Why does the presentation of persuasive evidence — even evidence of a scientific consensus — so often fail to resolve political debates? How is it, for example, that so much of the American public on the right refuses to accept the scientific consensus regarding the causes and risks of climate change?
For a while now, I’ve thought that Dan Kahan’s theory of “cultural cognition” offered the most persuasive answer to these questions. Kahan rejects the idea that the problem lies in Republicans’ lack of information about climate science. Offering more evidence isn’t going to resolve the issue at this point. It might even aggravate the problem.
Rather, Kahan offers empirical evidence that the Republican resistance to climate science is an example of a more general phenomenon: the human tendency to arrive at conclusions that are congenial to our cultural values, and to resist, dismiss, or attack conclusions that threaten our values and identities.
But the more I’ve learned about the specifics of the cultural cognition theory, the more I’ve felt like it leaves something out.
In this post, I’d like to propose a hypothesis that complements cultural cognition’s explanation for the frequent failures of evidence-based discussion to lead to increased agreement on politically charged issues. When I first heard about Kahan’s work, I thought that the theory I’m about to present was what he meant by “cultural cognition.” But as I’ve read more about his work, it’s become clear to me that the idea I have in mind is a distinct one.
I’ll call the hypothesis “tribal cognition.”
This is the hypothesis: we have a tendency not only to resist ideas that threaten our cultural values and identities, as Kahan argues, but to resist ideas that come from speakers we perceive as belonging to a tribe that is different from our tribe — especially a tribe that we perceive as being in conflict with our tribe.
To be clear: tribal and cultural cognition aren’t incompatible theories. It’s very possible that both tribal and cultural cognition play a role in thwarting democratic deliberation. But if my hypothesis is correct, then the solutions that Kahan proposes for addressing democratic communication failures may not be sufficient. We may need to take additional steps, or adopt a different approach entirely.
To understand the difference between cultural cognition and tribal cognition, consider the following observations about cultural cognition, from a largely sympathetic article by Chris Mooney in Mother Jones:
[T]hose who wish to communicate to the public about climate change will have to grapple with Kahan’s assertion that conservatives really aren’t ignorant about the issue—they’re just highly prone to defend their worldviews when asked certain kinds of questions. If Kahan is right, the implication is that we need to talk about climate science in a way that is entirely devoid of cultural meanings that will antagonize the right. [Emphasis added.]
Later in the paper, Kahan goes on to assert that precisely this strategy is working right now in Southeast Florida, where members of the Regional Climate Change Compact have brought on board politically diverse constituencies by studiously avoiding pushing anyone’s buttons. Kahan even shows polling data suggesting that questions like “local and state officials should be involved in identifying steps that local communities can take to reduce the risk posed by rising sea levels” do not provoke a polarized response in this region. Rather, liberals and conservatives alike in Southeast Florida agree with such a statement, which references a major consequence of climate change while ignoring the gigantic elephant in the room… its cause.
As Mooney suggests, Kahan et al.’s “Climate Science Communication and the Measurement Problem” (2014) explains the deliberative success of the Compact process as follows:
What makes Southeast Florida so different from the rest of the country? Indeed, why is Southeast Florida that engages climate change inside the Compact decisionmaking process so different from itself as a part of the country that is polarized on whether human activity is causing global warming?
The explanation is that the Compact process puts a different question from the one put in the national climate change debate. [Emphasis added.] The latter forces Southeast Floridians, like everyone else, to express “who they are, whose side they are on.” In contrast, the decisionmaking of the Compact is effectively, and insistently, testing what they know about how to live in a region that faces a serious climate problem.
In other words, cultural cognition theory suggests that if an issue has become politically polarizing, you should try describing it in different terms. Talk about it in a way that doesn’t raise political hackles or prod sensitive cultural values. Reframe the issue in terms of a different question. The problem is one of rhetoric and framing.
The tribal cognition hypothesis, by contrast, suggests that rhetorical changes might not be enough. Looking at the relative success of the Compact process in Southeast Florida, the tribal cognition hypothesis might ask whether the bipartisan agreement would have been possible if the “action items” in the Regional Climate Action Plan had been proposed by a panel of liberal experts — no matter how they were phrased.
Personally, if I know nothing about an issue of public policy, and I hear that Sean Hannity believes X, my default assumption is going to be that X is a bad idea — most likely, a very, very bad idea. I’m sure that a member of the Tea Party feels the same way after hearing that President Obama supports a policy. Hannity can do his best to present X rhetorically as something that liberals should agree with, but most likely I’ll still be deeply skeptical. In fact, his attempt to frame his favored policy as sympathetic to liberals will likely make me even more suspicious. Similarly, Obama’s frequent attempts to frame his proposals as nonpartisan, common sense ideas have almost entirely failed to persuade anyone who views him as the leader of the opposition.
The converse situation exists as well. If a member of my political tribe comes out in support of a policy that I assumed to be a Republican policy, I will be more likely to give the policy a second look — perhaps it is not so unreasonable after all? I might even have this response if the speaker frames the policy as an expression of cultural values that I ordinarily associate with political opponents.
If the tribal cognition hypothesis is correct, then rhetoric is not enough. The perceived identity of the speaker matters.
(The use of the first person in the preceding paragraphs feels a little strange, as though I should be distancing myself from such irrational deliberative processes. And yet — in all honesty — when it comes to evaluating public policy proposals, I find the default rejection of the ideas of the contemporary national Republican party to be a perfectly reasonable rule of thumb. I imagine many Republicans feel the same way, only with the partisan valence reversed — and their approach may be quite reasonable in light of the values they are attempting to uphold, which are probably very different from my own.)
If tribal cognition is a powerful factor in disrupting the formation of political consensus through reason-based deliberation, then it may not be possible to achieve a politically effective consensus simply by, for example, framing the problem of climate change as a national security threat in an attempt to appeal to the cultural values of the right. Or, to take another example: it may not be possible to reach agreement on generous welfare policies simply by framing those policies as a way of promoting Scandinavian levels of hard work. If Nancy Pelosi is doing the framing, the right will still find reasons to reject her conclusions.
The tribal cognition theory would suggest that it may be necessary to find someone from the opposing tribe to speak out in favor of the policies that your tribe supports. For example, the best way (perhaps the only way?) to persuade the American public on the right to accept the scientific consensus regarding climate change may be for public figures that the right identifies as solidly “one of their own” (not Jon Huntsman, not John McCain) to publicly support the scientific consensus. Maybe someone in a uniform with a lot of ribbons and medals on it?
What might it look like to test the existence and influence of tribal cognition? It’s not hard to think of experiments that could be conducted. Find an issue about which most members of the public are mostly oblivious. Collect a group of test subjects and sort them based on their political-tribal identities. Ask them a question about the issue to get their initial attitude. Show them a video of a polarizing public figure expressing an opinion about the issue. Then ask them again about their attitude, and see if it has changed — in particular, see if the members of the polarizing public figure’s tribe have moved toward the figure’s position, and the opponents of that tribe have moved away. Something like that. (Incidentally, I’m pretty sure I know how the test will turn out… But the obviousness of a result has certainly not prevented psychologists from carrying out experiments in the past.)
On a somewhat peripheral note: one illustration of the power of tribal perceptions to influence political views could be the rapid changes in public opinion regarding gay rights and gay marriage over the last thirty years, and especially over the last ten years. Some have argued that the changes resulted in significant part from members of the public learning that people they knew and respected were gay. In other words, as the public came to perceive that gays could be “one of us” — rather than some disconcerting, alien other — being gay no longer seemed so problematic, and opposition to gay rights, including gay marriage, began to dissolve.
Another rather peripheral note: an interesting contrast between tribal cognition and cultural cognition is that the former is a more concrete and determinate phenomenon in some ways, at least in the context of the current American political landscape. Tribal cognition is easier to test than cultural cognition, and it provides an easier basis for developing prescriptions. There’s often something slippery about determining whether a given piece of rhetoric appeals to or offends a given cultural value, or whether a given cultural value should be associated with one or another political pole, or whether a given policy has one or another political valence. (For example: what is more definitive of the “left” than the welfare state? Yet parties on what is usually seen as the far right in Europe have begun to support generous welfare benefits — for the natives.) Depending on what one wants to do with the theory of cultural cognition, it may be necessary to reach determinate conclusions about some of these slippery questions.
Tribal cognition, on the other hand, can be analyzed and used in more straightforward, easier ways. In the current American political landscape, where, as Ezra Klein points out in his article on Kahan, the parties are sorted by ideology to a degree without recent historical precedent, it’s pretty easy to tell what the tribes are, and who belongs to which tribe. Economic and social conservatives are almost entirely Republicans, and Republicans are almost entirely economic and social conservatives — and probably most of their friends, neighbors, and colleagues are as well. And there’s usually very little disagreement in people’s perceptions about who belongs to which tribe. No subtle rhetorical analysis needed.
On a related note, and as a final thought, the same facts that make tribal cognition a particularly tractable model in the United States today also make it particularly troubling. Because the political tribes are so well-sorted in terms of ideology and cultural attitudes, there will be few cross-over figures who can challenge members of their own tribe to take a second look at the opposing tribe’s beliefs and proposals. In the Senate of the late 1950s, as Robert Caro has shown in vivid detail, Lyndon Johnson was able to get civil rights legislation passed in part by presenting himself as a member of the Southern Democrats’ segregationist tribe. Johnson was a cultural Southerner, and a Democrat, but he also turned out to be in some ways a social liberal. He was a mix, and the mix allowed him to scramble his conservative Senate colleagues’ tribal-cognitive radars. The white supremacists thought Johnson was one of their own, and this allowed him to defuse white supremacist opposition to the Civil Rights Act of 1957.
Today, in the wake of the political sorting that (as Johnson predicted) followed the Democratic Party’s embrace of racial equality in the 1960s, there are very few prominent voices capable of scrambling our tribal-cognitive radars in a similar way. Take Jim Webb, the former Democratic Senator from Virginia, respected literary novelist, Secretary of the Navy (under Reagan), Navy Cross recipient, Vietnam veteran, and first-in-class graduate of the Marine Corps Officer Basic School. Some observers seem to feel that Jim Webb should appeal to Republican men based on his military background and cultural Southernness. But he doesn’t really, according to polls. Nor, I suspect, did Michael Bloomberg ever have a chance of appealing to Republicans, as centrist pundits sometimes liked to fantasize.
No, the final remnants of the Democratic South have now disappeared. Today, the parties have sorted themselves out in terms of ideology, and the American public has sorted itself out in terms of party. The main exception, for the time being, may be libertarians like Rand Paul — economic conservatives with social liberal preferences on certain issues such as military engagement, domestic surveillance, and drug policy.
The economist and political philosopher Amartya Sen once tried to undermine the power of community identity to lead to conflict and violence by emphasizing that each of us belongs to a variety of cross-cutting communities. Yes, you may be Muslim and I may be Sikh, but perhaps we are both women, or entrepreneurs. Yes, we may both be Christians, but perhaps I’m an egalitarian economic and social leftist while you’re a conservative who believes that economic and social hierarchies are both natural and necessary.
Today, in the United States, at least when it comes to political identity, Sen’s argument doesn’t seem to apply very well. The same people tend to be in each of the various tribes that define one’s political identity; and the people who belong to the tribe of the partisan opposition also tend to be the people who belong to the tribe of the ideological and cultural opposition. There are of course ways that a liberal can find common ground with a conservative, and vice versa — for example, barbecue — but I suspect these sources of common ground will rarely be deeply rooted or significant enough to scramble perceptions of tribal identity for the purposes of encouraging reason-based, non-tribal political dialogue.