often cannot, otherwise tolerant people may be justified in killing them in self-defense.
The ‘tortured daughter’ scenario makes the uncontroversial point that acquiring new beliefs can have a profound effect on our emotional state. (Note that Harris makes clear that he is talking about “beliefs that are communicated, and acquired, linguistically”.) It’s not clear what point it serves in Harris’s broader argument; presumably the goal was to get the reader thinking, “If I found out that my daughter was being tortured, I’d spring into action instantly” — thus establishing a tight connection between belief and behaviour.
However, in Harris’s example, any behaviour that followed, such as trying to rescue your daughter and perhaps killing the torturer if necessary, would be driven by your love for your daughter, and your motivation to protect her — neither of which are linguistically acquired beliefs.
Harris then delivers the two lines that have generated such opprobrium: “The link between belief and behavior raises the stakes considerably. Some propositions are so dangerous that it may even be ethical to kill people for believing them.” Many people, myself included, have read this to mean that if someone accepts certain dangerous propositions or beliefs, then that could provide sufficient grounds for ethically killing them. Harris says this is an outright misreading. How so?
Harris argues that his critics believe he’s overlooking the importance of behaviour, which is the direct and immediate cause of harm to others – and the prevention of which may warrant killing (a conclusion that Harris notes many people already agree with, for example in self-defence or the prevention of genocide).
Harris replies emphatically that he’s not ignoring the belief-behaviour connection; in fact, he’s concerned with beliefs precisely because they’re tightly connected with behaviour. As he clarified after The End of Faith was published, “It should be clear that I am not at all ignoring the link between belief and behavior. The fact that belief determines behavior is what makes certain beliefs so dangerous.”
This reply suggests Harris is oblivious to the possibility that the problem is not that he ignores the link between belief and behaviour, but mischaracterises the link. He claims that “belief determines behaviour”, and ‘determine’, according to the Oxford Dictionary on my shelf, means “to fix or decide causally, condition as a cause or antecedent, be a deciding or the decisive factor in … establish the nature of”. This dictionary definition may not be what philosophers mean by determine, but it does accord with everyday usage of the word. To say that beliefs determine behaviour is much stronger than saying they influence, guide or affect behaviour.
So it’s natural to read Harris as saying if, say, someone holds the belief that violent jihad and the killing of apostates is a good thing, this will determine their behaviour: they will act on this belief, going on to kill people, and that’s what makes the belief dangerous.
Although Harris never states that there’s an ironclad link between belief and behaviour, this view is implied by the whole tenor of his discussion, and is the only way to make sense of talking about killing people for their beliefs as opposed to their overt behaviour. To repeat, the point is not that Harris is ignoring behaviour, but that he thinks you can talk about beliefs as a proxy for behaviour because the contents of someone’s beliefs determine the character of their behaviour.
The Achilles’ heel of this line of thinking is that our behaviour is clearly not driven solely by beliefs. What we do is the product of a complex interplay not only of beliefs, but also of values, attitudes, desires, and motivations — and these are not just another species of linguistically acquired belief. Take a soldier who throws himself on a hand grenade to save his brothers in arms. Is his behaviour just the product of belief? Or is such heroism better explained by an emotional commitment that transcends belief in Harris’s sense?
Or take another example. In his fine book War, Sebastian Junger recounts the following story:
During the air war of 1944, a four-man combat crew on a B-17 bomber took a vow to never abandon one another no matter how desperate the situation. (A fifth team member, the top turret gunner, was not part of the pact.) The aircraft was hit by flak during a mission and went into a terminal dive, and the pilot ordered everyone to bail out. The top turret gunner obeyed the order, but the ball turret gunner discovered that a piece of flak had jammed his turret and he could not get out. The other three men in his pact could have bailed out with parachutes, but they stayed with him until the plane hit the ground and exploded. They all died.
I won’t prejudge what you’ll make of this. Perhaps you think it was stupid, or even selfish, for the three men who could have escaped to not do so (after all, they may have had family and other loved ones waiting for their safe return at home). Or maybe this story fills you with awe at the commitment and loyalty these men showed to each other.
Regardless of your take on this story, you cannot understand this kind of sacrifice — or acts of heroism or courage more broadly — simply by thinking about the beliefs stored in the heads of the men in the plane. Sure, they may have had different beliefs than you or I, specifically those related to the importance of group loyalty. But do you really think that the mere intellectual acceptance of this belief provides the motivational force to sit in a plane that’s plummeting to the ground, with your death a near certainty?
Or consider a less dramatic example. Imagine you’re at a party and some guy, without provocation, comes over to you and starts insulting you, pushing you around and squaring up for a fight. You try to defuse the situation, but he’s coming at you, chest pumped up, fists clenched. It looks like this is getting physical whether you like it or not.
In this kind of situation I believe — and I’m pretty sure Harris would agree — it’s ethically acceptable to deliver the first strike in self-defense: if you wait until he’s thrown a haymaker before you defend yourself, it may be too late.
Yet it’s one thing to have that belief, and quite another to actually summon the moxie to throw the first punch in a real conflict situation. You need to drill this reaction into your muscle memory by practising it over and over in realistic scenarios. Even if you’re physically capable of throwing a good punch because you’ve been training hard on a punch bag, belief in the rightness of your action is not enough to help you act as needed when push comes to shove. You need the right attitude, not just the right beliefs. Or, to put it another way, how you behave is this situation is not simply a product of your beliefs.
What’s more, holding certain beliefs in no way guarantees you’ll act on them. You probably believe it would be good to give more to charity, exercise more, and eat better, but that doesn’t mean you’ll actually do these things. (When, on the basis of your beliefs, you decide to forego donuts to make you healthier and happier, but then succumb to temptation and scoff one down, is it some anti-dieting belief that drives your behaviour?Or is it better seen as the product of a dopamine-fuelled reward system to just makes your crave the donut?) Beliefs can clearly motivate action, but they often, if not usually, need some non-belief-based input to really get the behavioural engine running.
Psychopaths dramatically illustrate this point. Psychopaths are not, by and large, delusional or disordered in their thought processes – they can be perfectly rational while being cold, cruel or outright evil. Psychopaths can reason through moral dilemmas and make judgements about what the right and wrong thing to do is much like everyone else – it’s just that they don’t care about doing what’s right. (In fact, when pushed to make moral judgements psychopaths tend to be adopt a more utilitarian framework of the sort Harris advocates elsewhere.) Psychopaths, in other words, can hold the same moral beliefs as non-psychopaths, but simply ignore them in practice. If you’re ever attacked by a psychopath, you won’t be defending yourself against his or her beliefs.
Belief and behaviour once more
In the latest defence his argument, Harris asks:
Why would it be ethical to drop a bomb on the leaders of ISIS at this moment? Because of all the harm they’ve caused? No. Killing them will do nothing to alleviate that harm. It would be ethical to kill these men—once again, only if we couldn’t capture them—because of all the death and suffering they intend to cause in the future. Why do they intend this? Because of what they believe about infidels, apostates, women, paradise, prophecy, America, and so forth.
Although Harris doesn’t spell it out, the desired conclusion is clear: we’re really killing them for their beliefs (or the propositions they hold in their heads), and this is all that was meant by the original claim “Some propositions are so dangerous that it may even be ethical to kill people for believing them”.
Really? Yes, beliefs are part of the causal story of why ISIS does what it does. But as I’ve argued, these beliefs are not the full causal story. Does one really become willing to die fighting for a group and a cause just through linguistically acquired beliefs? Is the only difference between someone who is willing to behead people, and someone who would not or could not, that they hold different beliefs? (By analogy, many people believe it’s ethically acceptable to kill chickens for food, but lots of these people would balk at the prospect of personally chopping a chicken’s head off.)
Even if beliefs were the sole or key determinant of what ISIS does, then it would still be odd to say we were killing them for their beliefs rather than their actions. We can acknowledge that their beliefs are part of the causal chain leading to their behaviour, but these beliefs are themselves the just a step on the causal path to behaviour. Beliefs, like other mental states, are causally dependent on brain activity, so why not go back another step in the causal chain, look at the brain basis of the beliefs they hold, and say “We’re killing ISIS because of their neurochemistry”?
Because nobody talks like this, and for good reason. When a group of hooded teens tried to mug me and a friend as we walked home with pizzas in London, I wasn’t trying to escape the causal factors that led to their behaviour, such as their beliefs, neurochemistry, hunger or social deprivation. I was trying to get away from an imminent threat, keep my belongings and not get beaten up or stabbed. (We were mostly successful, and after jumping us they only managed to get the pizzas without us suffering any serious harm).
Despite Harris’s nod to the importance of behaviour as the reason why we fight groups like ISIS in his most recent version of the ‘Infamous Argument’, he still focuses on beliefs as if action followed from them ineluctably. And this reinforces the interpretation of his original argument as saying that beliefs per se may be sufficient grounds for killing someone.
One way Harris could defuse a lot of the fear around his argument is to assent to the following: “It is ethically unacceptable to kill someone solely on the basis of the beliefs they hold. The only ethical justification for killing others is in self-defence, or to prevent people from inflicting harm on a third-party, regardless of the role specific beliefs play in creating that threat”. If that’s not a formulation Harris would accept, just how would he phrase it?
In sum, Harris has either made an outrageous ethical claim — if someone holds certain beliefs, that is sufficient warrant to kill them — that he’s now trying to deny, back away from, or revise. Or he made a widely accepted ethical point — that killing in self-defence against a violent aggressor is ethically permissible — in such an obtuse way that many readers came away thinking the opposite of what he intended. Which is it?