Sam Harris and the Myth of Perfectly Rational Thought

This piece was originally published in Wired on May 17, 2018.

Sam Harris, one of the original members of the group dubbed the “New Atheists” (by Wired!) 12 years ago, says he doesn’t like tribalism. During his recent, much-discussed debate with Vox founder Ezra Klein about race and IQ, Harris declared that tribalism “is a problem we must outgrow.”

But apparently Harris doesn’t think he is part of that “we.” After he accused Klein of fomenting a “really indissoluble kind of tribalism” in the form of identity politics, and Klein replied that Harris exhibits his own form of tribalism, Harris said coolly, “I know I’m not thinking tribally in this respect.”

Not only is Harris capable of transcending tribalism—so is his tribe! Reflecting on his debate with Klein, Harris said that his own followers care “massively about following the logic of a conversation” and probe his arguments for signs of weakness, whereas Klein’s followers have more primitive concerns: “Are you making political points that are massaging the outraged parts of our brains? Do you have your hands on our amygdala and are you pushing the right buttons?”

Of the various things that critics of the New Atheists find annoying about them—and here I speak from personal experience—this ranks near the top: the air of rationalist superiority they often exude. Whereas the great mass of humankind remains mired in pernicious forms of illogical thought—chief among them, of course, religion—people like Sam Harris beckon from above: All of us, if we will just transcend our raw emotions and rank superstitions, can be like him, even if precious few of us are now.

We all need role models, and I’m not opposed in principle to Harris’s being mine. But I think his view of himself as someone who can transcend tribalism—and can know for sure that he’s transcending it—may reflect a crude conception of what tribalism is. The psychology of tribalism doesn’t consist just of rage and contempt and comparably conspicuous things. If it did, then many of humankind’s messes—including the mess American politics is in right now—would be easier to clean up.

What makes the psychology of tribalism so stubbornly powerful is that it consists mainly of cognitive biases that easily evade our awareness. Indeed, evading our awareness is something cognitive biases are precision-engineered by natural selection to do. They are designed to convince us that we’re seeing clearly, and thinking rationally, when we’re not. And Harris’s work features plenty of examples of his cognitive biases working as designed, warping his thought without his awareness. He is a case study in the difficulty of transcending tribal psychology, the importance of trying to, and the folly of ever feeling sure we’ve succeeded.

To be clear: I’m not saying Harris’s cognition is any more warped by tribalism than, say, mine or Ezra Klein’s. But somebody’s got to serve as an example of how deluded we all are, and who better than someone who thinks he’s not a good example?

There’s another reason Harris makes a good Exhibit A. This month Bari Weiss, in a now famous (and, on the left, infamous) New York Times piece, celebrated a coalescing group of thinkers dubbed the “Intellectual Dark Web”—people like Harris and Jordan Peterson and Christina Hoff Sommers, people for whom, apparently, the ideal of fearless truth telling trumps tribal allegiance. Andrew Sullivan, writing in support of Weiss and in praise of the IDW, says it consists of “nontribal thinkers.” OK, let’s take a look at one of these thinkers and see how nontribal he is.

Examples of Harris’s tribal psychology date back to the book that put him on the map: The End of Faith. The book exuded his conviction that the reason 9/11 happened—and the reason for terrorism committed by Muslims in general—was simple: the religious beliefs of Muslims. As he has put it: “We are not at war with ‘terrorism.’ We are at war with Islam.”

Believing that the root of terrorism is religion requires ruling out other root causes, so Harris set about doing that. In his book he listed such posited causes as “the Israeli occupation of the West Bank and Gaza…the collusion of Western powers with corrupt dictatorships…the endemic poverty and lack of economic opportunity that now plague the Arab world.”

Then he dismissed them. He wrote that “we can ignore all of these things—or treat them only to place them safely on the shelf—because the world is filled with poor, uneducated, and exploited peoples who do not commit acts of terrorism, indeed who would never commit terrorism of the sort that has become so commonplace among Muslims.”

If you’re tempted to find this argument persuasive, I recommend that you first take a look at a different instance of the same logic. Suppose I said, “We can ignore the claim that smoking causes lung cancer because the world is full of people who smoke and don’t get lung cancer.” You’d spot the fallacy right away: Maybe smoking causes lung cancer under some circumstances but not others; maybe there are multiple causal factors—all necessary, but none sufficient—that, when they coincide, exert decisive causal force.

Or, to put Harris’s fallacy in a form that he would definitely recognize: Religion can’t be a cause of terrorism, because the world is full of religious people who aren’t terrorists.

Harris isn’t stupid. So when he commits a logical error this glaring—and when he rests a good chunk of his world view on the error—it’s hard to escape the conclusion that something has biased his cognition.

As for which cognitive bias to blame: A leading candidate would be “attribution error.” Attribution error leads us to resist attempts to explain the bad behavior of people in the enemy tribe by reference to “situational” factors—poverty, enemy occupation, humiliation, peer group pressure, whatever. We’d rather think our enemies and rivals do bad things because that’s the kind of people they are: bad.

With our friends and allies, attribution error works in the other direction. We try to explain their bad behavior in situational terms, rather than attribute it to “disposition,” to the kind of people they are.

You can see why attribution error is an important ingredient of tribalism. It nourishes our conviction that the other tribe is full of deeply bad, and therefore morally culpable, people, whereas members of our tribe deserve little if any blame for the bad things they do.

This asymmetrical attribution of blame was visible in the defense of Israel that Harris famously mounted during Israel’s 2014 conflict with Gaza, in which some 70 Israelis and 2,300 Palestinians died.

Granted, Harris said, Israeli soldiers may have committed war crimes, but that’s because they have “been brutalized…that is, made brutal by” all the fighting they’ve had to do. And this brutalization “is largely due to the character of their enemies.”

Get the distinction? When Israelis do bad things, it’s because of the circumstances they face—in this case repeated horrific conflict that is caused by the bitter hatred emanating from Palestinians. But when Palestinians do bad things—like bitterly hate Israelis—this isn’t the result of circumstance (the long Israeli occupation of Gaza, say, or the subsequent, impoverishing, economic blockade); rather, it’s a matter of the “character” of the Palestinians.

This is attribution error working as designed. It sustains your conviction that, though your team may do bad things, it’s only the other team that’s actually bad. Your badness is “situational,” theirs is “dispositional.”

After Harris said this, and the predictable blowback ensued, he published an annotated version of his remarks in which he hastened to add that he wasn’t justifying war crimes and hadn’t meant to discount “the degree to which the occupation, along with collateral damage suffered in war, has fueled Palestinian rage.”

That’s progress. “But,” he immediately added, “Palestinian terrorism (and Muslim anti-Semitism) is what has made peaceful coexistence thus far impossible.” In other words: Even when the bad disposition of the enemy tribe is supplemented by situational factors, the buck still stops with the enemy tribe. Even when Harris struggles mightily against his cognitive biases, a more symmetrical allocation of blame remains elusive.

Another cognitive bias—probably the most famous—is confirmation bias, the tendency to embrace, perhaps uncritically, evidence that supports your side of an argument and to either not notice, reject, or forget evidence that undermines it. This bias can assume various forms, and one was exhibited by Harris in his exchange with Ezra Klein over political scientist Charles Murray’s controversial views on race and IQ.

Harris and Klein were discussing the “Flynn effect”—the fact that average IQ scores have tended to grow over the decades. No one knows why, but such factors as nutrition and better education are possibilities, and many of the other possibilities also fall under the heading of “improved living conditions.”

So the Flynn effect would seem to underscore the power of environment. Accordingly, people who see the black-white IQ gap as having no genetic component have cited it as reason to expect that the gap could move toward zero as average black living conditions approach average white living conditions. The gap has indeed narrowed, but people like Murray, who believe a genetic component is likely, have asked why it hasn’t narrowed more.

This is the line Harris pursued in an email exchange with Klein before their debate. He wrote that, in light of the Flynn effect, “the mean IQs of African American children who are second- and third-generation upper middle class should have converged with those of the children of upper-middle-class whites, but (as far as I understand) they haven’t.”

Harris’s expectation of such a convergence may seem reasonable at first, but on reflection you realize that it assumes a lot.

It assumes that when African Americans enter the upper middle class—when their income reaches some specified level—their learning environments are in all relevant respects like the environments of whites at the same income level: Their public schools are as good, their neighborhoods are as safe, their social milieus reward learning just as much, their parents are as well educated, they have no more exposure to performance-impairing drugs like marijuana and no less access to performance-enhancing (for test-taking purposes, at least) drugs like ritalin. And so on.

Klein alluded to this kink in Harris’s argument in an email to Harris: “We know, for instance, that African American families making $100,000 a year tend to live in neighborhoods with the same income demographics as white families making $30,000 a year.”

Harris was here exhibiting a pretty subtle form of confirmation bias. He had seen a fact that seemed to support his side of the argument—the failure of IQ scores of two groups to fully converge—and had embraced it uncritically; he accepted its superficial support of his position without delving deeper and asking any skeptical questions about the support.

I want to emphasize that Klein may here also be under the influence of confirmation bias. He saw a fact that seemed to threaten his views—the failure of IQ scores to fully converge—and didn’t embrace it, but rather viewed it warily, looking for things that might undermine its significance. And when he found such a thing—the study he cited—he embraced that.

And maybe he embraced it uncritically. For all I know it suffers from flaws that he would have looked for and found had it undermined his views. That’s my point: Cognitive biases are so pervasive and subtle that it’s hubristic to ever claim we’ve escaped them entirely.

In addition to exhibiting one side of confirmation bias—uncritically embracing evidence congenial to your world view—Harris recently exhibited a version of the flip side: straining to cast doubt on evidence you find unsettling. He did so in discussing the plight of physicist and popular writer Lawrence Krauss, who was recently suspended by Arizona State University after multiple women accused him of sexual predation.

Krauss is an ally of Harris’s in the sense of being not just an atheist, but a “new” atheist. He considers religion not just confused but pernicious and therefore in urgent need of disrespect and ridicule, which he is good at providing.

After the allegations against Krauss emerged, Harris warned against rushing to judgment. I’m in favor of such warnings, but Harris didn’t stop there. He said the following about the website that had first reported the allegations against Krauss: “Buzzfeed is on the continuum of journalistic integrity and unscrupulousness somewhere toward the unscrupulous side.”

So far as I can tell, this isn’t true in any relevant sense. Yes, Buzzfeed has had the kinds of issues that afflict even the most elite journalistic outlets: a firing over plagiarism, an undue-advertiser-influence incident, a you-didn’t-explicitly-warn-us-that-this-conversation-was-on-the-record complaint. And there was a time when Buzzfeed wasn’t really a journalistic outlet at all, but more of a spawning ground for cheaply viral content—a legacy that lives on as a major part of Buzzfeed’s business model and as a parody site called clickhole.

Still, since 2011, when Buzzfeed got serious about news coverage and hired Ben Smith as editor, the journalistic part of its operation has earned mainstream respect. And its investigative piece about Krauss was as thoroughly sourced as #metoo pieces that have appeared in places like the New York Times and the New Yorker.

But you probably shouldn’t take my word for that. I’ve had my contentious conversations with Krauss, and maybe this tension left me inclined to judge allegations against him too generously. In any event, I suspect that if the Buzzfeed piece were about someone Harris has had tensions with (Ezra Klein, maybe, or me), he might have just read it, found it pretty damning, and left it at that. But it was about Krauss—who is, if Harris will pardon the expression, a member of Harris’s tribe. (I should add that Harris wasn’t wholly dismissive of the allegations against Krauss. He said that, based on conversations he’d had with knowledgeable people, he thought some of the allegations were very likely true, but that didn’t mean that the worst of them—the ones alleging actual sexual assault—were true. In any event, my point is just that if Krauss hadn’t been a member of Harris’s new atheist tribe, I doubt Harris would have  spent so much time trying to undermine the credibility of the media outlet in which the allegations of sexual assault appeared. Indeed, in the course of casting doubt on the article, he said, “This wasn't all about Lawrence. This was yet another attack on atheism and secularism.”)

Most of these examples of tribal thinking are pretty pedestrian—the kinds of biases we all exhibit, usually with less than catastrophic results. Still, it is these and other such pedestrian distortions of thought and perception that drive America’s political polarization today.

For example: How different is what Harris said about Buzzfeed from Donald Trump talking about “fake news CNN”? It’s certainly different in degree. But is it different in kind? I would submit that it’s not.

When a society is healthy, it is saved from all this by robust communication. Individual people still embrace or reject evidence too hastily, still apportion blame tribally, but civil contact with people of different perspectives can keep the resulting distortions within bounds. There is enough constructive cross-tribal communication—and enough agreement on what the credible sources of information are—to preserve some overlap of, and some fruitful interaction between, world views.

Now, of course, we’re in a technological environment that makes it easy for tribes to not talk to each other and seems to incentivize the ridiculing of one another. Maybe there will be long-term fixes for this. Maybe, for example, we’ll judiciously amend our social media algorithms, or promulgate practices that can help tame cognitive biases.

Meanwhile, the closest thing to a cure may be for all of us to try to remember that natural selection has saddled us with these biases—and also to remember that, however hard we try, we’re probably not entirely escaping them. In this view, the biggest threat to America and to the world may be a simple lack of intellectual humility.

Harris, though, seems to think that the biggest threat to the world is religion. I guess these two views could be reconciled if it turned out that only religious people are lacking in intellectual humility. But there’s reason to believe that’s not the case.

Note: After this article was published in Wired, I got complaints about how I had described Harris’s reaction to the Lawrence Krauss incident. In response to this feedback I’ve made two changes. I added the parenthetical about Harris acknowledging the likely truth of some of the allegations against Krauss. And, five paragraphs above that, I changed the phrase “straining to reject” to “straining to cast doubt on.”