Can incivility be stopped?
What are its causes, and what might it take to reduce incivility? Quite a lot.
Is there a crisis of incivility? The marketing firm Weber Shandwick in 2013, five years ago now, conducted a survey which concluded that incivility had reached crisis levels. Research published by the American Psychological Association in the same year concluded that incidents of incivility had lasting detrimental effects. Incivility in other areas of society has been bad for even longer. Books were being written about incivility in academia in 2008. According to a study published in the “Journal of Occupational Health Psychology” in 2001, 71 percent of employees had reported incivility issues in the workplace. Things seem to be even worse in politics and on the internet. If incivility has gotten much worse recently, basic aspects of human psychology, which remain stable across generations, are unlikely to be the culprit. What’s going on, really? Can it be fixed?
It’s easy to make superficial analyses and repeat outrageous stories that don’t lead to deeper understanding, don’t provide any insight into whether fixing incivility is even possible, and don’t provide hints about what to fix to make lasting changes. Identifying root causes of problems like incivility is not easy, but I think I have it mostly figured out. There are a lot of causes that reinforce each other. Changes in any single aspect might be tolerable, but when they are all put together, their cumulative effect creates a serious problem. It will take fixing things on many fronts to improve the situation, and some subtle but profound changes in the way things are done will help. A single panacea is not available, but none of the required changes are impossible. The list I’ve compiled here is of course not exhaustive. Social systems are incomprehensibly complex, and no amount of analysis or remediation will address all the problems. But we have to start somewhere, and the aspects listed here should bend the curve in the direction of civility.
First, politics
Many news media outlets view politics as a spectator sport. You choose a team, and root for them and against their opponents. Just as in other sports, to the political commentators working for profit-maximizing companies who want to maximize their audiences, there’s no right & wrong. There’s only winning and losing; better or worse is defined not in moral or ethical terms, but in terms of winning more or losing more. The best election is the one with the most drama, not one that most effectively determines the will of the electorate.

To maximize the drama, media provide free exposure, labeled as “news”, to the most outrageous candidates and campaign incidents.
“First past the post” elections with runoff elections in case neither candidate obtains a majority make negative campaigns effective. It’s not necessary for candidates to have any good qualities, they just need to be less evil than their respective opponents. Although scurrilous characterizations of political opponents have been around as long as elections, the superior power of negative campaign ads to sway voter opinions seems to have been discovered in the 1980s, and campaign advertising has gone downhill ever since.
The system of choosing candidates by primary elections has driven campaign positions towards extremes. In the bad old days when candidates were selected by party officials in smoke-filled conference rooms, those candidates tended to represent the center of their respective party attitudes. In a primary system, candidates have to differentiate their positions from other candidates in the same party. This means that the party’s agreeable, uninteresting center is left empty, and each candidate is more extreme. Then in general elections, voters are given a choice between extremists that don’t even represent their own parties, much less the non-partisan center.
On the internet social media
The advertising-driven imperative of engagement means that social apps use outrage to attract and keep your attention. One of the alternatives to Twitter, Mastodon, stated it well in their “elevator pitch” writeup (Mastodon has its own problems, but they got this analysis right) “often, tweets are at their most engaging when they’re at their most enraging. Any tweet that keeps people clicking and talking — even if they’re talking about how awful the other side is — is a tweet that keeps people around to view ads. Outrage is good for Twitter’s bottom line.”
Short statements such as the 240 characters of Twitter require writers to focus on the point, and not surround it with a context of soft, civilizing verbiage. Short, hard-hitting statements are intrinsically uncivil.
Once the size of a person’s online “social network” exceeds their Dunbar number, messages to and from significant portions that network are effectively anonymous. The best that can be done to communicate with a large number of followers is to create an artificial, mediated persona that provides the illusion of personal engagement, and takes pains to keep sensitive affairs of private real life from leaking through into that persona.

This quasi-anonymity creates an additional degree of immunity to repercussions from uncivil statements, in addition to the actual anonymity that comes from the physical distance between people online and from the pseudonyms that are pervasive in online interactions.
Pseudonyms have a long history of literary noms de plume, which were used when the content of a document might be offensive or embarrssing if associated with the actual author, but those were rare. Pervasive pseudonyms online arose from the memory limitations of early computers, which forced every user to be identified by a short ID code. In days when a decently large mainframe computer was equipped with what seems today to be a tiny 300 megabytes of disk storage and memory was measured in kilobytes, it was generous to allow a username to be as long as 8 or 12 characters. Showing my age: the first multi-user computer I used had usernames composed of a “project” component and a “programmer name” component, of 3 characters each, stored in two memory “halfwords”. [Note 1] Decades later, the physical limitations that drove the use of handles rather than real names no longer exist, but the practice remains.
Facebook’s real names policy might have reduced incivility when it was a small, college-scaled network, and the physical limits of architecture and geography constrained communities to a few thousand or hundred thousand members. When those limits are exceeded, prohibiting pseudonyms is just not powerful enough to have much effect. Instead of harassing just a few passers-by from under a bridge, a troll can now annoy people around the world.
On society in general
Most society is not uncivil, except to politicized snowflakes who get outraged over every microaggression. Most people have thicker skins, and choose not to pull out their concealed-carry pistols and have a shootout with everyone who cuts them off in traffic.[Note 2] As a guy with longish hair in a ponytail, I haven’t been publicly discriminated against since the 1970s.
News sources tell us that things are worse than they really are, because when nothing bad happens, it isn’t news. Headline-hungry news organizations must chase the latest disaster and give “breaking news” airtime to trivial or distant events in order to fill their continous news cycle when there’s nothing really immportant to talk or write about.
It’s not just news organizations, it’s all of us. Researchers have recently documented “prevalence creep”. The editors of Science magazine characterized it this way, summarizing a study published in June 2018: “Do we think that a problem persists even when it has become less frequent? David Levari et al. show experimentally that when the “signal” a person is searching for becomes rare, the person naturally responds by broadening his or her definition of the signal — and therefore continues to find it even when it is not there. From low-level perception of color to higher-level judgments of ethics, there is a robust tendency for perceptual and judgmental standards to “creep” when they ought not to. For example, when blue dots become rare, participants start calling purple dots blue, and when threatening faces become rare, participants start calling neutral faces threatening. This phenomenon has broad implications that may help explain why people whose job is to find and eliminate problems in the world often cannot tell when their work is done.”
Can incivility be cured?
How will we know when civility is sufficiently restored to public life? Prevalence creep ensures that any institutions that arise to fight incivility will work to amplify any remaining incidents, even beyond their natural institutional imperative to preserve the conditions that justify their existence. Although enforcers would surely be very considerate in their enforcement, a world that supports “politeness police” is not one where I would care to live. “Miss Manners’” column in the daily newspaper is quite enough, thank you kindly.
Just as institutions like the OECD, the UN and a handful of countries have developed measures of “gross national happiness”, social scientists will need to create measures of gross national civility, and track them regularly. Five years after Weber Shandwick’s study, have things gotten worse (probably) or better?
Fixing political polarization
To fix political polarization, and the incivility that follows from it, we have to replace the current two-party system with one that does not lead to fixed dominance of extreme positions. We cannot eliminate extreme attitudes, but we can make sure that institutional forces work against them becoming established, entrenched positions. A certain amount of political instablity is a good thing. It means encouraging turnover at the top via mechanisms such as term limits, so that it takes multiple administrations to accomplish any large change in policy. Those administrations will have differing perspectives, and policy changes will thus become smoothed to something less extreme. But we don’t want governments that are instantly responsive to the whims of the latest political fad, either. In this context, the organizational agility that is so important for corporations is actually a defect when it appears in governmental functions, and the intransigence to change of the “deep state” is a feature.
The “championship playoffs” system of selecting candidates needs to change from the current single-elimination tournament to one that rewards consistency instead of streaks of wins. In sports, this is accomplished with round-robin tournaments where every team plays against every other team at least once. In elections, this can be accomplished by preference voting systems such as ranked choice, where voters compare every candidate against every other one. These systems have been shown to elect candidates with more moderate positions than “first past the post” elections.
With preference voting, candidates can’t win by targeting any particular opponent, they need to present themselves as the best of them all. In this kind of election, it’s much more effective then to focus on your good qualities than to dilute campaign resources with multiple attacks on each of the many others contending for the same office. There are many different ways to do preference voting; none of them are perfect and they can still be abused [Note 3], but nearly every other system is better than the current one.
Fixing internet incivility
To fix incivility on the internet, we need to deal with the economic incentives that keep platform providers from implementing measures that control incivility. This means introducing more effective ways to monetize customer satisfaction than engagement measures that reward dark user interface patterns that lead to addictive and antisocial behaviors. It means driving advertising back to the minor role it had in the print world. There seem to be three paths to doing this: subscriptions with aggregation, patronage, and micropayments.
The astoundingly low cost of packaging and distributing content via the internet has made it possible for creators to produce their work in ever smaller chunks. For hundreds of years, the smallest viable chunk of content was the book. Then in the 18th and 19th centuries high-speed presses reduced the cost per copy enough to permit the appearance of magazines and newspapers. Television and radio productions originally were so expensive that it required national networks of stations to aggregate a large enough audience to make soap operas economically viable. Now, while it is still possible for producers to spend hundreds of millions of dollars on a summer blockbuster, and for publishers to spend millions on the promotion of a bestselling novel, platforms like the one that delivers this essay to you have reduced the cost of production to a miniscule level, and unfortunately reduced the revenue per copy to a correspondingly low amount. As the cost of production approaches zero, the cost of collecting a payment comes to dominate the per copy media economics. This cost includes not only the dollar price of the individual experience, but the less tangible transaction friction of paying for it.
In its original form, advertising was the most friction-free consumer experience possible. It was just there on the page, and you could look at it or not. Then it began to be abused. TV commercials were inserted at cliff-hanger moments in stories, so that you had to watch them before experiencing the dramatic climax. Ads on webpages began to popup or popunder the content windows that triggered them, so that you had to click on them to make them go away, thus providing measurable evidence of at least minimal attention, and full-screen ads took over your screen completely. Malicious ads appeared that installed software on your PC without permission in order to track your viewing more comprehensively and force you to read ads regardless of your intentions or even steal information from your hard drive or spy on your keystrokes as you typed in credit card numbers. People began to use ad blockers in order to stifle bad behavior, and sites in response would detect ad-blockers and withhold their content as punishment. The capitalist limit of content as nothing more than a medium to deliver ads became explicit, the transaction cost of viewing blocked content went to infinity, and the system of an ad-supported web began to break down.
Paywalled sites appeared, incorporating the promise of the dealer of addictive drugs: “the first one is free”. The complex of dark patterns that sustain paywalls is too elaborate to go into here: it involves artificial monopolies, quasi-addiction, and the sunk cost fallacy, among other cognitive and economic defects. The breakdown of the paywall model is not as apparent as the breakdown of the advertising model, but it’s difficult to see how it can be sustained indefinitely.
From a publisher’s point of view, paywalls are an effort to create online the subscription model that has been successful in print media. But paywalled sites ignore the transformative power of low-overhead internet distribution that makes it possible for an individual creator to also be their own publisher. The added value historically provided by content editors, copy editors, advertising managers, and distribution and production managers has been eroded by automation until it no longer justifies the high prices needed to pay their salaries.
One solution to problems caused by automation of previously manual activities is to scale up those functions: instead of having an artisanal role in each published item, editors become architects of editing support systems that are implemented across multiple branded collections. These collections can share a single subscriber base, sharing the cost of managing payments and subscriber retention functions at a lower cost than is possible for any individual publication. Platforms such as medium.com are pioneering this model, although thir current price point is still to high for me. Ad-free Premium channels on YouTube and on some podcast studios are another step in this direction. Traditional publishing houses such as Condé Nast have an opportunity to provide an aggregated subscription that can increase the reach of polite advertising by spanning the silos of their 19 individual titles and deliver lower costs, increasing readership even more by reducing prices.
Patronage has supported the arts for longer than the 500 years since the Popes in Rome and the Medicis in Florence funded the Italian Renaissance, and wealthy supporters continue to this day. Organizing large numbers of benefactors has become a well-developed industry for institutions like museums and symphony orchestras. Public televsion and public radio in the US started out as government funded organizations, and have evolved into an audience patronage and philanthropic patronage funding model. Most scientific research is funded by patronage in the form of grants from government agencies and philanthropic institutions. But now, the internet has enabled a new kind of economy of scale in the way that patrons can connect with creators, in the form of crowdfunding systems. Instead of searching for large contributions from a few individually wealthy patrons, creators can solicit support from a large quantity of small donations using sites like Patreon, RocketHub, GoFundMe or Kickstarter.
Patronage and aggregated publishing subscriptions can support individual creators, but they cannot break the hold of advertising on individual creations. They do not provide the self-amplifying feedback loop that enables the emergence of bestsellers or hits or viral items. To get to that level of cultural impact without the help of ads requires micropayments.
The idea of micropayments has been around for as long as the idea of electronic commerce, and many micropayment systems have been tried, and they have all failed. It turns out that making the consumer PC secure enough to support seamless payment transactions is far more difficult than micropayment visionaries expected. Simply securing the user’s endpoint device is not enough, the entire ecosystem of device, application, content creation backend, and payment infrastructure has to be secured end to end, and implemented with enough diversity in each architectural component to allow the free market to optimize costs to a low enough level that the payment process is profitable for all participants. This is a tall order, and we’re not there yet. The tendency of free markets towards monopolistic overpricing means that the existing concentrated financial system may permit payment settlement systems to keep their prices so high that any emerging micropayment system will not survive long enough to grow to profitable scale.
Even micropayments may not be enough to counteract the forces of incivility and dishonesty in individual online items. Advertisements are only the vanguard of an engagement process whose goal is to get the customer to commit funds to a seller in return for some value. In most online transactions, that value is not provided until after the payment has been committed, that is, the customer is buying a promise of some value rather than the fully delivered value. This gives the seller an opportunity to hedge or even break that promise when it’s too late for the customer to do anything about it other than feel exploited.
Moving the payment transaction from before delivery to afterwards transforms the nature of the payment from a promise of some future value to an award for good work. This significantly reduces the opportunity for fraud and near-fraud and alters the balance of trust between buyer and seller to favor the buyer. When a producer has a high risk of non-payment by an unsatisfied customer, the producer has much less motivation to trade quality for cheap production and deliver work that customers will find unsatisfactory.
Brick and mortar stores can be successful with payment before delivery because they actually provide partial delivery while the buyer is in the store examining the product before purchasing it. Examination ensures that the promise of performance upon delivery after payment is significantly more trustworthy. When trustworthiness of product claims can’t be effectively addressed, like with used cars, government regulations such as lemon laws that create post-delivery recovery options will arise. Most things sold online don’t have a government assurance of quality so online sellers have had to devise creative alternatives. For example, Amazon provides a kind of pre-delivery quality assurance with its customer reviews, and this provides it with a competitive advantage over online retailers whose review systems are not as sophisticated, or which may not provide reviews at all.
The economic structure of social media transactions is much more complex than a simple marketplace of things to buy, but social media already provide the seeds of award payments. Nearly every one of them already has “like” or “thumbs up” or “clap” or “upvote” buttons that can add the user’s vote to raise or lower the prominence of an item. Economists like to say that demand is not valued by how passionately a customer feels the need for a product, but by their “revealed preference” when they actually give up their hard money in a payment transaction. “Likes” are about the weakest form of preference revelation one can imagine; suppose that a tap on a like button actually came with tangible award — a one cent or ten cent micropayment, or more.
Award micropayments could transform the economics of content creation. Rather than being dependent on creating large numbers of pieces in order to provide a large number of ad opportunities, the lowered barriers to entry enabled by award micropayments would open up opportunities for less productive creators to produce small numbers of pieces and still obtain enough income to justify their efforts.
With award micropayments, “likes” for uncivil, offensive statements could be penalized without blocking or deleting them completely, by making them expensive on an exponential sliding scale. The mobbing behavior seen on platforms like Twitter could be moderated by pricing algorithms in the same way that collisions in broadcast networks like classic ethernet automatically resolve themselves with an exponential backoff algorithm, in order to prevent network congestion.
Fixing social incivility
Incivility in the offline world has existed since time immemorial; it takes a better historian than me to summarize its roots. But we often forget that only a few hundred years ago, cities were so dangerous that everyone carried swords or knives, and what are now considered minor offenses such as pickpocketing were punishable by hanging. Wikipedia traces organized police forces back to at least the Roman Empire two thousand years ago. Despite the fearful conditions in bad parts of town today, things were much worse in those days.
It’s hard to identify broad causes for the improvement in civility over history, but my impression, without specific study, is that widespread wealth and education are good candidates for overarching drivers.
To my psychologizing mind, street incivility is a proactive ego-maintenance response to situations with higher likelihood of violation of cultural norms for smooth interactions. Just like a jazz musician altering his line in response to a missed note, a person can respond to a social error by moves that indicate “I meant to do that all along.” When conditions are crowded and people from many different cultures become thrown together into a single community, even minor social mishaps become more common, and uncivil responses can proliferate.
Notoriously uncivil cities like New York suffer from all of these problems. It would be interesting to know how this pattern holds for other cities around the world. Tokyo, for example is large and crowded, but has a high degree of cultural uniformity. London seems to have gotten less civil on its way to being the most diverse city in the world. Houston and Los Angeles have a high degree of diversity, but may escape widespread incivility by being low density.
So increases in ability to adapt to a variety of cultural norms and reductions in the frequency of violation of those norms via reduced crowding and better education should lead to more civility. Better education leads to more wealth, which leads to lower rates of population growth and greater ability to avoid unsettling public interactions. Broadly distributed wealth and cultural homogeneity seem to have led to more civility for the Scandanavian countries.
Can wealth and education become widespread enough to counteract the demographic tendencies towards urban crowding and inter-group friction that lead to growing incivility? With the amplifying relationship between wealth and power that leads to ever more concentrated wealth, it may take some reworking of capitalism to get there.
Notes
[Note 1] It was a PDP-10 from Digital Equipment Corporation running a heavily customized version of the TOPS-10 operating system.
[Note 2] Gun advocates like to say that “an armed society is a polite society”. The phenomenon of road rage belies this slogan. Even in moderately armed cities like Houston, there’s a recognizable pattern of shooting incidents where someone gets into an argument at a club, goes home, gets their gun, returns to the club and shoots at the person they were arguing with. A belligerent drunk with a gun is still a belligerent drunk, he does not magically become a polite drunk just because there are others around him who are armed.
The 1942 award winning science fiction novel “Beyond this Horizon” dramatizes this situation: everyone is armed except conscientious objectors who announce their philosophy by wearing armbands (“brassards”) at all times, which gives them immunity to armed methods of conflict resolution at the cost of reduced social status. (In this society, as in the real-life ’40s, womens’ status is already minimal, they don’t need guns to maintain their low position.) Dueling is common, and shootouts in crowded restaurants are so routine that diners don’t even bother to take cover. Awkward pauses in social visits to friends in government offices are eased by casual quick-draw practice against targets on the office walls; coworkers don’t seem to be distracted by the sounds of random gunshots. An entrepreneur would welcome the opportunity to sell crossfire injury insurance in such a society, but the rest of us would be appalled at a realization of this vision.
[Note 3] “Arrow’s Impossibility Theorem” mathematically shows how every voting system has pathological situations where the wrong candidate will be elected.
