Глава 1
The Great Firewall's Hidden Power: Distraction and Diversion
When Chinese internet users try to access Google, they often face a frustrating experience-sometimes the site loads perfectly, other times it's frustratingly slow, and occasionally it doesn't load at all. This isn't a technical glitch; it's a deliberate strategy. Welcome to China's sophisticated censorship system, where the government doesn't simply ban information-it taxes it through inconvenience. Margaret Roberts' groundbreaking book "Censored: Distraction and Diversion Inside China's Great Firewall" reveals how this system works with stunning clarity. The book has become required reading in political science departments worldwide and was awarded the 2019 Goldsmith Book Prize from Harvard Kennedy School for its exceptional contribution to our understanding of media, politics and public policy. Even tech leaders like Mark Zuckerberg reportedly keep it on their nightstands as they navigate the complex relationship between information platforms and government control. What makes Roberts' analysis so compelling is that it challenges our fundamental assumptions about how censorship functions in the digital age-and raises uncomfortable questions about information control in democracies as well.
Глава 2
The Paradox of Porous Censorship
Why would authoritarian regimes invest billions in censorship systems that can be easily circumvented? This central puzzle drives Roberts' investigation. China's Great Firewall can be bypassed with VPNs, which millions of Chinese citizens use daily. Keyword filtering can be evaded by using similar words, images, or creative workarounds like mixing characters and emojis. Website throttling makes sites load slowly rather than blocking them entirely, creating frustration without outright denial. This pattern extends beyond China-Iran systematically reduces internet speeds during elections and protests, Russia floods opposition hashtags with bot-generated content and coordinated disinformation campaigns, and historically, British censors banned translated French literature while allowing the original versions, knowing few could read French.
The answer lies in understanding how information actually flows through society. Censorship doesn't need to be airtight to be effective-it just needs to create different barriers for different populations. By requiring citizens to spend extra time, money, or technical expertise accessing certain information, governments effectively filter out casual users while allowing determined citizens through. Most people, being politically disinterested and overwhelmed by daily information, are easily diverted by small access costs. Studies show that even a 3-second delay in page loading can cause users to abandon their search, while requiring an extra click reduces engagement by up to 50%.
Roberts identifies three mechanisms of censorship: fear (threats of punishment), friction (increasing costs of accessing information), and flooding (coordinated distraction through propaganda). While fear is problematic because it must be observable to work and can cause backlash, as seen in Hong Kong's protests against security laws, friction and flooding have proven more effective in the digital age. These subtle methods don't need to be observable and can be easily explained away as technical glitches, server issues, or algorithmic quirks. For example, when Twitter-like platform Weibo slows during sensitive political events, it's often attributed to high traffic rather than deliberate intervention.
These mechanisms drive a wedge between elites who can circumvent censorship and the masses who cannot, preventing the coordination between the core and periphery necessary for collective action. Tech-savvy users with resources can use VPNs, mirror sites, and encryption tools, while average citizens face just enough inconvenience to discourage deeper investigation. By making sensitive information slightly harder to find rather than impossible to access, the government creates a system where most citizens simply don't bother looking. This "friction-based censorship" proves remarkably effective - studies show that when Chinese social media platforms implement even minor delays for sensitive topics, discussion of these topics drops by over 80%.
The genius of porous censorship lies in its plausible deniability and cost-effectiveness. Rather than maintaining perfect control, it achieves practical control through subtle manipulation of user behavior and information flows, creating what Roberts calls a "graduated system of information control" that segments the population without obvious repression.
Глава 3
Beyond Fear: The Subtlety of Modern Censorship
Traditional views of censorship focus on fear-the threat of punishment for accessing forbidden information. However, Roberts demonstrates that fear is actually the least effective mechanism in the digital age. When users encounter obvious censorship online, they often react with anger or curiosity rather than fear. Through empirical studies of both content producers and consumers who experience censorship, Roberts found that censorship signals don't deter engagement with sensitive topics-they may actually increase it.
In a matched comparison study using Weiboscope data, Roberts compared pairs of users who posted identical content on the same day, but where only one user's post was censored. Surprisingly, censored users were more likely to persist in discussing sensitive topics rather than self-censoring. They wrote posts more similar to their censored content, used sensitive words at similar or higher rates, and were twice as likely to complain about censorship.
One user named Zhang, whose post advocating for decreased internet censorship was ironically censored, responded by tripling his posts about the internet in the following days. He viewed censorship as evidence of corruption, writing "It's because the corrupted officials are worried that Weibo will spill all their (negative) secrets."
A broader survey of Chinese internet users confirms these findings-only 5% reported they would feel fearful after experiencing censorship, while 23% would feel angry. This explains why the Chinese government tries to make censorship less observable-visible censorship can backfire by increasing interest in sensitive topics while damaging public support for information control policies.
Instead of relying on fear, modern censorship operates through more subtle means-making certain information slightly harder to access (friction) or drowning it out with competing narratives (flooding). These methods don't require citizens to be aware they're being censored, and they scale efficiently across millions of internet users.
Глава 4
The Power of Small Inconveniences
If fear doesn't effectively control information flow, what does? The answer lies in the remarkable power of small inconveniences. Content filtering significantly impacts information spread in China through multiple mechanisms. While determined users could theoretically overcome censorship barriers, these small costs add up to create a multiplying effect on information accessibility.
The slightly increased effort required to find censored information dramatically reduces how many people encounter it. This pattern is clearly demonstrated in the case of Tibetan self-immolation protests. Between 2011-2013, 120 Tibetans self-immolated within China, representing a major political challenge to the government. Analysis shows that self-immolations occurring on weekends-when fewer censors are working-generated significantly longer bursts of online discussion compared to weekday events.
The Great Firewall provides the most striking example of friction's effectiveness. Despite being relatively easy to circumvent using VPNs, only 5% of urban Chinese internet users reported having jumped the Firewall. Even among those who knew circumvention was possible, 48% didn't understand what "jumping the Firewall" meant.
Notably, fear isn't the primary deterrent. When asked why they don't circumvent censorship, only 2% cited fear and 9% mentioned legal concerns. Instead, 45% said they had no reason to jump the Firewall, 15% didn't know how, and 14% found it too bothersome. These small frictions, rather than fear or deterrence, keep most Chinese netizens from accessing blocked information.
Those who do circumvent the Great Firewall represent a distinct demographic-highly educated, politically engaged citizens with sufficient technical skills and resources. These users are typically younger (under 35), with 75% having college education compared to just 25% among non-evaders. They have significantly higher incomes (averaging 141,800 or $22,839 annually) compared to those who know about but don't jump the Firewall (87,100 or $14,029).
This small, politically sophisticated group is effectively separated from the general public by the Firewall's friction, preventing potential political mobilization without requiring direct repression.
Глава 5
The Divided Internet: Elites vs. Masses
Beyond survey data, behavioral evidence from Twitter provides compelling confirmation of the Great Firewall's effectiveness in segregating Chinese internet users. By analyzing geo-located tweets from mainland China (which requires VPN use since Twitter is blocked), researchers estimated that only about 0.02% of Chinese internet users post on Twitter daily. This stark contrast becomes even more apparent when comparing to Hong Kong, where Twitter isn't blocked and approximately 1.7% of internet users post daily. This comparison suggests China has only 1-2% of the Twitter users it would naturally have without Firewall restrictions, representing a massive reduction in international social media engagement.
Content analysis reveals profound differences between Chinese Twitter users and those on the domestic platform Sina Weibo. Twitter users from mainland China frequently discuss sensitive political topics, using terms like "peace," "freedom," and "democracy," and even mention controversial activists like Ilham Tohti and Liu Xiaobo. They also demonstrate a distinctly international outlook, regularly using English words, technology-related terms, and discussing global affairs. Their posting patterns suggest they are predominantly urban, educated, and professionally successful - China's digital elite.
In contrast, Weibo users exhibit markedly different behavior, remaining starkly apolitical in their discussions. Their content centers primarily on entertainment, everyday feelings, personal relationships, and celebrities. Popular topics include TV shows, food recommendations, relationship advice, and lifestyle content. This dramatic content difference illustrates how the Firewall effectively creates two distinct online populations: a small, politically engaged group with access to global platforms, and the vast majority limited to domestic sites.
The Great Firewall's effectiveness stems not from being impenetrable - VPNs can easily circumvent it - but from creating just enough friction to discourage casual users. When websites are blocked, Chinese traffic to them drops dramatically, often by 90% or more. Three notable case studies demonstrate this effect:
1. Google's experience proves particularly instructive. The government employed a strategy of gradual throttling beginning in 2010, making the site increasingly frustrating to use before fully blocking it in 2014. This caused traffic to decline from 10% of Google's global traffic to 5% before the final block reduced it below 1%.
2. Wikipedia faced intermittent blocks throughout the 2010s before being completely blocked in 2015, causing Chinese-language contributions to plummet by over 75%.
3. Instagram's sudden block during the 2014 Hong Kong protests demonstrated how quickly the government could sever access to platforms sharing sensitive content, with Chinese user engagement dropping from millions to thousands overnight.
The government's strategic approach to blocking varies by platform and circumstance. The gradual throttling approach used with Google proved particularly effective, as it created uncertainty about whether poor performance stemmed from censorship or technical issues, thereby minimizing public backlash. This sophisticated approach to digital control has created a two-tier internet system that effectively contains political discourse while maintaining the appearance of open access.
Глава 6
When Censorship Fails: Crisis and Motivation
While friction-based censorship is remarkably effective under normal circumstances, it has significant limitations. Two key factors determine when users will evade censorship: the value of the blocked information and how suddenly censorship is applied. This dynamic creates a delicate balance between control and potential backlash that authorities must carefully manage.
Crisis events significantly complicate the government's ability to effectively use friction-based censorship. During crises, citizens become more politically aware and motivated to seek information regardless of cost. This heightened awareness stems from immediate concerns about safety, economic impact, and social stability. Sudden censorship during these periods alerts people to what the government is trying to hide, creating the Streisand effect where attempts to suppress information increase interest in it. The more aggressive the censorship, the more citizens suspect important information is being withheld.
The Tianjin chemical explosion of August 12, 2015 provides compelling evidence of this phenomenon. The massive industrial accident in the port city created a fireball visible from space and raised serious concerns about public safety and environmental contamination. Survey data collected before and after the explosion showed that only 5% of internet users in Tianjin reported using VPNs before the disaster, compared to nearly 30% afterward. Being interviewed after the explosion predicted a 22 percentage point increase in the probability of using circumvention tools, even controlling for demographic and political characteristics. Citizens actively sought alternative information sources when official channels proved insufficient or untrustworthy.
Similarly, the Instagram block during Hong Kong protests not only drove VPN adoption but also increased signups for previously blocked platforms like Twitter and Facebook. The sudden restriction of a popular social media platform created a ripple effect of digital migration. Users who continued using Instagram through VPNs tended to be more active users before the block, suggesting their established habits motivated them to seek workarounds, inadvertently exposing them to other censored information. This demonstrated how censorship can backfire by pushing users toward even more politically sensitive platforms and communities.
During these sensitive periods when friction becomes less effective, the government turns to flooding to manage rather than stall conversations. This strategy shift acknowledges that complete information control becomes nearly impossible during crises, leading authorities to focus instead on shaping narratives through overwhelming volume rather than restriction. The approach includes deploying massive numbers of posts, comments, and articles that promote official viewpoints while drowning out alternative perspectives.
The limitations of friction-based censorship during crises reveal a fundamental weakness in information control systems: when people are sufficiently motivated by immediate concerns, they will find ways around technical barriers. This pattern holds true across different types of crises, from natural disasters to political upheavals, suggesting it represents a persistent challenge for censorship regimes.
Глава 7
Flooding: The Art of Distraction
When friction fails, flooding becomes the government's preferred strategy. When a 6.5-magnitude earthquake struck Yunnan province in August 2014, killing hundreds and injuring thousands, the Chinese government deployed a sophisticated information control strategy. Eight hours after the disaster, state media orchestrated a coordinated campaign about Internet celebrity Guo Meimei's confession to gambling and prostitution-a three-year-old scandal suddenly revived. Multiple major outlets simultaneously published detailed accounts following government directives to "prominently display Xinhua and CCTV coverage of Guo Meimei, and to actively organize and direct commentary."
The difference between friction and flooding can be explained through a tax analogy: governments can either directly tax gas-guzzling cars (friction) or subsidize fuel-efficient ones (flooding) to achieve the same effect of making inefficient cars relatively more expensive. Similarly, flooding makes competing information cheaper and off-limits information relatively more expensive to access.
China's propaganda system is structured to facilitate coordination at every level. The Propaganda Department, one of the most extensive bureaucracies within the CCP, manages messaging directly from the top levels of the party through every news medium. The government controls personnel in all major media organizations and requires government certification for journalists.
To identify newspaper propaganda in China, researchers combined leaked propaganda directives with detection of coordination across government newspapers. Since the Chinese government uses directives to facilitate coordination, they reverse engineered propaganda by identifying days when newspapers published identical articles. Analysis of 111,789 articles from 25 provincial newspapers in 2012 showed that coordination peaked during the power transition from Hu Jintao to Xi Jinping and other sensitive events like Bo Xilai's wife's trial.
Online, contrary to popular belief that China's "Fifty Cent Party" online commentators primarily argue with government critics, analysis of leaked emails reveals their main strategy is distraction through cheerleading. Rather than engaging in argumentative defense of policies or attacking critics, 85% of posts focused on patriotism, encouragement, inspirational quotes, or celebrations of Chinese culture and historical figures.
This coordinated flooding significantly influences what appears in social media. When coordination increases, government-sanctioned information spreads more widely online, either through sheer volume or because netizens unwittingly reshare it. This shifts the ratio of government to citizen information, making people more likely to encounter and share propaganda rather than alternative viewpoints.
Глава 8
Implications for Democracy in the Digital Age
The digital revolution has fundamentally transformed the landscape of free speech, blurring traditional distinctions between democratic and authoritarian information control. While historical analysis focused primarily on overt censorship and the basic right to speak publicly, today's challenges center on the more nuanced question of visibility and reach in an overwhelmingly crowded digital space. Democratic societies, which once distinguished themselves through broad speech protections, now grapple with new forms of information control that mirror some aspects of authoritarian systems.
The nature of censorship has evolved to encompass sophisticated mechanisms of content prioritization and presentation. Search algorithms, controlled by major tech corporations like Google, Facebook, and Twitter, wield unprecedented influence over electoral information flow and public discourse. Government bureaucracies across political systems increasingly shape the information landscape through selective data releases, classification decisions, and regulatory frameworks. Meanwhile, well-funded interest groups and political organizations exploit digital platforms to flood communication channels with targeted messaging, effectively drowning out competing voices through sheer volume rather than direct suppression.
The crisis in traditional journalism has exacerbated these challenges. As advertising revenues decline and newsrooms shrink, the vacuum is increasingly filled by politically motivated content producers. Local news outlets, particularly vulnerable to financial pressures, are being replaced by partisan networks and social media echo chambers. This transformation demands a fundamental reimagining of free speech protections for the digital age, including:
• Mandatory transparency requirements for search and recommendation algorithms
• Anti-monopoly measures to ensure diversity in digital platforms
• Standardized protocols for government data release and accessibility
• New regulatory frameworks protecting digital information rights
• Support mechanisms for independent journalism
The long-term implications of modern censorship extend far beyond immediate information access. In China's case, the Great Firewall has created a parallel internet ecosystem, with companies like Sina Weibo, Baidu, and WeChat thriving in the absence of international competition. However, this protection comes at a significant cost - Chinese tech companies must maintain extensive censorship operations, often employing thousands of content moderators. Startups face additional hurdles, requiring VPN services to access global development tools and platforms, increasing operational costs and limiting innovation.
The ideological consequences are particularly concerning for global democracy. Information asymmetry between elites (who often have means to circumvent restrictions) and ordinary citizens creates dangerous social tensions. In China, this manifests as a growing disconnect between internationally educated urban professionals and the broader population. Moreover, censorship increasingly impedes cross-cultural understanding between Chinese and Western societies, limiting opportunities for diplomatic engagement and mutual comprehension. This cultural isolation risks exacerbating international tensions and complicating efforts at global cooperation on critical issues like climate change, public health, and economic stability.
These challenges suggest that defending democracy in the digital age requires new approaches that go beyond traditional free speech protections, focusing instead on ensuring meaningful access to diverse information sources and protecting the integrity of public discourse in an increasingly complex information environment.
Глава 9
The Future of Information Control
As China's censorship model gains traction globally, there is a pressing need for comprehensive research examining how fear, friction, and flooding mechanisms operate across diverse authoritarian environments. The specific implementation of these control methods varies significantly based on three key factors: the regime's political structure (from single-party systems to military dictatorships), economic development level (which affects internet penetration and digital literacy), and technological capabilities (including surveillance infrastructure and content monitoring systems). Understanding these predictors would enable better anticipation of how different regimes might adapt and implement censorship strategies.
The remarkable resilience of authoritarian regimes derives from a sophisticated combination of traditional repression, responsive governance, and increasingly subtle tools that impede collective action coordination. Modern censorship methods emphasizing friction and flooding don't rely on brute force or complete information blockades. Instead, they create a carefully engineered environment that subtly steers citizens away from activist networks and alternative viewpoints through psychological barriers and information overload. These techniques prove particularly effective in digital spaces where attention is increasingly fragmented.
This strategy succeeds because citizens behave as strategic information consumers in an environment of overwhelming data abundance. Faced with limited time and cognitive resources, people naturally seek efficient signals of information importance and reliability. Unlike direct censorship or violent repression which can backfire by drawing attention to sensitive issues, throttling access and flooding channels with distracting content effectively guides citizens toward state-approved narratives without obvious coercion.
Aldous Huxley's prescient observations in Brave New World become increasingly relevant - the most effective form of control comes not through force but through the manipulation of desires and attention. Modern censorship operates primarily through sophisticated information taxes that alter incentive structures, making it cognitively and emotionally costly to seek out alternative viewpoints while making state-approved content more readily accessible and appealing.
The implications of this evolution in information control extend far beyond China's borders. In an era where information abundance rather than scarcity presents the primary challenge, the ability to direct attention through subtle psychological nudges rather than outright bans becomes increasingly powerful. This shift affects both authoritarian regimes and democracies, as those who control the friction and flooding of information increasingly shape public discourse and social reality. The battle for control no longer centers on restricting access to information but on influencing what information people are likely to encounter, engage with, and ultimately believe. This creates new challenges for preserving democratic discourse and independent thought in both free and controlled societies.
These developments raise crucial questions about the future of information freedom, democratic deliberation, and the role of technology in shaping public consciousness. As these subtle control mechanisms become more sophisticated and widespread, understanding and countering them becomes essential for preserving genuine public discourse and democratic values.
