Arnold Kling on the Three Languages of Politics
What this covers
Arnold Kling joins Russ Roberts to explore how Progressives, Conservatives, and Libertarians process political disagreement through distinct moral frameworks. Kling argues that each group organizes the world along a separate axis (political language))—the oppressor/oppressed axis for Progressives, civilization/barbarism axis for Conservatives, and freedom/coercion axis for Libertarians—and that these axes function primarily as tools for reinforcing group identity and dismissing opponents rather than as genuine explanations of belief. The conversation pivots on a counterintuitive claim: recognizing these separate languages offers a path toward intellectual humility and reduced tribalism, not because it guarantees persuasion, but because it prevents the automatic mistake of placing ideological opponents at the opposite pole of your own axis when they operate on an entirely different one.
The episode ranges across economics, psychology, and media strategy to show how the framework works in practice. Roberts and Kling examine why more political knowledge correlates with greater polarization, why exposure to contrary facts often strengthens rather than weakens existing views, and how punditry often aims to cement tribal loyalty rather than bridge disagreement. Specific terrain includes the structure of public schools, health insurance outcomes, and how economists like Gary Becker lose audience share to figures willing to abandon economic reasoning and adopt tribal language. Throughout, Kling positions Constructive Reasoning as a learnable habit—the capacity to examine ideas objectively without reflexively filtering them through one's preferred axis—and proposes that a shift from Engaged vs Engaged Pundit toward detached observation could reduce political intolerance without requiring anyone to change their actual beliefs.
Kling argues that Progressives, Conservatives, and Libertarians each communicate along a distinct moral axis (oppressor/oppressed, civilization/barbarism, freedom/coercion), and that these axes are tools for reinforcing tribal solidarity and demonizing opponents rather than explanations of belief, so recognizing them can reduce political intolerance.
- Each tribe reframes every issue into its preferred axis and treats opponents as occupying the bad end of that axis
- The axes describe how people express and process beliefs, not why they hold them (contrast with Sowell's 'why' explanation)
- Understanding others' axes is offered as a self-help tool for detachment, not a guaranteed persuasion technique
Opponents operate on different axes than you; engaging them requires entering their framework, not proving them wrong.
- When someone disagrees with you, the natural habit is to place them at the opposite end of your single axis, but in reality opponents are often orthogonal to you—operating on a different axis entirely—which is a place they would never put themselves.
“we have this natural habit of dividing the world into one axis, really, with one end that's good and one that's bad; and to shove people into the--they disagree with us, they must be at the other end. But they are orthogonal to us, it turns out.”
- The common assumption that you change someone's worldview by proving them wrong does not work—it fails in proselytizing for religion and ideology alike—so encouraging someone to think as you do requires putting yourself in their shoes and using their axes rather than simply refuting them.
“most people think the way to do that is you just prove they are wrong... And that doesn't work. It doesn't work in proselytizing for religion; it doesn't work in proselytizing for ideology... if you want to encourage someone to think the way you do, you ought to put yourself in their shoes and use their axes.”
Factual observations get misinterpreted when mapped onto an opponent's moral axis instead of engaged directly.
- When citing the RAND study (showing larger health-care subsidies led to more usage but not better outcomes) to a journalist, Russ found the journalist react with fury, treating 'the RAND study' as code words meaning 'you don't care about poor people getting health insurance'—an example of placing a factual observation onto an opponent's moral axis rather than engaging it.
“he got furious... Those are code words; that means you don't care about poor people getting health insurance. That's really what the point of your book is: he's on that axis, and I was on some other axis.”
People symmetrically overreact by treating single studies as proof of absolute conclusions across the political spectrum.
- There is a symmetric tendency to overreact to single studies: just as some treated the Reinhart-Rogoff error as proof that any level of debt is now acceptable, others treat the Oregon Medicaid Study as proof that Medicaid has no effect and can be eliminated—both are overreactions to one study.
“we are having a similar blogosphere explosion overreacting to one study; therefore, the Reinhart-Rogoff thing is wrong so any kind of debt is allowed... There are some results for the pro-Medicaid side; but most of the results are for the anti-Medicaid side. So now it's: Well, Medicaid doesn't have any effect; okay, we can eliminate it. A little bit of an overreaction.”