Research: Wrong Narratives About Internet Culture

For Book 3: Lurk More

Topic 1: “All Trolling Is Harassment”

What the Common Narrative Says

Trolling is synonymous with online harassment, cyberbullying, and hate speech.

What Actually Happened

Judith Donath’s 1999 paper “Identity and Deception in the Virtual Community” defined trolling as “a game about identity deception” — posting provocative content to test community boundaries and reveal assumptions (MIT Media Lab). The original trolling taxonomy includes:

• Classic trolling: Posting bait to see who responds sincerely (fishing metaphor — “trolling” as in dragging bait through water)
• Concern trolling: Pretending to sympathize while actually undermining
• Sealioning: Politely demanding endless evidence
• Griefing: Disrupting shared spaces for entertainment

Whitney Phillips (This Is Why We Can’t Have Nice Things, MIT Press 2015) documented how media coverage collapsed all forms of trolling into “harassment,” losing critical distinctions (OpenLibrary). Her 2019 follow-up “It Wasn’t Just the Trolls” argued that media amplification of trolling was itself a form of the problem — journalists gave trolls oxygen while condemning them (Data & Society).

The conflation matters because: (1) it makes genuine harassment harder to identify when everything is “trolling,” (2) it erases the tradition of productive provocation (Socratic method, satire, culture jamming), (3) it gives institutions a weapon to dismiss any uncomfortable speech as “trolling.”

Named Sources

• Judith Donath — “Identity and Deception in the Virtual Community” (1999)
• Whitney Phillips — This Is Why We Can’t Have Nice Things (2015), “It Wasn’t Just the Trolls” (2019), “The Oxygen of Amplification” (Data & Society, 2018)
• Stephanie Ortiz — “Trolling as a Collective Form of Harassment” (Social Media + Society, 2020)

Topic 2: “4chan Is a Radicalization Pipeline”

What the Common Narrative Says

4chan is a breeding ground for extremism, a radicalization pipeline that turns normal people into white supremacists.

What Actually Happened

4chan hosts over 70 boards covering topics from cooking (/ck/) to origami (/po/) to music production (/mu/). The radicalization narrative focuses almost exclusively on /pol/ (Politically Incorrect), which is one board among many.

Colley & Moore (“The challenges of studying 4chan and the Alt-Right,” New Media & Society, 2020/2022) found: “We can neither confirm nor deny a direct link between the online boards of 4chan, radicalisation, and real-world violence. There is an empirical gap between 4chan and political violence, and researchers lack evidence about the development of actual relationships between them.”

The academic literature on 4chan is overwhelmingly about one board (/pol/), treated as representative of seventy-plus boards.

Actual radicalization pathways are more complex:

• The SPLC’s “McInnes, Molyneux, and 4chan” found multiple common paths — including 4chan AND exposure to white nationalist YouTube figures. Pathways are plural and cross-platform.
• Thorleifsson’s 2022 paper “From cyberfascism to terrorism” (Nations and Nationalism) focuses specifically on /pol/, not 4chan as a whole.
• Researchers note that responsibility isn’t exclusive to 4chan — radicalization moved across multiple platforms, with media attention, political actors, and monetized content ecosystems all contributing.

The Monolith Problem: Treating 4chan as a monolith produces bad policy. A person discussing Warhammer 40K on /tg/ is not on a radicalization pathway. 4chan’s own communities frequently despise /pol/. The monolith narrative directs attention away from actual mechanisms of radicalization (algorithmic recommendation, political media ecosystems, real-world grievances) toward a bogeyman.

Named Sources

• Colley & Moore — “The challenges of studying 4chan and the Alt-Right” (New Media & Society, 2020/2022)
• SPLC — “McInnes, Molyneux, and 4chan: Investigating pathways to the alt-right”
• Thorleifsson — “From cyberfascism to terrorism” (Nations and Nationalism, 2022)

Topic 3: “The Internet Made People Worse”

What the Common Narrative Says

The internet, and social media in particular, made people meaner, more hostile, more isolated, and more tribal. Before the internet, people were more civil.

What Actually Happened

Social Isolation Was Already Increasing: Robert Putnam’s Bowling Alone (2000) documented the decline of American social capital — civic engagement, community organizations, social trust — beginning in the 1950s and accelerating through the 1970s-1990s (Simon & Schuster). This preceded Facebook (2004), Twitter (2006), and smartphones. Putnam pointed to television, suburban sprawl, changing work patterns, and economic inequality as drivers. Steve Bannon publicly stated that when strategizing Trump’s election, they studied Bowling Alone because “when people are socially isolated, they become vulnerable to populist appeals” (Fortune).

Mean World Syndrome: Visibility, Not Creation: George Gerbner coined “mean world syndrome” in the 1970s. He found a direct correlation between television consumption and the belief that the world is dangerous. Crime rates dropped in the 1990s, but on-screen violence didn’t — a 2015 Stetson University study found on-screen murders spiked precisely as real-life murders declined (Journal of Communication). The internet extends this dynamic: social media amplifies hostile content through algorithmic engagement optimization, making existing hostility more visible, not creating new hostility from nothing.

Online Disinhibition: More Complex Than “People Are Meaner”: John Suler’s 2004 paper “The Online Disinhibition Effect” identified six factors (CyberPsychology & Behavior). Critically, Suler identified TWO types of disinhibition: toxic (trolling, aggression, hate speech) AND benign (greater emotional expression, openness, self-disclosure). The narrative focuses exclusively on the toxic form. Research also suggests that lack of eye contact, not anonymity per se, is “the chief contributor to the negative effects of online disinhibition.” Some findings indicate “people might actually be more polite to strangers in anonymous contexts than to their friends.”

The Counterevidence: A literature review in General Hospital Psychiatry found peer-support interventions online decreased depression symptoms. A randomized controlled trial in PLOS One found internet support groups reduced depression at 6- and 12-month follow-ups. Fan communities, creative writing communities (SCP Foundation), open-source communities, and mutual aid networks demonstrate genuine connection, creativity, and support.

Why the Difference Matters

The “internet made people worse” narrative reverses causation. Social isolation, political polarization, and declining trust were increasing for decades before social media. The internet made these existing trends visible and, in some cases, accelerated them through algorithmic amplification. But it also created new forms of connection, support, and creative collaboration that didn’t exist before.

Named Sources

• Robert Putnam — Bowling Alone (2000)
• George Gerbner — cultivation theory and mean world syndrome (1970s)
• John Suler — “The Online Disinhibition Effect” (CyberPsychology & Behavior, 2004)

Topic 4: “Anonymity Is Inherently Dangerous”

What the Common Narrative Says

Anonymous users are the source of online toxicity. Real names produce accountability and civility. If people had to stand behind their words, the internet would be better.

What Actually Happened

Facebook’s Real-Name Policy: Zuckerberg asserted it “helps keep people safe.” Facebook, the world’s largest real-name platform, has been one of the internet’s most significant vectors for misinformation, political manipulation, and organized harassment. Research shows “real names are part of mechanisms that drive misinformation in settings marked by affective polarization.” The policy disproportionately harmed trans people, abuse survivors, and Native Americans.

South Korea: The Definitive Natural Experiment: South Korea implemented an online real-name verification law in 2007 for sites with 100,000+ daily visitors. Results:

• Malicious comments dropped by only 0.9 percentage points (13.9% to 13.0%)
• In August 2012, the Korean Constitutional Court unanimously ruled the policy unconstitutional, finding “evidence presented was insufficient to show a decrease in hateful comments” (The Korea Herald)
• The policy created treasure troves of identity data for hackers

The Disqus Data: Pseudonymity Wins: 61% of pseudonymous comments were positive, vs. 34% anonymous and 51% real-name. Pseudonymous commenters produced higher-quality content.

Huffington Post Study (2013-2015): Analyzed 45 million comments across transitions from easy anonymity to registered pseudonyms to Facebook-only. “Comments made under conditions of durable pseudonymity were rated by other users as having the highest quality.” The key insight: “What matters is not so much whether you are commenting anonymously, but whether you are invested in your persona and accountable for its behavior in that particular forum.”

Historical Pseudonymity: Hamilton, Madison, and Jay wrote the Federalist Papers as “Publius.” Thomas Paine published Common Sense pseudonymously. Mark Twain, George Eliot, Lewis Carroll — all pseudonymous.

Something Awful’s $10 Paywall: Charged $10 for accounts — not to verify identity but to create accountability through investment. Combined with active moderation, produced “one of the more controlled message boards.” For roughly five years (2000-2005), SA’s pseudonymous community “operated the engine that made internet culture go.”

Why the Difference Matters

The evidence consistently shows that the variable producing quality discourse is not identity verification but investment and structural accountability. Pseudonymous communities with good moderation and barriers to re-entry consistently outperform both anonymous free-for-alls AND real-name platforms. Worse, real-name requirements actively endanger dissidents, abuse survivors, and marginalized people whose safety depends on controlling their identifiable information.

Topic 5: “Social Media Caused Political Polarization”

What the Common Narrative Says

Social media, through filter bubbles and algorithmic amplification, is the primary driver of American political polarization.

What Actually Happened

Boxell, Gentzkow, and Shapiro (PNAS, 2017): The single most important study. Using nine measures of polarization from the American National Election Study:

• “The growth in polarization in recent years is largest for the demographic groups least likely to use the internet and social media.” (PNAS)
• Eight of nine measures show greater increases for those older than 75 than for those aged 18-39.
• Less than 20% of 65+ used social media in 2012, vs. 80% of 18-29. Yet the elderly polarized more.
• Conclusion: “These facts argue against the hypothesis that the internet is a primary driver of rising political polarization.”

Cable News: The Pre-Social-Media Polarization Driver: Removing Fox News during the 2000 election “would have reduced the average county’s Republican vote share by 1.6 percentage points” (Quarterly Journal of Economics). Cable news stations “gained popularity because of an increasingly divisive political atmosphere, and these political divisions fueled the demand for partisan news.”

The Filter Bubble Thesis: Largely Debunked: “Despite Pariser’s founding anecdote there is a severe lack of convincing evidence for the existence of significant filter bubbles in search results.” Studies show “substantial overlaps in the search results seen by users of different political interests” and “the diversity of most citizens’ media uses.”

What Social Media Actually Did: Accelerated outrage cycles, created incentive structures rewarding extreme content, enabled coordination of collective action, made existing divisions more emotionally salient. But blaming social media lets the actual drivers off the hook: cable news, partisan media ecosystems, geographic self-sorting, economic inequality, and structural political factors.

Named Sources

• Boxell, Gentzkow, Shapiro — PNAS, 2017
• Eli Pariser — The Filter Bubble (2011) — and subsequent debunking studies
• Nolan McCarty — Princeton, gerrymandering research

Topic 6: “The Dark Web Is Mostly Criminal”

What the Common Narrative Says

The dark web is a lawless criminal marketplace. Tor is a tool for drug dealers, human traffickers, and terrorists.

What Actually Happened

What Tor Actually Is: Developed in the mid-1990s by United States Naval Research Laboratory employees to protect American intelligence communications. DARPA funding 1997. Alpha launched September 20, 2002. As of 2012, 80% of the Tor Project’s $2M annual budget came from the US government (State Department, Broadcasting Board of Governors, NSF) (Wikipedia). The irony: the tool politicians condemn was developed and funded by the US military/government.

Actual Usage Statistics:

• Only ~6.7% of Tor users connect to Hidden Services (the “dark web”). The other 93.3% use Tor for regular browsing with privacy.
• The dark web “comprises only 3% of the traffic in the Tor network.”
• In “not free” countries, possible malicious use drops to 4.8%. In “free” countries it rises to 7.8% — suggesting Tor’s primary use in authoritarian contexts is circumventing censorship.
• During Hong Kong 2019 protests, Tor usage increased by 3,000%.

Legitimate Users: SecureDrop (designed by Aaron Swartz and Kevin Poulsen) runs as a Tor onion service, used by 35+ news organizations including NYT, WaPo, The Guardian, ProPublica, The Intercept (SecureDrop). Journalists, activists, abuse survivors, people in authoritarian countries, privacy-conscious individuals represent the vast majority of Tor’s user base.

The Silk Road Distortion: Media coverage dominated by Silk Road (2011-2013), treated as representative of all Tor usage. The real influx of Silk Road users came after Gawker published an in-depth story (June 1, 2011) — the very coverage that sensationalized it also drove growth (Wikipedia). One marketplace, a fraction of Tor activity, became the public face of technology used by millions for legitimate purposes.

Why the Difference Matters

Demonizing Tor threatens journalists, whistleblowers, and dissidents worldwide. Policy proposals to weaken Tor wouldn’t meaningfully affect criminal networks (which would migrate to alternatives) but would devastate legitimate users. The narrative serves authoritarian governments that want to eliminate anonymous communication.

Topic 7: Section 230 — “The Law That Broke the Internet”

What Section 230 Actually Says

The key 26 words: “No provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider.” (Legal Information Institute)

The Origin: Stratton Oakmont v. Prodigy (1995) — Jordan Belfort’s firm (yes, Wolf of Wall Street) sued Prodigy for $200M. Judge ruled because Prodigy moderated, it was liable for ALL content (Wikipedia). Perverse result: the way to avoid liability was zero moderation. Section 230 (1996) fixed this.

What Politicians Claim:

• Republicans: Claim it requires platforms to be “neutral.” It doesn’t. Section 230 explicitly protects moderation in good faith.
• Democrats: Claim it prevents accountability for harmful content. It doesn’t override federal criminal law.

As Techdirt noted: “One side is furious that platforms can moderate, while the other side is furious that platforms don’t have to moderate.” Both want it gone “for contradictory reasons.”

If Repealed: MIT professor Sinan Aral: “one of two things will happen. Either platforms will decide they don’t want to moderate anything, or platforms will moderate everything.” Wikipedia would face existential liability. CCIA estimated repeal cost: $2.2 trillion+. Comment sections, review platforms, forums, and any user-generated content service would face elimination of user participation or burial in litigation.

Named Sources

• Jeff Kosseff — The Twenty-Six Words That Created the Internet (2019)
• Stratton Oakmont v. Prodigy (1995)
• Bipartisan Policy Center, Brookings, FIRE analyses

Topic 8: “Hackers Are Criminals”

What the Common Narrative Says

Hackers are criminals who break into computer systems to steal data, cause damage, and commit fraud. “Hacker” means “cybercriminal.”

What Actually Happened

The Original Meaning: Steven Levy’s Hackers: Heroes of the Computer Revolution (1984) documented the original hacker ethic from MIT’s Tech Model Railroad Club (late 1950s) (Project Gutenberg). Core principles: access to computers should be unlimited; all information should be free; mistrust authority, promote decentralization; hackers should be judged by their hacking, not “bogus criteria such as degrees, age, race or position”; you can create art and beauty on a computer.

The Security Researcher Reality: HackerOne paid out $81M in bug bounty rewards in a single year (BleepingComputer). The DOD runs bug bounty programs. Bug bounties have “averted north of $20 billion in hack damage.” White hat hackers have submitted 29,000+ vulnerability reports to government, with 70%+ determined valid.

DEF CON / Black Hat: DEF CON (1993, Jeff Moss) and Black Hat (1997) form “Hacker Summer Camp” (Wikipedia). Federal agencies including NSA and DHS attend and recruit. The same people media call “hackers” in a criminal context are the people defending critical infrastructure.

The CFAA: A Law of Absurd Breadth: “Prohibits intentionally accessing a computer without authorization or in excess of authorization, but fails to define what ‘without authorization’ means.” First-time offenses: up to 5 years. Other violations: 10, 20, or life.

Aaron Swartz: Co-founder of Reddit, creator of RSS, co-designer of SecureDrop. Charged under CFAA for downloading 4.8M academic articles from JSTOR through MIT’s network, using an account he had legitimate access to as a Harvard research fellow. Despite JSTOR declining to pursue charges, federal prosecutors charged 13 felony counts (35 years, $1M fines). Died by suicide January 2013 at 26 (Wikipedia). Aaron’s Law proposed but never passed.

Van Buren v. United States (2021): Supreme Court narrowed CFAA in 6-3 ruling — “exceeds authorized access” means accessing off-limits areas, not accessing permitted information for unauthorized purposes (Supreme Court). EFF called it “a victory for all Internet users” (Electronic Frontier Foundation). But CFAA remains a blunt instrument.

Why the Difference Matters

The equation of “hacker” with “criminal” justifies prosecutorial overreach (Swartz case), chills security research protecting critical infrastructure, and treats curiosity and exploration as inherently suspicious. The distinction between curiosity-driven exploration and malicious intrusion is the difference between the people who find vulnerabilities and those who exploit them — treating them identically makes everyone less safe.

Cross-Cutting Themes

1. The truth is always more complex than the narrative: Every topic has a simple version that’s wrong and a complex version that’s inconvenient.
2. Platform architecture shapes behavior: Reddit, 4chan, Twitter, Facebook each create specific dynamics — the platforms are environments that select for certain behaviors.
3. Censorship creates legitimacy for the censored: Every attempt to suppress discussion made it more intense.
4. The media consistently fails to cover internet culture accurately: Pre-existing frames forced onto complex, multi-faction events.
5. The digital native perspective is structurally different: People who grew up online understand multiple motivations coexist in the same spaces. Institutional observers flatten this complexity.

Key Sources Summary

Source URLs

• Angela Nagle, Kill All Normies (2017)
• Dale Beran, It Came from Something Awful (2019)
• Gabriella Coleman, Hacker, Hoaxer, Whistleblower, Spy (2014)
• Whitney Phillips, This Is Why We Can’t Have Nice Things (2015)
• Evgeny Morozov, The Net Delusion (2011)