Harvard Law faculty share 'Why I Changed My Mind'

ELI5/TLDR

Three Harvard Law professors sit down and each confess a big idea they used to believe but no longer do. Samantha Power thought you should never dignify a lie with a response — now she thinks you have to fight back. Nikolas Bowie came to law school believing courts were the way to change the world, then realized litigation only works when judges already agree with you. And Yochai Benkler spent 17 years betting that open-source, volunteer-built technology could route around capitalism — then watched Facebook, Apple, and Amazon swallow the commons whole.

The Full Story

Don’t Feed the Lie (Until You Must)

Samantha Power’s changed mind is about a piece of conventional wisdom so widespread it barely registers as a belief: don’t give oxygen to a lie. The logic is clean. If someone tells a falsehood about you, rebutting it just amplifies the original claim. Social scientists call this the illusory truth effect — the more you hear something, the more real it sounds, regardless of whether the repetition comes wrapped in a correction.

Power watched this play out at close range. During the Obama campaign, the birther conspiracy was treated as too absurd to dignify with a response. The birth certificate eventually came out, but late. During Trump’s first term, Power herself was accused of “unmasking” — an intelligence community practice she had never heard of and never done. She Googled it just to understand the accusation. By the time she testified under oath to deny it, the lie had already lapped the truth several times over.

As Mark Twain says, the lie had traveled halfway around the world before the truth had the chance to put its boots on.

Then at USAID, she faced it again: the agency was accused of smuggling weapons into Ethiopia’s Tigray conflict under cover of humanitarian aid. Same instinct — don’t give oxygen. Same result — the lie picked up speed anyway.

What finally moved her was research her husband, Cass Sunstein, pointed her toward. The so-called “backfire effect” — the idea that correcting a falsehood just makes believers double down — turns out to be far less robust than originally claimed. Even the original researchers have walked it back. It is possible to get ahead of a lie, though success depends heavily on who delivers the correction and how emotionally invested the audience already is.

Power landed on George Loewenstein’s “truth sandwich” as a practical framework: lead with the truth, reference the falsehood, then close with the truth again, always repeating the accurate version more than the false one. But she is careful to note the limits. In an era of deepfakes, any clip can be manipulated. In a hostile political climate, even allies who know the truth may stay silent out of fear. USAID was 90% earmarked by a Republican Congress, and those same Republicans ducked and covered when the lies started flying because they feared being targeted by Elon Musk. Bandwidth, she says, is the scarcest commodity in public life.

Litigation Is an Elite Theory of Change (and a Fragile One)

Nikolas Bowie arrived at law school with a straightforward plan. He had spent a summer with Unite Here organizers in Northern Virginia, watching the night-and-day difference between unionized DC hotel workers and their non-union counterparts across the river in Crystal City — different pay, different shoes, different sense of dignity. Virginia’s right-to-work laws made organizing dangerous. Law school, he figured, would teach him how to change those laws.

And it did — but only through one channel. Everything at Harvard Law was oriented toward judges. Read judicial opinions. Write like a judge. Argue before a judge. The heroes on the walls — Thurgood Marshall, Constance Baker Motley, Ruth Bader Ginsburg — all made their mark through litigation. Bowie was good at it. He won the Ames Moot Court competition. He clerked on the Supreme Court.

But the theory kept creaking. Two articles by Derrick Bell, himself an NAACP Legal Defense Fund lawyer, planted seeds of doubt. In “Serving Two Masters,” Bell described the tension between using litigation to advance a legal ideal (integration) while representing clients who wanted something more immediate (better-funded schools for their kids right now). In “The Interest Convergence Dilemma,” Bell asked why the same Supreme Court that decided Brown v. Board in 1954 was actively hostile to desegregation by the 1970s. His answer: judges have interests shaped by their identity and social world. When those interests converge with the people arguing before them, they rule favorably. When they diverge, they don’t. Brown succeeded partly because Cold War embarrassment made segregation costly for elites — the US government itself filed an amicus brief noting the Soviet Union was mocking American hypocrisy.

Litigation as a theory of change is just really contingent. It could end up being something of a dead language, like learning Latin — if you want to become a Catholic priest, it’s really important. But if you want to communicate with people in your neighborhood, maybe not totally relevant.

The final push came when Bowie was clerking on the Supreme Court in 2016 and Justice Scalia died. For one brief moment, a 4-4 court seemed to open a window. Then Trump won. Bowie realized he would likely never see a liberal Supreme Court in his lifetime. The question became: what kind of democracy is it when your ability to shape society depends on someone dying?

He found his way to a different theory of change through the Civil Rights Act of 1866 — a law in which Congress, not the courts, defined a list of rights states could not violate, then proposed the 14th Amendment to lock it in. The same Congress. The same session. Today the Supreme Court interprets the 14th Amendment in a way that would prohibit the very law that gave birth to it. That historical irony became the seed of Bowie’s current work: law schools should teach legislation and organizing, not just litigation. Students of every political stripe should learn to write laws, build coalitions, and exercise democratic power — not just make clever arguments to judges who may or may not be listening.

The Internet Did Not Route Around Capitalism

Yochai Benkler’s intellectual journey has the longest arc and perhaps the sharpest reversal. For 17 years — roughly 1996 to 2012 — he built his career on a vision he now considers fundamentally mistaken. The core idea: in the networked age, the most important inputs to advanced economic activity (knowledge, creativity, computing power) were for the first time since the Industrial Revolution widely distributed among ordinary people. This meant voluntary, cooperative production — open-source software, Wikipedia, Creative Commons — could bypass both the state and the market.

Benkler called this framework something between left libertarianism and anarchism. He was deeply skeptical of the capitalist state and doubted markets could serve most people’s interests. But he was equally skeptical that the state could be reformed from within. The answer, he believed, was to build around both — to create technical and institutional infrastructure (open standards, free software, decentralized networks) that eliminated control points and toll booths.

He spent those years writing amicus briefs against copyright expansion, advising regulators, working with technologists on wireless standards and network architecture. Some of it was popular (defending fan fiction). Some was not (arguing WikiLeaks was journalism and the State Department had overreacted to the embassy cables).

The reckoning came in 2012, when Erik Olin Wright, the Marxist sociologist behind the “Real Utopias” project, invited Benkler to write about what a real utopia looks like. Benkler looked around for new examples of voluntary cooperative production succeeding at scale. The list was the same as a decade earlier: Wikipedia, Linux, Apache. Maybe Signal. Meanwhile, the threat had arrived from a direction he never anticipated. He expected incumbents to fight back through law — intellectual property, telecom regulation, antitrust. Instead, Facebook, Apple, and Amazon built new structures of control and concentration through novel business models and technical layers above and below the open commons.

It’s not so much that the internet interprets censorship as damage and routes around it, as much as capitalism interprets anarchic, voluntaristic self-production as damage and routes around it — in fact, using it now as a resource to mine. Hello, Wikipedia.

The lesson, as Benkler frames it: they lost because they had the wrong theory of capitalism. They underestimated its creative, generative capacity to absorb and co-opt. And they had the wrong theory of the state — dismissing it as merely capital’s enforcer when it was, in fact, the only plausible source of countervailing power. He has since shifted his work toward law and political economy, focusing on the institutional structures that enforce market dependence. Even in conversations about AI, his answer circles back to the same structural point: freedom from want is the precondition for any meaningful autonomy, and labor markets for most people are less like contracts than like clicking “I agree.”

The Common Thread

The moderator, Jonathan Zittrain, notes — with some surprise — that the three panelists harmonized without coordination. All three arrived at variations of the same insight: your theory of how to change the world matters more than your theory of what the world should look like. Idealism without a realistic mechanism is, as Benkler puts it, just cycling without being connected to any gears.

An audience member distills it neatly: the tension between idealism and pragmatism. Benkler pushes back on the framing — pragmatism is not compromise. It is concrete, reality-based learning. Following the best available tools, quantitative and experiential, to understand how the world actually works, even when the answer is costly or psychologically uncomfortable.

Power closes with a thought borrowed from Bryan Stevenson and, unexpectedly, from Toyota’s manufacturing philosophy: get close. Go to the gemba — the actual place. Theories of change evolve when you bump up against realities that are hard to grasp from inside an institution.

Bowie gets the last word, and it is a quiet defense of ideas themselves. The first time he read Derrick Bell, he was not convinced. But years later, with experience, those articles became a lens he could not put down. Ideas may not convince anyone right away. But they plant seeds that experience can water.

Key Takeaways

The “backfire effect” is weaker than advertised. The widely cited finding that correcting a lie just makes people believe it harder has been substantially walked back — even by its original authors. Rebuttal can work, but effectiveness depends on source credibility, audience investment, and framing.
The truth sandwich: Lead with the truth, mention the falsehood, close with the truth. Always repeat the accurate version more than the lie. A practical heuristic from George Loewenstein for anyone in public life.
Litigation requires a friendly judiciary. It is a contingent, elite theory of change — Harvard graduates arguing before other Harvard graduates about a Constitution written by a Princeton graduate. When the bench is hostile, it becomes a dead language.
Interest convergence explains judicial behavior. Derrick Bell’s framework: courts side with you when judges’ own social interests align with your argument. Brown v. Board succeeded partly because Cold War geopolitics made segregation embarrassing for elites.
Law schools teach litigation but not legislation. Bowie’s observation that Harvard Law teaches students to argue before judges but not to draft laws or organize coalitions is a structural critique of legal education itself.
Capitalism routes around open-source commons. Benkler’s key insight: the threat to Wikipedia, Linux, and the open internet did not come from old incumbents using law. It came from new platform companies building control layers above and below the commons.
Voluntarism alone cannot counter structural power. The left-libertarian bet that decentralized cooperation could bypass both state and market was wrong. The state, for all its flaws, remains the only plausible source of countervailing power against concentrated capital.
Experience changes minds more than evidence. In five years of the “Why I Changed My Mind” series, almost no one has said “I read an empirical study and changed my view.” Lived experience does the work. Ideas plant seeds; experience waters them.
Pragmatism is not compromise. It is epistemological honesty — following the best available methods to understand reality, even when the conclusions are inconvenient or require abandoning years of prior work.

Claude’s Take

This is a genuinely interesting panel, elevated above the usual academic discussion format by the fact that all three speakers are confessing to being wrong about something fundamental — not just refining a position, but admitting that a core operating assumption led them astray. That takes more courage than it looks like from the outside, especially for people whose careers and reputations were built on the very ideas they are now questioning.

The unplanned convergence is the real story. Power, Bowie, and Benkler come from different corners of the legal and political world, but they all landed on the same structural insight: having the right values is not enough if your theory of how change actually happens is wrong. Power learned it from watching lies outrun truth. Bowie learned it from watching courts become hostile territory. Benkler learned it from watching Silicon Valley swallow the open internet. Three different data points, same conclusion: understand the mechanism, not just the goal.

Benkler’s arc is the most intellectually dramatic. Spending 17 years building a body of work on left-libertarian techno-optimism and then publicly saying “we lost, and here is why” is not something most academics do. His reframing of the famous internet adage — capitalism interprets voluntaristic self-production as damage and routes around it — is the single best line in the conversation.

The panel is not without limits. It is very much Harvard talking to Harvard about Harvard. The moderator nearly says as much. And there is a certain comfort in confessing to intellectual evolution when you have tenure at the most prestigious law school in the country. The stakes of being wrong are lower when your job is secure. But that is a criticism of the venue, not the ideas.

Score: 7/10. Rich, honest intellectual content with real self-criticism from people who have earned the right to be self-critical. Not groundbreaking for anyone already familiar with critical legal studies, interest convergence theory, or the platform capitalism literature. But the personal dimension — three people tracing how experience forced them to revise deeply held beliefs — makes it more than the sum of its academic parts.