In what is becoming a nearly routine event, we learned today of yet another mass killing in the US: A young man bearing multiple firearms entered the campus of a community college in the Pacific Northwest and killed at least 9 people – apparently strangers. The regularity of these tragic events in the US has put the nation into a conflicted fury, with blame being laid in equal proportions on guns, mental illness, and extremist beliefs.
Unfortunately, scientists – including psychologists like myself – have surprisingly little to say about why these events happen. This is in large part because despite seeming to be omnipresent, heinous acts of evil are in fact quite rare overall (leaving aside garden variety human brutishness), and are either stable in their frequency or on the decline. Their rarity means that most measurable factors that we might use to predict major violent acts like a mass shooting are hopelessly coarse. To cite one example, a recent Op-Ed in the LA Times noted that only 12% of individuals who committed mass shootings between 2005 and 2009 had a previous psychiatric diagnosis. Meanwhile, roughly 40 million Americans suffer from anxiety disorders, and nearly 10% of all americans have suffered from some form of mood disturbance in the past 12 months.
Despite this difficult problem of causal attribution, Sam Harris and others have recently made considerable noise by trying to account for one class of violent acts – those associated with religious extremism. And now, as the most recent killing has apparently been perpetrated by an atheist – against those who had religious faith – Harris has fallen under a massive social media attack.
Groups like ISIS have made a very salient link between their Islamic convictions and what can only be described as hideous acts of violence against innocent victims. I’ll spare you the links that provide evidence for this, as it’s not my main objective here to convince you of the evils of ISIS, or the link between their acts and religious belief. Suffice to say that Sam Harris and others have made the argument that certain types of violent acts – if not all – are best predicted by beliefs. Namely the beliefs that reside in the heads of violent perpetrators.
According to a recent blog post by Harris in which he responds to controversy over his position:
“Many peoples have been conquered by foreign powers or otherwise mistreated and show no propensity for the type of violence that is commonplace among Muslims. Where are the Tibetan Buddhist suicide bombers? The Tibetans have suffered an occupation every bit as oppressive as any ever imposed on a Muslim country. At least one million Tibetans have died as a result, and their culture has been systematically eradicated. Even their language has been taken from them. Recently, they have begun to practice self-immolation in protest. The difference between self-immolation and blowing oneself up in a crowd of children, or at the entrance to a hospital, is impossible to overstate, and reveals a great difference in moral attitude between Vajrayana Buddhism and Islam. This is not to say that Tibetan Buddhist suicide bombers couldn’t exist. Tibetans, generally speaking, are not pacifists—nor are most Buddhists elsewhere. In fact, during WWII, the Japanese Kamikaze pilots were influenced by the doctrine of Zen Buddhism. But there are important differences between Zen and Vajrayana that seem relevant here. Vajrayana emphasizes compassion in a way that Zen does not, and Zen generally maintains a more martial and more paradoxical view of ethics…. My point, of course, is that beliefs matter. And it is not an accident that so many Muslims believe that jihad and martyrdom are the highest callings in human life, while many Tibetans believe that compassion and self-transcendence are. This is what Islam and Vajrayana Buddhism, respectively, teach.”
As Daniel Dennett and many others have argued, appealing to beliefs and desires is surely the best tool we have for predicting human behavior in most contexts, and we all do exactly this every day. In fact, there’s increasing evidence that even human infants default to explaining human actions as the product of beliefs and desires (though I remain skeptical about what such studies mean). Dennett calls this mode of behavioral explanation the “Intentional Stance”. Psychologists – and especially developmental psychologists like me – refer to this as “Theory of Mind” or “ToM”. ToM, though present in bits and pieces in infancy, is slow to mature, impaired in children with Autism Spectrum Disorder, and is either absent or greatly restricted in other species, perhaps because it depends in part on the ability to learn and use natural language.
Roughly, explaining behaviors via the intentional stance involves something like the following steps:
(1) Decide to treat an entity as a rational agent.
(2) Decide which beliefs and desires the agent should have given their past experiences and current situation.
(3) Predict the entity’s behaviors by assuming that it will attempt to satisfy its desires given its beliefs.
Pulling things apart in this way is important because it allows us to see ways in which explaining behaviors by appeal to beliefs can break down.
First, the intentional stance may not apply if we judge that an individual is not rational. For example, we might judge that an individual cannot fulfill his or her desires given their knowledge state because their knowledge state does not reflect reality, due to some form of psycho-pathology. In other words, the individual may be delusional or otherwise incapable of veridically perceiving reality. Not a rational agent, in short.
Second, perpetrators can act on bad information – and thus on the basis of false beliefs. The otherwise evil-seeming actions of an agent might be discounted as less than evil if it turns out that they have been misled. For example, an act that seems random and evil – like beating a stranger to submission – may seem less evil, though perhaps still brutish, if we learn that the perpetrator has been misled to believe that their victim has assaulted a loved one.
Finally, the intentional stance assumes that beliefs and desires are more or less determinate: That if we have one belief, “I like cake” then we probably don’t also have the opposite belief, “I don’t like cake”. This last assumption also poses interpretative problems to the intentional stance, since often our beliefs don’t seem all that determinate after all. Anyone who has ever wanted cake, but also wanted to maintain a constant waste line, might understand what it means to have more than one attitude towards a proposition. It’s not that weird to want to have one’s cake and eat it too.
So much for belief-desire psychology. Humans interpret the actions of others according to beliefs and desires all the time. And to not do so would be plainly bizarre. At least in Dennett’s framework, we either explain the behaviors of things in our environment by appeal to mental states, or we are left with lower level mechanisms of explanation like the one we use to predict toasters and alarm clocks (Dennett’s design stance), or the one we use to predict ping pong balls and falling apples (the physical stance). These latter two modes of explanation are of course not mutually exclusive with the first (beliefs and desires are surely restricted by the physical systems that they are implemented in). But the way that we routinely explain human action is via intentional ascription, rather than these other two much more indirect channels.
Returning to Harris, does it make sense to conclude that unusual acts of cruelty and violence are sometimes the result of religious beliefs? Many critics of Harris, like Reza Aslan and Ben Affleck, find it offensive to even consider this proposition that religious beliefs play a causal role in acts of violence. This seems like a difficult position to hold, unless we’re willing to give up beliefs and desires altogether. How could religious beliefs not impact behavior – whereas all other beliefs – like the belief that apples are stored in the fridge – do? If we are willing to attribute acts of charity and benevolence to religious belief, we also need to accept that acts of evil could be causally related to such beliefs too. Either beliefs are causally related to behaviors or they aren’t. It’s frankly difficult to imagine a coherent public discourse that didn’t assume a role for beliefs in causing actions. And I honestly don’t know of any other sphere of debate in which this idea is seriously considered.
On the other hand, how useful is it, really, to go from the plainly obvious assertion that beliefs and desires are causally related to actions to the conclusion that we can thereby explain mass shootings, acts of terror, and so on? Harris is on firm ground when he makes the plainly obvious point that beliefs are causally efficacious, but mistakes this single important observation for a theory of human behavior. Here it is instructive to again consider Dennett’s intentional stance. First, notice that the intentional stance does not itself explain why some people but not others come to acquire the beliefs that they do. Second, the theory does not explain why individuals differ with respect to the depth of their convictions. Third, the intentional stance does not explain why two people with the same conflicted desires (e.g., regarding cake) ultimately behave differently – one staying lean and the other putting on pounds – even though both would prefer to have the cake rather than eat it. In order to provide a complete explanation of behavior – such that we can predict when humans will act one way rather than another – we need to know all three pieces: What causes the adoption of beliefs and desires, which actions are consistent with particular beliefs and desires, and finally, what motivates particular humans to actually act on them.
And here, as research on the link between mental illness and mass shootings demonstrates, is where things fall apart for Harris. In broad strokes, beliefs matter, but they are only one small piece of the puzzle.
Let’s make things clear with a familiar example. Personally, I believe the conclusions reached by climate science, and so do most people that I interact with daily – scientists mainly. So, I drive an electric car, avoid almonds, and am astonished by people who don’t believe, and, well… that’s about it. My actions in response to climate change are quite limited. Meanwhile, an increasingly bleak future awaits me and my family if we remain in Southern California long term. It is not the case that climate science deniers are clustered along the coast lines – in fact coastal inhabitants are substantially more likely to endorse climate science than those who live in the middle of the country. And yet the coasts – which soon will be under water, ravaged by super storms, and parched by drought – are densely populated, incredibly expensive, and getting more of each by the day. These coastal dwellers, like me, believe-ish. And together, we scold doubters (formerly skeptics) for not believing-ish. But most of us don’t really double-dog-dare-you believe. Or else we’d be selling our electric cars, riding bicycles, going vegan, and buying real estate in the pacific north west. Like James Cameron.
In many domains, like climate science and religion, human belief is not terribly predictive of behavior. Some go to church on Sundays and commit adultery on Mondays. Or believe in the climate end times on Sundays and fly to scientific conferences in carbon coughing jet planes on Mondays, while snacking on complimentary nuts (ok, well not anymore). But then there are those who really believe. In the case of climate science, believers build bunkers, hoard food and water, learn to operate fire arms, and assume the worst very generally. In the case of religion, they look forward to the afterlife, feel elevated by martyring themselves and innocents around them, and feel no compunction about doing God’s work even if it involves bloodshed, destruction of ancient artifacts, and inspiring fear.
Harris – together with fellow atheist Bill Maher – has noted that a sizable proportion of ordinary Muslims endorse beliefs which, if acted upon, would result in what many consider to be atrocious acts. And yet only a vanishingly small minority of Muslims actually carry out such acts. The problem is that, in this instance, Harris treats the explanation of behavior as a single factor enterprise, while giving very little thought to what causes beliefs to arise, and why some people act on beliefs while others do not. Many factors need to be understood to predict how a particular person will behave, including beliefs and desires, but also including factors like poverty, the availability of weapons, mental illness, and environmental triggers for action like the cruelty and violence of other people.
This type of reduction of moral questions to single-factor forced-choices consistently leads Harris into controversy: Is it religion or poverty that causes terrorism? What’s worse: rape or religion? Harris’ questions – and ultimately the positions he takes – are about as sensible as the game “Would you rather”, in which 6-year-olds entertain absurd thought experiments such as, “Would you rather be born with an elephant trunk or a giraffe neck?” or “Would you rather eat poison ivy or a handful of bumblebees?”(or for popular and equally absurd online versions for adults, see here or here). It’s pretty obvious that not every question that can be asked, however interesting to contemplate, turns out to be all that useful to answer. This seems especially true as we try to make sense of yet another mass killing. No single factor will ultimately tell us what led a young man to kill a group of innocent strangers. Was he unhappy? Disenfranchised? Was it the availability of weapons? Or was it because he hated religion? Quite likely each factor played some role, with none alone either necessary or sufficient. Pinning all of the focus on a beliefs alone – in this particular case, or in any other mass shooting – can only tell us part of the story, and ultimately risks overlooking a true understanding of how we can prevent acts of violence in future.