The Tyranny of the Minority—and How to Prevent It

When the internet first became part of mainstream public life, one of the great hopes it seemed to hold up was that of an expansion of democracy. From now on, the great media conglomerates would no longer dominate the conversation. Instead, the public sphere would be irrigated by new voices, the voices of ordinary people. And that’s exactly what happened—partly thanks to social media platforms like Facebook and Twitter, which allow people to post their views on pretty much anything, from what to do about Brexit to whether the earth is flat.

But the conclusion that the internet has made things more democratic is less clear than it might seem. As well as allowing more people to have a say, the web has enabled small bands of especially vocal activists to dominate the conversation. And not only that—persistent complaints by activists have even led to books being withdrawn from publication, shows being cancelled and people being sacked.

The number of complainants can often be quite small. For centuries, theorists have worried about the potential of unrestrained democracy to lead to a tyranny of the majority, in which majority groups ride roughshod over the rights of minorities. What we often see today is instead a kind of tyranny of the minority: a system in which a particularly extreme and motivated fraction of the populace can wield outsized power in the face of a majority which is either too indifferent or too scared to oppose it.

The claims of the activist minority often draw much of their strength from a tacit assumption that they represent a far larger body of opinion. Complaints about cultural appropriation, for example, rely on the usually unchallenged idea that one representative of a group can speak for all or most of that group. If someone says, I don’t like white people wearing sombreros, we have no reason to treat it as anything more than an individual opinion. If the claim is instead, As a Mexican, I can tell you that by wearing a sombrero to a Halloween party you’re insulting Mexicans, that might seem to justify further action, even if the crucial claim that all or most Mexicans would care about this isn’t backed up.

But the question of how many people the complainants actually have on their side is even more fundamental than that. That’s because numbers are the only thing that can ultimately adjudicate one of the key principles of liberalism: the harm principle, formulated by J. S. Mill. Put simply, the harm principle states that we should all be able to do whatever we want, as long as it doesn’t harm anyone else. As generations of Mill’s critics have pointed out, what counts as harm is often a matter of interpretation. If I tell a dirty joke in public, and you complain about it, have I harmed you or not? Who’s to say?

The answer, ultimately, is the people. That is, on a pragmatic level, we deal with the ambiguity of Mill’s principle by passing laws which reflect most people’s idea of what constitutes harm. That’s why it’s not against the law to say something with which you might disagree, but it is against the law for you to punch me in the face. The idea that getting smacked in the face constitutes a harm enjoys widespread agreement, whereas the notion that you saying something that I might not think is true constitutes a harm is not something that most people would agree with, at least not at present.

And the procedures we use to make the law are designed to give us a more or less accurate sense of what people’s views really are. In our representative democracies, that means that laws are made through voting, by politicians who have themselves been selected by some method that is responsive to the popular will. Of course, there’s no absolutely perfect way of doing this, and you might well think that the electoral systems that we have at the moment fall especially far short of perfection. But the system is designed to give us a sense of the balance of opinion in society, and some of its slightly unusual features (secret ballot, for example) help it do that better than some of the less formal methods we might turn to.

What other things might help us get a better sense of what people are really thinking? One crucial issue for pollsters is sampling: that is, making sure that they get information from as broad and representative a cross-section of society as possible. They need to do that to get a sense of what the average person is thinking, not just the small sub-set of people who are particularly motivated to speak up on a particular issue. Self-selected polls, which rely on people coming forward, are widely seen as junk for precisely this reason. The people who step forward voluntarily may well have a heartfelt view to express, but they may also make up only a tiny fraction of the population.

Of course, all this assumes that what we should care about is the number of people who want something, not how much they want it. That assumption has long been a bedrock assumption of our democracies, but we could run our affairs a different way. Various historical societies have relied on voting procedures that take account of the intensity of people’s feelings as well as their number. In the assemblies of some Celtic tribes, for example, warriors would vote by banging their weapons together, with victory going to the side that produced the most noise. In ancient Sparta, some officials were chosen by having people shout for the candidate they liked the most. Obviously, this made it possible for the side that was smaller in terms of number, but more passionate about their cause, to win the day.

But, if the scarcity of societies that still use such methods isn’t indicative enough, we might recall that there are very good reasons why we make decisions based purely on how many people like a certain option, not on how strongly its supporters feel about it. With one person one vote, there’s no possibility that anyone will be able to get more power than anybody else (at least as far as the vote is concerned). If intensity of feeling is allowed into the picture, people might start to become more passionate just to steal a march on their opponents. That would create an unhealthy dynamic in which voters were incentivized to become ever more zealous about their own viewpoints. It might also mean that moderate, unexciting policies would lose out to options that got some people’s blood up. And, crucially, people who didn’t feel as passionately about issues would have less of a say, meaning that it would be more difficult to call a system like that democratic.

Arguably, though, such a system is already in place in certain areas of public life. Not at the level of the state, of course—there, the formal procedures of representative democracy still hold sway. But, in our everyday lives, especially in the world of culture, a less formal system has apparently grown up, on the social media sites that once seemed to hold so much promise for democracy. Which books get published (or withdrawn), who gets hired (or fired) as a presenter and so on—decisions on questions like these are increasingly influenced by self-selected minorities, who accrue authority not through their numbers, but through the intensity and persistence of their complaints.

What do those of us who care about democracy do about this? One thing we can do is try to figure out, before responding to a complaint, whether it actually commands broad support or not. So, for example, a comedy club might get a complaint about an offensive act. Fair enough—what constitutes offensive is about as ambiguous as what constitutes harm. So let’s allow the people to decide: in this instance, the audience for that particular act. How many of them agree that it was offensive?

It may seem like a chore to reach out to every audience member and ask them all. But the alternative is that a club is rushed into taking action because a tiny minority of the audience felt strongly. That will lead to the kinds of problems that we’ve already seen come with self-selected politics: it encourages extremism and ignores the average audience member’s desires. It will also have costs for the comedian who has action taken against him, for the culture of contemporary comedy more broadly—and, in all likelihood, for the bottom line of the club. After all, if a club takes action against someone whom 90% of the audience thought was funny, that will hardly be good for business.

That just underlines how important it is to get these things right. But is it really practical to ask every audience member what they thought of a joke? Well, yes—in fact, the same advances in technology that have made it so much easier for people to make complaints should also make it easier for us to ask others how they felt. All we have to do is make use of them. So, for example, a theatre could, in the case of a serious complaint, use its record of ticket purchases to send a quick survey to everyone who was there on the night in question via email or social media. If a majority of the audience agree with the complaint, the theatre might take action; if not, it could reject the complaint with the added authority of the majority will.

Still skeptical? If so, it’s worth bearing in mind that something of this sort has been going on for a while. What is a counter-petition, if not an attempt to show that more people oppose a change than support it? Some companies have started trying to gauge the broader public’s views on certain issues in a more systematic way: for example, a US radio station recently put the wintertime classic “Baby, It’s Cold Outside” back onto its playlist after running a poll that found that 95% of respondents had no problem with the song. (It had previously been labeled problematic). Of course, petitions and polls, if sloppily designed, can fall into the same self-selection trap that we’ve been discussing. Which is why I’m suggesting we try to design procedures that can give us a quick but accurate picture of what people’s actual views are in these types of cases.

Nothing I’ve said here should be taken to imply that people who want to make a complaint about something don’t deserve to be heard. Of course they do—but, unless they can demonstrate criminal wrongdoing, they have no automatic right to have their complaint acted upon. In our democratic societies, politics and culture should be shaped by what all of us want, not by the whims of a few particularly riled-up activists. The tyranny of the minority has made too many inroads already. Allowing it to continue would constitute a serious erosion of our democratic culture. The way to defeat it is not to try to roll back or place limits on the cacophony of voices that the internet has made possible. It is, instead, to use the technology we now have at our disposal to make sure that it’s the people’s wants that become reality. What we need, in other words, is not less democracy, but more.

If you enjoy our articles, be a part of our growth and help us produce more writing for you:


  1. “The claims of the activist minority often draw much of their strength from a tacit assumption that they represent a far larger body of opinion.”

    This is nothing new. Consider the unsupported ‘moral majority’ claims of religious conservatives and the efforts they sometimes made to ban or restrict (say) music albums. Whether a majority or a minority, a ruthlessly vocal group that manages to get what it wants at the expense of others can be considered a ‘mob’ and it was the risk of ‘mob rule’ that concerned old liberal-democratic theorists.

    You could reflexively use ‘more democracy’ as you suggest for a comedy venue. You could also preemptively use the liberal value of greater transparency. Those religious conservatives I mentioned who objected to particular albums got a compromise – parents-advisory labels. This allowed everyone to make informed choices. Similarly we could re-assert the usefulness of ratings and warnings for other creative endeavours. They currently serve the interests of children (assuming that parents ever pay them any heed) and could do the same for sensitive adults. The point however would be that those who object can stay away so that others can continue to enjoy themselves.

    1. Interesting – I think this suggestion would definitely work in certain cases. So, for example, with smoking in bars and cafes. If bars that allowed smoking just had signs on the window signalling that, and bars that banned smoking had signs signalling that, people could choose which bars to go to, and the number of smoking and non-smoking bars would presumably adjust to a level roughly in line with demand.

      The problem with ‘offensiveness’ is that it’s not as clear-cut as whether people are smoking or not. I guess you could create a kind of parental-advisory label (but for adults) for comedy venues, where an act with a lot of swear-words or daring jokes would get an R and less audacious acts would get a PG or something equivalent. I wouldn’t be massively opposed to that, though you’d have to be careful in how you set up the panels that handed out the labels to various acts.

      I’m not sure, though, that a labelling system would necessarily solve the problem I’m discussing in this piece. To put it in 90s terms, the problem these days isn’t that kids are watching things they didn’t want to watch and getting scared – it’s that some kids are seeing things they didn’t want to see and then insisting that the producers are fined, fired, or disgraced, or sometimes even erased from all previous artistic productions. And it isn’t even that the kids are seeing these things by mistake – they’re actively seeking out content they don’t like so they can persecute its makers.

      This is what I see as a real danger, and even a sort of tyranny, especially when it violates basic rights like the freedom of expression. If the activists/vigilantes try to use the argument that the majority will should prevail (which I agree is a central democratic principle) it’s important that we actually test their implication that the majority will is on their side. Effectively what’s going on is that elements within civil society are becoming able to inflict the kind of punishments that were once reserved to the state, but without the voting procedures and safeguards that were designed to make sure that the state’s decisions had some democratic legitimacy. So extending proper voting procedures into civil society might be a way to make sure democratic principles are properly applied in our everyday cultural lives.

  2. Do we really need to gather such feedback about an act’s offensiveness? If enough people find the act distasteful, word will get out and people will stop spending their money on it. Employers of any kind should not be called upon to act as morality police: their concern should remain fixed on simply making money.

    1. I basically agree with this, but the problem is that employers effectively are being called upon to act as morality police – by activists insisting that so-and-so should be punished in whatever way for what they said. I’m suggesting that employers shouldn’t feel forced into playing out the role of morality cops, and the idea of holding votes of the relevant audiences is meant to help them make decisions based on what people are actually thinking. Companies or event managers can’t often predict what’s going to happen, so if they see someone on Twitter saying ‘This is really offensive, and I’m telling all my followers to boycott your products/events!’ it makes sense that they’ll be scared about the future of their businesses. Arranging a poll would give them more information about how many people agree with the complainants. Finding out that a clear majority of people don’t find something offensive allows them to face the future with confidence, and also informs complainants that (though their complaint might have been made honestly) they’re in a small minority on the issue.

  3. I believe the issue is further complicated by definitions. If the majority agrees that the purpose of comedy is to offend and that does not make it ‘wrong’, then it contrasts with those who claim that it was offensive and wrong. So we have two meanings of the word offensive. Similarly, if you complain about a workout being tough you are not necessarily denouncing the exercise if you agree that a workout should be difficult. Thus, we encounter a clash of worldviews at the fundamental level of purpose and intention.

    1. Thanks, Vincent. But I think this is precisely the kind of issue we can vote on. I may find something inoffensive and you may find it offensive because we have different definitions; that’s fine. We can even argue that our definition is better than alternatives. Then everyone can vote (one vote each, of course), and their own view about what ‘offensiveness’ is will feed into the result.

  4. While I agree with the main thrust of this article, I worry that the proposed solution risks turning back into a tyranny of the majority. While that’s better than a tyranny of the minority, it’s still far from ideal.

    Often in liberal democracies, concessions are made to certain minorities, even on things that the majority of people do not view as harmful. For example, if a religious minority views a certain thing as harmful on religious grounds then, even if most people don’t see it as harmful, often accomodations will be made for the minority, as long as said accomodations don’t violate someone else’s idea of harm. If the majority of people don’t care very strongly about something, then it seems likely that any proposed accomodation wouldn’t meet their definition of harm, so it might make sense to side with the minority who does see the thing in question as harmful.

    The classic example of this is 3 people voting on where to go for dinner. 2 people would prefer a certain restaurant, but the 3rd is allergic to almost everything on the menu. Of course, if the other 2 are decent people, you’d expect them to change their votes to accomodate the 3rd. However, if they’re not aware of the reason why the 3rd person objects to going to that restaurant, they might vote for it anyway, likely leading to a miserable, and possibly dangerous, experience for the 3rd.

    Now, offensive jokes aren’t going to kill people (at least not directly). However, you could see a similar sort of situation; if you just send a survey to the audience after a show asking if they found certain jokes offensive, they might say they were ok with them, when they might have had more sympathy for the people who found them offensive if they knew the specific reasons why. I guess you could call this a “tyranny of the indifferent”.

    I do agree however, that when someone claims to speak for a certain community, they should show some evidence that a majority of said community actually agrees with their position.

    1. The story you describe is the tyranny of the _intolerant_ as described by Nicolas Taleb. Clearly when someone can die like an allergy the harm is clear and if he is a friend clearly one is willing to adapt.,However how does this translate to a single vegan in a group of rugby players? Should all German kids eat halal food at school because 5-10% of them are muslim? The problem with your position is that all the choices that any minority makes further restricts the life of the majority that wants to eat wurst. In Holland there is a yearly discussion about Zwarte Piet, an old tradition that has a very old (500 year or so) saint with a black faced helper. A rather vocal group in Holland wants the majority population abandon that because they deems the blackface racist and offensive. (Legally it has been decided now several times it is not racist.) Mainstream press argues that Zwarte Piet could also be blue or orange but forcing people to give up an old tradition is a non-trivial thing to ask traditional people since it requires a recoloring of your childhood. A majority can give in to this intolerance of a minority but in a multicultural society there are a lot of minorities! The current PC culture is a direct consequence of the majority that desperately tries to take all these intolerant demands into account. Each one of them is not big enough to resist but chinese torture is also one drop at a time. At the same time the majority is (justly) demanded to be tolerant of the habits of minorities.

      How can a multicultural society can survive if any minority can force the majority to adapt (however little) to their intolerances?

      If a joke is offensive to you, shrug your shoulders, walk away, and never visit the theater again. That is the power you have. Using mob power to force other to conform to your intolerance is imho quite despicable.

    2. People do indeed get killed for offensive jokes. Ask Charlie Hebdo or Danish cartoonists. How quickly we forget!

      This should illustrate the problem with your contention that the majority should cave to religious demands. Elevating a religion as above criticism despite serious cultural problems is poisonous to a society and need not be tolerated by the majority.

      1. Thanks, Ben – that’s a very thoughtful comment and an understandable worry to have. As you know, a lot of democracies (most famously the US) have tried to deal with the threat of ‘the tyranny of the majority’ by writing a set of basic rights into their constitutions. Some radicals see that as itself undemocratic, since it basically puts some people’s conceptions of rights beyond the reach of the decision-making powers of the people as a whole. And, in any case (they’d say), one person one vote has already granted everyone political equality, so what complaint can defeated minorities have? On the other hand, some left-liberal democrats like Joshua Cohen have argued that ‘liberal’ rights should themselves be seen as part of what’s needed to make democracy function. How, they might ask, could there be a real democracy without everyone having an inalienable right to free speech? I basically agree with that view, though I think that left-liberal democrats sometimes fall into a kind of ‘mission creep’ that leads them to defending rights that can’t plausibly be seen as necessary to democracy tout court (e.g. various sorts of welfare entitlements – not that I’m against these per se; I’m just not sure they’re essential to democratic systems of government). So I’d defend a package of liberal rights, but a more minimal one than some left-liberals.

        With regard to how this affects the proposal I’ve made in this piece (basically, for voting about what’s offensive or not), I’d say this. Rights are powerful things – as Ronald Dworkin put it, they can be seen as ‘trumps,’ that is, as interests that over-ride and defeat any other considerations that might come up. That means that they’ve very properly been the subject of a lot of legal and political contention – legal and political contention that often involves a good number of people, and reflects the will of the nation as expressed in elections, referenda, and so on. I think we should pay attention to that, and not make claims about rights lightly. If someone does, for example by saying ‘It’s my human right not to be at a party where white people are wearing sombreros!’ I think the right response is simply to gently remind them of the fact that that’s not how rights work. They can’t just be generated by individual proclamation – if they could, everyone would be proclaiming new rights all the time, and we’d all be forced to honour them, even though that might not be logically consistent or even possible.

        As to your story about the friends going out to dinner, I basically agree with Peter and Stephanie’s reactions. As you say yourself, offensive jokes aren’t the same as an actual human-rights violation. That’s partly why they seem to me to be very much the kind of thing we can vote on. And when you say that people might want or need to get a better understanding of why some found something offensive, I think that’s a valuable point, and I have nothing against it as a suggestion. I think I was always imagining that this would play out over a few hours or days: a show would happen, someone would complain online, and then some time later there’d be some discussion before a final vote called by the event-organiser. That would hopefully allow some of the indifferent people to consider the reasons the complainant put forward for complaining. Of course, whether those people accepted those reasons or not, or thought them strong enough to justify action – that would be up to them.

  5. Lenny Bruce addressed this problem of comedians’ offending people back in the 1950s. At considerable cost to himself, Lenny handled this by continuing to be himself and tell his jokes, critics be damned. We are all the better for his defiance. People today should follow his example.

    1. I agree that modern comics should just plow ahead, as Bruce did. However, I do wonder what would have happened to him in the age of Twitter mobs and employers (in this case, comedy club managers) buckling under to their demands.

      1. Thanks, Marian – I agree that comedians are often great heroes for free speech – Lenny Bruce in the 50s and people like Richard Pryor, Chris Morris, Ricky Gervais, Sarah Silverman, Bill Burr, and Andrew Doyle in the decades since. The zany and ribald humour of comics like Aristophanes also played a central role in the culture of democratic Athens (I’ve written about this here: https://www.academia.edu/37373191/Deliberative_Democracy_and_Aristophanic_Comedy).
        S. Vitkovitsky, I agree that the problem nowadays is that employers (and even authorities) react too quickly and strongly to complaints. That’s why I’ve tried to suggest a quick way (voting, basically) for such people to ascertain how many people are really offended. I think they think that reacting to complaints is kind, or they feel they might as well cave into the pressure. But it’s not kind to ruin one person’s career and a lot of people’s enjoyment, and caving into the pressure in the short term only incentivizes bullies, creating an atmosphere where anyone can be mobbed by activists with a lot of time on their hands.

Leave a Reply