When George W. Bush became president in 2001, it marked the first time in 70 years that conservative Republicans controlled all three branches of government. By the time Bush left office, we were all reminded why. The financial crisis and resulting global economic meltdown Bush left us with were eerily reminiscent of the Great Depression, but there was also 9/11, the Iraq War and Katrina—a multifaceted record of spectacular failure so stunning that it should have disqualified conservative Republicans from holding power for at least another seven decades. Yet, the Democrats' political response to the many messes Bush left behind has been so spectacularly inept that they've not only lost both houses of Congress, they've also lost more state legislative seats than any time since before the Great Recession.
There are many ways one might explain this state of affairs—and certainly the rise of Wall Street Democrats and the decline of labor played crucial roles. But beyond any particular issue area, there’s also the matter of differences in how liberals and conservatives think—and how they act and organize as a result.
As I’ve written before, a growing body of literature reveals that liberals and conservatives think differently from one another in ways that can even be traced back, in part, to the level of instinctual response, reflecting conservatives’ heightened sensitivity to threat bias. This work is congruent with an integrated multi-factor account offered by John Jost and three co-authors in the 2003 meta-analysis “Political Conservatism as Motivated Social Cognition.” In their abstract, they explained, “Analyzing political conservatism as motivated social cognition integrates theories of personality (authoritarianism, dogmatism–intolerance of ambiguity), epistemic and existential needs (for closure, regulatory focus, terror management), and ideological rationalization (social dominance, system justification).” Their meta-analysis integrated findings from 88 sample studies in 12 countries, with 22,818 individual subjects—meaning it drew on a substantial body of work by others.
Yet, once publicized, it drew such a hostile response there was even talk of Congress defunding the entire field of research into political attitudes. In response, Jost and one co-author wrote a Washington Post Op-Ed, which defused the crisis. In it, they wrote:
True, we find some support for the traditional "rigidity-of-the-right" hypothesis, but it is also true that liberals could be characterized on the basis of our overall profile as relatively disorganized, indecisive and perhaps overly drawn to ambiguity -- all of which may be liabilities in mass politics and other public and professional domains.
This statement underscores the point that liberal cognitive tendencies can be as problematic in their way as conservative ones are.
The multi-factor distinction Jost and his colleagues analyzed is roughly congruent with a broader distinction, discussed by Chris Mooney in"The Republican Brain" (which I wrote about here), related to two of the “Big Five” personality traits—conservatives score higher on conscientiousness, while liberals score higher on openness to new experience.
As these few examples suggest, there are multiple ways to characterize the differences in how liberals and conservatives think. For instance, Mooney argued that liberals, still fundamentally inspired by the Enlightenment promise of ever-growing knowledge about the world, are fundamentally mistaken about the nature of human reason, which they see as knowledge- and truth-seeking. But modern cognitive science teaches us that our brains are much more fundamentally shaped by the need to make persuasive arguments, which only require the appearance of rational argument.
In "The Battle for God," Karen Armstrong illuminates a slightly different, though related, difference, contrasting the modalities of mythos and logos. As Armstrong explains, logos is concerned with the practical understanding of how things work in the world, while mythos is concerned with ultimate meaning. Either modality can be used by liberals and conservatives alike in their everyday lives. But macro-historically, there’s been a distinct bias—and weird twist on top of it—at least since the dawn of the modern era. That’s when logos began becoming so all-pervasive that it seemed destined to dislodge mythos, and some defenders of mythos (now commonly known as fundamentalists) fought back paradoxically by assuming the framework of logos, and arguing that their mythos was literally true—a move that true traditionalists would have found to be deeply in error, because it devalued the essential purpose of mythos.
The congruence with Mooney’s argument is obvious: There’s a clear kinship between logos and the Enlightenment model of reason on the one hand, and mythos and persuasion on the other. If conservatives under George W. Bush once again proved themselves incompetent in the logos of governing, liberals under Obama proved themselves incompetent in its mythos.
Or so I hypothesized. But I wanted to check things out with perhaps the world’s leading expert on incompetence, psychologist David Dunning, the senior researcher in the team that discovered the Dunning-Kruger effect, which Wikpedepia defines as “a cognitive bias whereby unskilled individuals suffer from illusory superiority, mistakenly rating their ability much higher than is accurate.” Wikipedia added that “This bias is attributed to a metacognitive inability of the unskilled to recognize their ineptitude.” Or, as Dunning explained to Errol Morris, writing an essay series, "The Anosognosic's Dilemma: Something's Wrong but You'll Never Know What It Is," for the New York Times, “If you’re incompetent, you can’t know you’re incompetent … [T]he skills you need to produce a right answer are exactly the skills you need to recognize what a right answer is.” A recent article by Dunning, “We Are All Confident Idiots,” provides both humorous and serious examples showing just how pervasive the problem is.
Like many, I first learned of the Dunning-Kruger effect from that NYT series—and made some observations based on it at the time. There are obvious conclusions one can draw from the Dunning-Kruger effect: perhaps most important, that none of those obvious conclusions will apply to your own shortcomings, even though those are the ones that ought to concern you most. But this is specifically an individual effect, and my observation was about groups—and rather large ones, at that. So in reaching out to talk with Dunning, behind any specifics, I had two questions in mind: Could it apply to groups as well as individuals? And was it possible to do something about it?
In both cases, he answered yes, but some of the specifics surprised me. Which is just what I should have expected—to discover some limits of my own understanding. (Dunning himself has referenced Donald Rumsfeld’s phrase “unknown unknowns” to describe what we’re up against, just by the very nature of being human. But don’t have a cow, man. He’s also referenced Socrates, as well.)
To begin with, I wanted to make sure we were on the same page. An example that came readily to mind was the GOP’s claims to have 46 jobs bills that had passed the House, and were languishing in the Senate. If only Obama and Harry Reid would act on them! The reality, of course, is that these bills would not actually do very much in the way of job creation, as critics have pointed out repeatedly over the past several years. In late October, the New York Times even interviewed some top GOP economists who admitted as much, along with independent analysts who said it would be hard to measure much impact.
In short, the GOP “jobs bills” aren’t seriously intended to create jobs. They’re intended to create talking points about creating jobs — and to counter Democratic talking points (while also doing favors for GOP donors, of course). They reflect both the persuasive nature of human cognition highlighted by Chris Mooney, and the meaning-making function of mythos described by Karen Armstrong. They might not create many jobs, I noted early in my conversation with Dunning—it’s aggregate demand that’s the primary driver in doing that—but they do resonate with the “job creator” mythos, which has been so prominent in conservative circles these past several years, and which makes perfect sense in the world of small businessmen I’ve known.
Dunning thought it was an apt example. He noted that people are often perplexed over where a never-ending, chicken-and-egg cycle begins. “You have business people, they don't just decide there is going to be a market, they respond to the market, they respond to a demand,” Dunning said. “But they start the process where they enter the picture … People tend to think of themselves sort of as creators who come in and are imposing their will and their desires on the environment, and sort of filter out the conditions that they are really reacting to. They can recognize it pretty accurately for everybody else, they just miss that for themselves. Which I think is interesting.”
Understanding an example of how conservatives’ thinking leads them astray is the easy part, however. It helped to get our thinking in sync. But the real challenge would be making sense of how liberals and Democrats make comparable kinds of errors—errors they cannot see. And here is where things had to get a bit tricky, since I had some ideas of what the errors might be, but given the Dunning-Kruger effect, I had to expect some ideas I’d never thought of, too.
The next thing out of Dunning’s mouth wasn’t quite that—but it did have some of that flavor. Above I mentioned a paper by John Jost that represents the integration of work done by hundreds, if not thousands of researchers over a period of several decades. That integration very much represents the broad consensus view of how liberal and conservative thought relate to one another—a consensus that has since been significantly strengthened with the addition of a related finding in the physiological dimension. But there are at least two notable voices who stand out with somewhat contrasting views—Jonathan Haidt and Dan Kahan—and Dunning quickly mentioned both of their work as being harmonious with what I was saying.
I could have gotten down into the details of their theories, but in a big picture sense, Dunning was perfectly right: They’re all saying something similar—that liberals and conservatives do think differently from one another, which means that they can both fall prey to the Dunning-Kruger effect in ways that may be different—in terms of cognitive processes—as well as similar—produced by similar sorts of situations, for example. What’s more, if one is concerned with trying to identify ways in which people are blind to their own shortcomings, then it’s helpful to have as many different accounts as possible.
Kahan’s work is particularly challenging, because he prefers to use a two-factor model of orientation—hierarchical/egalitarian and individualist/communitarian—rather than the single-factor measures used by most investigators, which can then be correlated with liberalism/conservatism. But he still does encounter and deal with real-world political questions where things tend to divide into two. One striking example is the “white male effect,” so-called because white males have been observed to be significantly less fearful about certain sorts of risks—most notably ones associated with guns and the environment. In a 2007 paper, for example, Kahan and his colleagues wrote, “The insensitivity to risk reflected in the white male effect can thus be seen as a defensive response to a form of cultural identity threat that afflicts hierarchical and individualistic white males.” Even though two factors were involved, the result was an effect most notable for setting one sub-population apart from the rest of the public—a cognitive bifurcation of sorts.
“Essentially you have each side interpreting what's going on through different lenses,” Dunning said of Kahan’s work, and what connects this to his work is “the idea that people either have no idea that the other side has a different lens—they literally don't know it, so that if you are looking at things the logos way, you really don't know that there's an alternative way of viewing the world is the mythos way. Or, if you are exposed to it, you think it's not real.” Expanding on the later alternative, Dunning said, “I can guess if you described the logos world to a person who was into mythos, and vice a versa, their response would be, well, that's a very interesting way of viewing the world, I wonder what took them off the right track.”
Although this made sense, my working thesis was not that most people don't have both mythos and logos in their experience—that may or may not be true—but that the political worlds of liberalism and conservatism are organized differently, so that only certain aspects of the personal resonate strongly with the political.
“Well, that makes sense,” Dunning told me, then quickly added, “I'm from the other world, though, where I find it very easy to focus on the individual level.” Which led directly to the next question on my mind: Is there specific evidence for the Dunning-Kruger effect overlapping from the individual level to larger social groups?
“The answer is, yes it can,” Dunning told me. He’s still working on writing up experimental results, but he was happy to share some broader observations. “It's likely to happen in two different ways,” he said. “The first is often you have organizations that are well set--you can say they're very competent--in their ways. They can have a problem when conditions on the ground shift, and you could say that in the last 10 or 20 years, the conditions on the grounds shifted demographically, in terms of the people, for example who vote in the midterm.” Twenty years ago, these older voters had come of age during the Great Depression, and leaned Democratic as a result. But older voters now are much more Republican, which tends to skew the midterms in the opposite direction from the past. (The age-based difference in participation rates has also grown over time, as well.) As a result, Dunning said, “You have Republicans in midterms, and the occasional voters now are the young, and they’re rather Democratic so there is oscillation between elections and that's a changed situation,” which is precisely the first sort of thing that tends to trip up organizations.
“The other thing is that the most recent work we've done suggests that the real cost of the Dunning-Kruger framework at the organizational level is that when a very smart idea or very smart person comes along, the organizations are not necessarily very skilled at recognizing that person's genius,” Dunning said. “We have lots of data showing that very top performers, their top performances are very much missed. The genius of their ideas are just missed by the group.”
Three examples on the Democratic side came readily to mind when Dunning said this. First was civil rights lawyer and Harvard law professor Lani Guinier, whose work on voting rights and representational fairness was easily demonized by the right when she was nominated to head the DOJ’s Civil Rights Division, in part because Joe Biden (then chair of the Senate Judiciary Committee) couldn’t understand it, and Bill Clinton didn’t even take the time to try. Yet, in retrospect, if they had stood behind her, most of the GOP’s voting rights mischief in the past 20 years—including the theft of the 2000 election—could have been avoided, or defeated, and there would be a much stronger foundation for multi-racial coalition politics at all levels of government.
Second was Princeton political scientist Jacob Hacker, who has done pioneering work on asymmetric polarization and the devastation of the middle class, as well as developing the public option as a complement to Medicare in providing insurance for those left uninsured as employer-provided health insurance becomes increasingly expensive and less common. In all three areas, Hacker’s thinking has involved considerations that are mostly not even understood by those who’ve ignored, dismissed, or, more rarely, criticized them.
Third was University of California, Berkeley, cognitive linguist George Lakoff, whose work illuminating the cognitive and communicative differences between liberals and conservatives—"Moral Politics," "Don’t Think of An Elephant," "Whose Freedom," etc.—has found a wide audience centered in the progressive activist base, but has yet to seriously impact the political professionals whose collective failure I alluded to in this story’s first paragraph.
When I interviewed him recently for Salon, Lakoff even highlighted the concept of hypocognition—that “we don’t have all the ideas we need.” One example he cited was the concept of reflexivity, “the fact that thought is part of the world. That when you’re thinking, it’s not separate from reality, it’s part of reality. And if your understanding of the world is reflected in what you do, then that thought comes into the world through your actions,” which helps to explain, in part, the power of conservative mythos, even when it’s mistaken as a matter of fact, a matter of logos.
Lakoff also pointed out that “Hypocognition itself is an idea that we need.” There are things going on in our social and political world that we don’t have names for—and because we don’t have names for them, we can’t think and talk about them coherently. So, we have conservatives on the one hand acting on their mythos, mistakenly believing it’s true as a matter of logos—which is one kind of incompetence—and yet, nonetheless reshaping reality through the power of reflexivity. (Think of how invading Iraq in response to 9/11 helped bring ISIS into existence, for example.) On the other hand, we have liberals seeing things only in terms of logos, who can’t understand how wildly mistaken conservatives can nonetheless reshape the world to reflect their paranoid fantasies, because they’re missing the crucial concept of reflexivity (and even the very concept of missing concepts, the concept of hypocognition)—which is another, very different, but very real form of incompetence.
So, when Dunning told me, “The genius of their ideas are just missed by the group,” Lakoff’s discussion of hypocognition naturally came to mind. What could be a worse idea to miss than the very idea of missing ideas? If you don’t think they’re out there, you’ll never go looking for them—never believe anyone who claims to have found one of them, either.
In fact, Democrats appear to face a situation in which both the phenomena Dunning pointed to are happening at once. The demographic shift in midterm voters has happened in the same time frame that Guinier, Hacker and Lakoff have been writing, arguing for new ways of doing things, and more often than not seeing the genius of their ideas being “just missed by the group.”
Dunning also suggested that collective cognitive differences could manifest in group blind spots. “If you're in a group that tends things in a logos box or a mythos box you may very well not know that the other box exists,” Dunning said. “It just doesn't occur to you even though, as you mentioned before, in everyday life you are probably switching from a mythos to logos person.”
“I think the best examples come from cross-cultural understanding,” he quickly added. “That is, what we don't know happens in other cultures—you know, their ideas, the meanings of certain actions, certain concerns that just don't occur to us because we grew up in the United States, as supposed to Japan or China or Africa, for example. So if you think about the mythos/logos distinction as cultural distinction, it wouldn't be surprising that there are ‘unknown unknowns.’”
Dunning then went on to cite “some empirical work being done by Rob Willer and Matthew Feinberg … showing that you can get ‘the other side’ to support your side more if you make sure to approach or political arguments into language or have it address the concerns of the other side. So, for example, if you talk about environmentalism as maintaining the purity of the earth, and get conservatives much much more excited about the idea of sustainability and environmentalism.” [Article press release here.] On the other hand, they also showed that emphasizing the military’s role in providing equal opportunities for minorities impacts liberals to make them more supportive of the military—so adopting different basic frameworks can reach people on both sides of the ideological divide. Their research doesn’t show that differences are erased, but they can be diminished, which is a start.
I started this article by taking note of the colossal failures of the Bush administration, in part because they’re so staggering that they’re impossible to miss. But Dunning cautioned against being misdirected. “The real effect of suffering from Dunning-Kruger is not that you suffer obvious losses but that there are so many opportunities you will never notice, or know about in your life,” Dunning said, “and that's absolutely true that the collective level ... People at collective level are going to make mistakes, even if they’re expertly interested in doing the best they possibly can.”
While Dunning already touched on one reason for this—a collective failure to appreciate new, breakthrough ideas—he now turned to an opposite kind of problem, failures that can come from a lack of group coherence, from people seeming to work together, while actually having significantly different sorts of goals. “I have often wondered, if everybody has a hymnal, is everybody’s hymnal the same,” Dunning said. “So for political operatives, sometimes I wonder if their task is to get candidates elected or to make sure that they earn enough money from the election that they can live a good life, and can continue to have their political business.” It’s a question many in netroots have raised repeatedly over the years.
It’s not limited to politics, of course. Dunning also cited a similar problem in the business world, where “if you’re a CEO, you get graded on how well you do this past quarter, so the task is to maximize profit for this quarter. And that may not be the best strategy in the long-term.”
More generally, Dunning said, “Whenever you see somebody acting in what looks like an incompetent way you have to ask yourself if they actually perceive themselves doing a different task than what you think they're doing, and being very competent in [that task].” He cited an example from his own experience, a dean at a university he had observed, who “would speak in non sequiturs,” hardly a sign of intellectual competence, “until I realized this task was to outlast everybody, just so he got his way, and that he was brilliant. But if you thought his task was to bring many minds in together, and come to an understanding … no, that was not the task.”
What all these examples seem to suggest is that we can identify potential sources of failure—even if we can’t see them all. We will always be limited in our understanding—and limited in our understanding of our understanding. But we can still make progress, nonetheless, particularly if we stay mindful of our limitations.
Dunning offered two broad observations of what groups can do in this regard when they encounter recognizable failure—which would seem to be the most opportune time for change. First, he said, was to do “an honest autopsy of what just happened, and then to actually apply the lessons of that autopsy, as opposed to just go through the exercise and then do what you did the last time.” Dunning specifically noted the need to “talk to a lot of people, including people who might be opponents, people who certainly have an opposing view to people of your own, politically.” The Republicans' post-2012 autopsy doesn’t exactly look like it meets these criteria—but the midterm electoral disconnect Dunning mentioned earlier “saved” them from suffering as a result, which might well only make it even more difficult for them to change in the long run.
The second way out, Dunning suggested, was turning to history—which can still be tricky, since one must decide which historical examples are relevant, and what the suggested answers from them are. It’s popular among pundits, for example, to say that Bill Clinton saved the Democratic Party by moving it to the center—a claim that rather strikingly ignores the massive, historic losses at all levels in the 1994 election, losses that were duplicated even more deeply in 2010, after Obama had similarly sought to bring both sides together—and been spurned, just as Clinton had been. Still, the general principle seems sound: by looking to history, we can examine a situation as outsiders, and see things in it that we may well be blind to as insiders. As with the autopsy example, the possibility of successful reorientation is present, there is no guarantee it will be achieved.
There’s another line of research Dunning’s been involved with more recently, in collaboration with Clayton Critcher of U.C. Berkeley, which struck me as having direct relevance to my interest in bipartisan incompetence. This involves gaining deeper insight into how people construe the inner workings of others, based on what they know (or believe) about themselves, what Critcher and Dunning call “egocentric pattern projection.” It was my take, as a layman, that this could play an interesting role in how groups come to theorize about one another—and a possible source of blindness as well as insight.
As I explained in an email to Dunning before the interview, “I see this as connected, because I believe that both liberals and conservatives tend to misunderstand one another in various ways, and that one of those ways is via projecting self-derived assumptions onto the other. These projections both feed into and derive strength from beliefs that their cognitive competencies are all that's needed to win politically, and that their cognitive incompetencies aren't incompetencies at all.”
As with the Dunning-Kruger effect, egocentric pattern projection is not a wholly new idea at bottom; it’s long been widely understood that we tend to view the world through our own particular framework of assumptions. But, again, it’s a matter of seeing something old in a new light. A 2009 paper they co-authored showed that “If two traits go together in the self, then they are assumed to go together in other people. If two traits clash in the self-concept, then they are presumed not to co-occur in other people.” The shift from projecting traits to projecting patterns is not only significantly more sophisticated, it opens the doorway for a whole progression of further steps.
These weren’t controversial or political traits, but commonplace, representative ones—idealistic, perceptive, generous, wordly, resigned, bashful, reserved, prideful, considerate, persistent and dependent. Which is precisely why we’ve got every reason to believe that it’s generally true. But how much of a leap beyond that is it to the sort of group misunderstanding I was wondering about?
“It's not very much,” Dunning told me. Egocentric pattern projection “turns out in the main to be a good thing,” he said, “because you're a human like other people, and it turns out you can make some good guesses about other people,” but “It can get you into trouble when people do think differently, like when you have this divide in terms of mythos and logos.” He then went on to tell me about a new paper, not yet published, studying how people build explanations for the patterns they perceive. As that paper explains, “Causal trait theories—created to explain trait co-occurrence in a single person—are exported to guide one’s implicit personality theories about people in general.”
Obviously, once people have such theories, they are one step closer to sharing them with others, and through sharing, building collective theories—just the sorts of theories that could lead liberals and conservatives to misunderstand one another. We’re still a long way from knowing for certain that this broadly general process underlies what’s happening in our political culture today, but it’s starting to look more and more likely that it is.
While Dunning and others referred to are doing remarkable work to illuminate both the nature and limits of our self-understanding, a nagging question remains: How much does it all matter? By understanding how we misunderstand each other, for example, we may find better ways to overcome misunderstanding, as the work by Willer and Feinberg suggests. But that’s still presuming a common desire for mutual understanding, which may be found in a laboratory setting. But is it truly a realistic presumption to hold onto in America as a whole today? Or is it just another broad expression of liberals’ problem-solving, logos-based orientation, which conservatives fundamentally reject? And does the hope of finding common ground willfully ignore the role of reflexivity linked with conservative counterfactual beliefs to create conditions in which bipartisan problem-solving simply isn’t possible?
In short, it may be heartening when cognitive research suggests roughly symmetrical mechanisms and ways of overcoming differences, but that could be just another example of liberal intellectuals projecting their framework of assumptions, blinding themselves to more fundamental and intractable differences, which conservatives are, in their own way, smart enough to stick with and exploit, while depending on liberals’ relative disorganization, indecision and attraction to ambiguity to allow them to win the day, even if they can’t win an outright majority in a presidential election any more.
Once you’re aware that the Dunning-Kruger effect is involved, it’s anybody’s guess, really, who is more incompetent than whom.