In May 1980, students in the South Korean city of Gwangju rebelled against the unpopular military regime. Many hundreds were brutally murdered by paratroopers sent in to quell the uprising. General Chun Doo-hwan, the leader of the military government, claimed that the students were North Korean revolutionary stooges.

Over the following two decades, South Korea became a democracy, and Chun was put in prison. While Korean liberals still mourn the students of Gwangju as martyrs to democracy, some conservatives believe that Chun was right to see the uprising as a North Korean plot. Now, South Korea’s current liberal president, Moon Jae-in, is pushing for laws to ban such views as “historical distortions.” Denying that the Gwangju uprising was anything but a quest for freedom can now land a person in jail for five years. Praising aspects of Japanese colonial rule in Korea can lead to an even longer prison sentence.

Proponents of such legislation in South Korea point to laws in several European countries that prohibit denial of the Jewish Holocaust. Opponents, meanwhile, regard such laws as an attack on free speech, arguing that governments should not be allowed to decide what is right or wrong in historical debates.

There are historical facts, of course: Auschwitz existed, atom bombs were dropped, and students were killed in Gwangju. But much is also open to interpretation. Bad arguments and falsehoods must be contested with better arguments and more accuracy.

That is the ideal case for free speech, anyway. In reality, legal and social constraints exist everywhere, often for good reason. Inciting hatred and discrimination on the grounds of race, creed, or sexuality is illegal in the European Union. Though the constitution of the United States is less restrictive, it still bans speech that directs or incites “imminent lawless action.” Similarly, US courts do not extend free-speech protections to child pornography or defamation.

Is this enough? Is the ideal of free speech not a little naive in an age when a US president can spread noxious lies to millions of voters via the internet? Should dangerous conspiracy theories that aggravate a global pandemic or undermine democratic institutions be banned from social media? Are better arguments and more accuracy enough to stop these falsehoods from doing serious damage?

Since I believe in free speech, I don’t like laws against Holocaust denial and other abhorrent opinions. But this position must be tested against the clear risks of letting some of the most poisonous views circulate. Many people believed that allowing Nazi propaganda to spread in Germany after World War II would have posed an imminent danger to the country’s fragile liberal democracy. This was not an unreasonable assumption. Banning such propaganda made sense at the time.

A common practical argument against outlawing crackpot theories once held that they were marginal, and thus relatively harmless. Before the age of internet and social media, the idea that Hillary Clinton and George Soros were running a global network of cannibalistic pedophiles would have been limited to a lunatic fringe. But now, millions of people around the world—including as many as 50% of Republicans in the US—say they believe such nonsense. A cult is not open to arguments; confronting believers with facts would be missing the point.

Several European countries, as well as the EU, are working on laws to regulate internet platforms. But asking governments or social-media platforms to censor irrational and harmful beliefs is unlikely to get rid of them. True believers will only be strengthened in their conviction that they are under siege from a malevolent establishment.

Even if bad ideas could be curtailed by censorship, would it be the right thing to do? Here, I think the famous Skokie case is still relevant. In 1977, the National Socialist Party of America wanted to demonstrate in a Chicago suburb where many Jews lived, including survivors of the Holocaust. Prompted by complaints from the local population, the municipal officials tried to stop it. The Nazis claimed their right to free speech, which included waving Swastika flags. This right was defended by lawyers of the American Civil Liberties Union. The case went all the way up to the Supreme Court, where the right to free speech was upheld. No matter how unpleasant, Swastika flags were deemed permissible, because they did not qualify as “fighting words”—a narrow category of speech that is denied the standard constitutional protections.

The argument put forward by the ACLU lawyers—some of whom were Jewish themselves, and none of whom had any Nazi sympathies—was simple: if you allow the state to ban opinions you oppose, you make it easier for the state to ban views you agree with. Protecting the right of Nazis to demonstrate was seen as a way to protect the right of others to take very different views. This argument still holds, even in our digital age.

But even in the US, which is more lenient than most countries, the principle cannot be absolute. Inciting imminent violence is not permitted. Donald Trump’s speech on January 6, urging the mob to storm the US Capitol, certainly came close to overstepping this boundary. It was a clear demonstration that language can be dangerous. What the internet media has done is raise the stakes; “fighting words” are spread around much faster and more widely than ever before. This will require a great deal of vigilance, to protect our freedom to express ourselves, while observing the social and legal bounds that stop words from turning into actual fighting.

—Project Syndicate