The Room Around the Argument: Audience Sorting at Outside the Beltway, 2003-2026

Posted on August 18, 2026 by Luke Ford

In a companion study I coded claims across twenty years of Outside the Beltway and asked whether heated assertions carried evidence, qualification, counterargument, concession or correction nearby. They increasingly do. Steven L. Taylor and James Joyner write with more sourcing and more visible updating than they did in 2007. That study could not see who was still present to test any of it.

This one looks at the room. The site turns out to have kept a record of its own sorting, most of it written by the proprietors.

My own exit belongs in the evidence. I stopped reading in 2023. I did not leave because I wanted friendlier coverage of Donald Trump (b. 1946); I pay over a hundred dollars a month for subscriptions to the Los Angeles Times, the New York Times, the Washington Post, the Financial Times, the WSJ, Apple News Plus, TNR, NYRB, TNY, and company, none of them tender toward him. So the conclusion was never the problem.

I need value to keep reading, and OTB stopped offering value.

I get it. One man’s value is another man’s burden.

What can be settled is whether the room changed. It did, and the site said so.

In 2012 Joyner published Where Have the Thoughtful Conservatives Gone?, answering a longtime commenter who argued that Joyner’s own drift and the changing comment section had driven older conservatives off. Joyner rejected parts of the account and accepted the demographic core of it: most of the better contributors now sat to his left, and thoughtful conservatives had thinned.

The post also contains Joyner’s account of why blogging changed him. He describes having his arguments constantly challenged, by commenters and by other bloggers, and having to defend what he wrote. He describes the hunt for material forcing him to read people unlike himself. He calls himself and Taylor a rough controlled experiment. That is close to the reciprocal process I argued for in the first study, professional habit migrating into the online voice, and it carries a condition Joyner does not state: it works while the challengers are there.

The thread then produces the best piece of analysis in either essay, and it comes from a commenter. Rodney Dill argued that numerical imbalance drives out the measured voices on the minority side first. The conservatives most likely to remain would be the ones willing to antagonize a hostile majority. Another commenter, Drew, said a conservative could go on reading the essays while concluding that commenting had become a fool’s errand.

Others in the same thread disputed the diagnosis. Russell said he came as a liberal specifically to hear thoughtful conservatives and avoid an echo chamber. KariQ said she read OTB to challenge her own Democratic reflexes. Anjin-san mocked the idea that liberals had hunted reasonable conservatives to extinction. In 2012 the site was still doing the thing that made it valuable, and its liberal readers could name what that was.

Seven years later, in Where Have All the Commenters Gone?, Joyner went into the site’s database and published comment totals and first and last dates for vanished regulars. Eighteen persona records carry usable numbers. They sum to something on the order of seventy-five thousand comments, with a median of about 3,250 per persona and a median observed tenure above six years. Every figure is a floor, since several commenters used multiple addresses, and the records are personas rather than people; one man appears twice under two known handles.

The list does not show conservative flight. HarvardLaw92, stonetools, anjin-san and mantis argued from the left, and six records end in 2017 alone, three of them within one week in June, which baffled Joyner at the time. Blogs shed regulars the way neighborhood restaurants do. One commenter had died. One had a plausible professional reason for silence. Ordinary churn is the larger process.

The ideological finding sits beside the data, in Joyner’s own voice. He reports a steady decline in the conservative commentariat. He gives the reasons without being asked: he had moved left, Taylor had moved the same way, the Republican party had changed, and conservative commenters came to feel besieged from both directions. He then describes what remained on the right as consisting, with rare exceptions, of trolls or people hard to distinguish from trolls.

Set that beside Rodney Dill’s 2012 prediction. Reasonable minority voices leave first; the combative ones stay; the majority comes to experience the surviving minority as unreasonable. The prediction and the observed endpoint were written seven years apart by people who were not collaborating.

In January 2023 Joyner published OTB at 20. Early OTB, he writes, ran on mostly Republican and conservative commenters arguing with similarly inclined front-pagers, with some Democrats and progressives in the mix. Twenty years on, the commentariat was almost entirely the latter. He attributes this partly to the hosts becoming less conservative and mostly to the national landscape.

That is the proprietor describing his own community, so it cannot be dismissed as a hostile reader’s impression. It should be dated carefully. The post marks the moment Joyner said the transformation out loud, not the moment it happened; his own evidence puts most of the movement between 2012 and 2019. My exit in 2023 coincides with the announcement.

Joyner says the comment community had become the part of the site that mattered most to him, more than traffic figures, and that OTB might run on largely for the forums alone.

The governance of the room tightened in datable steps.

2012 is lament: Joyner notices the thinning and discusses it in public. March 2019 is management: Feeding The Trolls Only Makes Them Stronger records repeated efforts to remove disruptive participants and urges regulars to stop rewarding them, and Joyner says he had finally concluded that two long-running commenters had to go, with a third, present almost from the site’s founding, close behind. November 7, 2024, the day after Trump’s second win, brings Taylor’s Don’t be a Troll, welcoming disagreement and setting three rules: no personal attacks on post authors, engagement with the post, and show your work. Joyner adds his own the next day in And Another Thing!, asking commenters to drop the nicknames for the president-elect. July 12, 2025 brings the Comment Policy Change: mandatory registration, so that bans stick against men who return under new names.

Every participation cost is still asymmetric in practice. Registration is nothing to a daily regular and can be the last nudge for a man already wondering why he bothers. Being asked to show your work is trivial when the room shares your premises and expensive when ten replies are waiting for the premise itself. A symmetrical rule can have an asymmetrical incidence.

The complaint I would make appears inside the site, made by named people, on dates. In 2019 john430 says the blog has grown darker, angrier and ugly. Eric Florack calls it venomously left-wing. Guarneri lists one-sidedness and faux outrage and reduces the front page to Trump around the clock. In 2025 Connor says the site was better a decade ago and now treats dissent as trolling.

That series is weaker than it looks. In the 2025 thread the hosts and regulars identify Connor as a returning banned participant previously posting as Drew and Guarneri, among others. If they are right, a thirteen-year run of independent complaints collapses partly into one persistent man under several names, and I have to discount my own best-looking evidence.

What survives the discount is strong. Rodney Dill’s 2012 mechanism stands on its own. Matt Bernius supplies an internal history in 2025: OTB drew left-leaning readers because expert front-pagers offered intelligent argument to their right, and the complaints of leftward drift grew through the Obama years and accelerated with Trump. And Joyner’s three descriptions, in 2012, 2019 and 2023, are worth more than any number of angry conservatives, because they are the house’s own account.

The parties agree that the composition changed. They disagree about why, and that disagreement is legitimate.

The most useful witness in the whole record is the man who did not leave.

In the July 2025 thread a commenter called Andy, who has read and commented for more than two decades and describes himself as ideologically eclectic, explains how he survives. He does not respond to certain people. He self-censors on topics where experience tells him engagement will cost half a day and thousands of words. He writes unusually long comments because short ones get mischaracterized, so he front-loads qualifications, which he calls throat-clearing. He says that as a newcomer rather than a twenty-year regular he could see himself being chased off.

That reinterprets my first study. I scored evidence, qualification and counterargument as epistemic regulation. In Andy’s case the same surface features come from somewhere else. The hedge is armor. The counterargument is preemption of the five people who would otherwise say he ignored it. The prose would score well, and the score would be measuring social cost.

Research on the spiral of silence has found for decades that people express political views less readily when they read the room as hostile, and studies of social media report the same asymmetry. Andy goes on writing, at greater length, about fewer things.

In late July 2025, two years after I stopped reading, Taylor began the series Meanwhile in North Korea, introducing an early installment as a new ongoing bit. By August 2026 it had passed twenty entries. In one of them Taylor states the purpose: the category demonstrates ways Trump and his allies engage in authoritarian symbolism and behavior.

The proposition is intelligible comparative politics, and it falls inside Taylor’s field. Personalist regimes do run on leader imagery, ritual praise and naming practices. The measurement problem lies elsewhere.

The title has already made the comparison. Suppose Taylor links a proposal to name a highway after Trump, supplies the reporting, and observes that it shows the president’s hold on his party. The claim can be sourced to the last comma. Nothing in the source establishes that North Korea is the proportionate frame. The category supplies that before the post begins, and every future entry inherits it.

This is the mirror of a problem I met in the first study. There, a disclaimer at the top of a long post threatened to launder unsupported claims downward. A rubric launders a frame sideways, across everything that will ever be filed under it. Claim-level coding sees neither.

The form matters too. I sampled seven spaced installments where Taylor’s own prose could be separated from embedded material and quotation. The median ran about 66 words. Four were at or under that, two under thirty, and the launch post was an embed with almost no authorial argument. Others in the series are long and analytical, so this is not a claim about all of them.

It is enough to correct the first study’s selection finding. I argued that the micro-post died at OTB, killed off by the 2018 removal of Quick Picks and by Joyner’s rule against posting without several paragraphs to contribute. The micro-post came back. A branded category lowers the threshold again, because a small observation no longer has to earn a standalone essay; it can be another specimen in a collection. A writer raises his publication threshold, then builds a rubric under which subthreshold items become publishable.

Claim regulation asks whether the assertion is sourced, whether uncertainty is marked, whether contrary evidence is conceded, whether errors get corrected. Taylor and Joyner do well on it, better than they did as young bloggers.

Frame regulation asks why this event is being read through this analogy, what would count against the analogy, and whether the frame distinguishes leader worship from ordinary sycophancy, personalist presidentialism from democratic backsliding, backsliding from authoritarianism, authoritarianism from totalitarianism.

Portfolio regulation asks which facts become posts at all, how much of the output organizes around one man, how often a theme recurs against its alternatives, and what kind of fact never gets written up.

A publication can be immaculate at the first level and thin at the other two. Every sentence sourced, every source accurate, every prediction later audited, and the whole thing monomaniacal to anyone who has not been in the room the whole time. When I called late OTB unhinged, that was a reaction, not a finding. These three levels are what I should have said, and they can be tested.

Taylor answers his critics at length. His reply to Connor in July 2025 runs longer than most of his posts: PEPFAR, immigration enforcement, due process, threats to citizenship, media settlements, the military, tariffs, each with reasons, ending with an invitation to explain why his concerns are wrong and a hope that Connor or some reader might change his mind.

Moderation may have preserved argument rather than suppressed it. Homogeneity does not establish error; a room of oncologists agrees about smoking. Forced heterogeneity may not help either, since exposure to opposing views online has been found to harden partisans rather than soften them. And a week after the registration change Joyner published Debating to Lose, criticizing debaters who made CPAC arguments to a hostile audience, under the line that preaching to the choir works only if the choir is in the room. He understands that argument is audience-dependent. The question is whether any writer can see his own choir.

Elias Aboujaoude’s model in Virtually You describes what the medium does to a person. A long-lived community changes people by selection instead.

Nobody has to become more extreme. The moderate dissenter leaves. The combative one stays. The agreeable reader becomes a regular, the ambivalent one a lurker, the surviving dissenter an armored writer who skips subjects and spends eight hundred words on what used to take a hundred and fifty. The authors then live in the room those choices produced. Approval arrives faster than contradiction. Analogies become shorthand. Arguments need less rebuilding because the audience remembers them. A writer can grow more disciplined sentence by sentence while losing track of how the total looks to someone who did not go through the same sorting.

An internalized opponent is not a live one. The author picks which objection to raise, in what form, at what length, and decides when it has been met. The real opponent answers the wrong question, attacks the premise instead of the evidence, insists that the marginal point is central, and is sometimes irritating because he is right. Joyner’s own 2012 account of what changed him names commenters and rival bloggers, not imagined interlocutors.

So the audit trail stays real while fewer people remain who want to audit it.

I have counted almost nothing myself. The 2019 totals are Joyner’s, unauditable, and are floors rather than counts. I did not assign ideologies to departed commenters, since retrospective coding from scattered comments would add judgment without adding information; the conservative decline rests on the hosts’ own statements. The moderation chronology shows increasing formalization and demonstrates no ideological discrimination. The rubric sample is seven purposive entries, enough to show the micro-post returned and not enough to characterize the series. Where the site’s participants link one commenter to earlier handles, I have used that to discount my own witnesses rather than to establish anything about him.

Two measurements would settle much of this and remain undone. A count of unique commenters per year would show the population contracting and being replaced. A content analysis separating posts aimed at Trump as a person from posts aimed at policies, institutions and other actors would test whether the writing changed or only the room. Neither should be invented, and both want a second coder.

The largest missing population cannot be reconstructed at all. Lurkers leave no trace. They are the people most relevant to sorting and the least likely to be recorded, which means the commentariat is not a sample of the readership. It is a survivor list.

I left in 2023 and nothing happened. No argument, no ban, no farewell comment. The database has no departure date for me, because I never commented. I don’t comment almost anywhere. I went on paying to read publications that reach the same conclusions about the same president.

What I stopped wanting was their particular contempt of people like me. A reader can be told he is wrong. He can read a writer who despises the politicians he likes. He can take sharp language and hard conclusions, and he can pay for them by the month.

What gets harder to sit with is a room where he can no longer imagine what evidence, argument or event would produce surprise. At some point in 2023, the operators of OTB stopped offering me anything I didn’t already know.

Posted in Blogging | Comments Off on The Room Around the Argument: Audience Sorting at Outside the Beltway, 2003-2026

Steven Taylor, James Joyner and the E-Personality: Aboujaoude’s Virtually You Tested on Twenty Years of Outside the Beltway

Posted on August 18, 2026 by Luke Ford

Elias Aboujaoude (b. 1971) published Virtually You: The Dangerous Powers of the E-Personality in 2011, when blogging still mattered, Twitter still imposed a 140-character limit, and the psychological consequences of living online were new enough to seem strange. The architecture of online life, he argued, encourages a distinctive e-personality: more assertive, less restrained, more impulsive, more certain of what it knows. The danger he emphasized was permeability. Habits learned online migrate into the rest of the personality.

He was right about several of the routes. He worried that blogging frees a writer from the patient construction of an argument, and that Twitter compresses serious matters into a form hostile to sophisticated exchange. He tied impulsivity to a decline in what he called reflection, meaning the collection and evaluation of information before reaching a conclusion.

The model has a gap. If the border between the online and offline self is porous, traffic should run in both directions. Habits acquired through professional scholarship may migrate down into the e-personality while impulsivity, certainty and aggression migrate up out of it.

A personal note. After Trump returned to politics in 2023, I can’t read the blog anymore. The unhinged hatred for one side of the political spectrum (Trump and his supporters) is not my cup of tea. I don’t see any value in the blog’s obsession with chronicling how America under Trump is becoming like North Korea. To me, this is the unhinged e-personality.

Three things.

First, the following study cannot see that North Korea series. The anchor samples run May 2007, May 2013, May 2020 and May 2025. The North Korea rubric starts in late July 2025 and runs at least twenty-three numbered installments through August 15, 2026, nearly all Taylor, with Joyner contributing adjacent items like “Donald J. Trump Institute of Peace” and “The Greatest President in History, Some Say.” The whole run sits outside the frame by a few weeks. So the essay’s picture of 2025 Taylor rests on a window that closed right before the most repetitive thing he does now.

Second, and worse for my method: a standing rubric does argumentative work that no individual post has to defend. This is the same defect I flag below in “The List,” running the other direction. There, a disclaimer at the top threatened to launder unsupported predictions into qualified ones. Here, a recurring title installs a comparison as a premise, and each new post inherits it without paying for it. Local brakes measure whether a hot claim carries evidence in its immediate neighborhood. They cannot detect a frame that has been amortized across two dozen posts and a category page. My measure rewards the form of argument and is blind to its proportion. A writer can attach a source to every heated sentence and still run a monomaniacal frame, and my numbers would call him disciplined.

Third, the series undercuts the selection-adaptation finding. The essay’s strongest claim about Taylor is that the micro-post died and the developed essay replaced it. A numbered, branded, image-and-caption rubric published every few weeks is the micro-post, revived with a house style. If those posts run short, the 2018 disappearance of the InstaPundit-style item looks less like a permanent change in threshold and more like a form that returned once it acquired a name.

I don’t understand how anyone who does not hate Trump with every fiber of his being can read the blog. I enjoyed the blog until some point in 2023. I spend over $100 a month subscribing to elite media that uniformly loathes Trump (the LAT, NYT, WP, FT, NYRB, TNR, TNY, etc). I am fine with reading hinged hatred of Trump. There are many good reasons to hate Trump. I don’t find any value from OTB’s unhinged hatred posing as truth-optimizing analysis.

Steven L. Taylor and James Joyner (b. 1965) started Outside the Beltway in January 2003. Joyner published thirteen posts on the first day. He later described early OTB as links, excerpts and often snarky commentary. Taylor, who had been blogging at PoliBlog, joined the OTB front page in November 2006. The site covers most of the lifespan of the political blogosphere, written by two men who built substantial professional identities outside it. Joyner became a professor of security studies at Marine Corps University’s Command and Staff College, working on national security, military affairs and civil-military relations. Taylor became a professor of political science and then dean at Troy University, publishing on parties, elections, electoral systems and comparative democratic institutions. He retired in July 2024. Joyner’s retrospective OTB at 20 supplies much of the site’s own account of its history.

That combination permits a harder question than whether two men got angrier over twenty years. I coded claims and tested four competing explanations against each other.

The first is Aboujaoude’s medium effect. Lower-friction, more interactive forms should produce more rhetorical heat and weaker epistemic regulation. The second is compensatory regulation. Experienced writers may keep making heated claims while attaching evidence, qualification, counterargument, concession or correction to them. The third is selection adaptation. Rather than learning to write every instantaneous reaction more carefully, an experienced blogger may learn not to publish many of those reactions at all. The fourth is arousal override. Political events of unusual emotional intensity may temporarily overwhelm whatever regulatory habits a writer has acquired.

The four do not receive equal support, and the two authors do not behave the same way.

For Joyner, the evidence favors a medium effect and gives partial support to compensatory regulation. For Taylor, the developmental story fails at the start. His blog writing already shows strong epistemic regulation in the earliest comparable material, especially when he writes inside political science. His long-term adaptation is selection. He publishes less of the old rapid-fire blogging and more developed analysis. Taylor also supplies the best evidence for arousal override, in the hours and days after Donald Trump’s (b. 1946) victory in November 2024.

The theoretical result may outlast the empirical one. Aboujaoude got the permeability right and the direction incomplete. At Outside the Beltway, the professional personality keeps invading the e-personality.

The best evidence in the study comes from one night in 2012.

During the Republican National Convention on August 29, Joyner watched Paul Ryan’s (b. 1970) acceptance speech and commented live on Twitter. He then embedded the stream in an OTB post called RNC Night 2: This Is More Like It. The opening line works as a manipulation check: he promises more coherent thoughts in the morning.

The Twitter stream is fast, funny and evaluative. It delivers instant judgments of speakers, jokes and sarcastic reactions. When Ryan blamed Barack Obama (b. 1961) for the closing of the General Motors plant in Janesville, Wisconsin, Joyner’s immediate response amounted to incredulity. The medium rewarded the reaction ahead of the research.

The next morning the same writer returned to the same claim about the same speech in Paul Ryan’s GM Plant: Fact Checking the Fact Checks. The structure changes. Joyner walks through the plant’s chronology, Obama’s statement and the competing fact checks. He separates the technically defensible portion of Ryan’s sentence from the implication Ryan wanted the audience to draw. He concludes that Obama cannot be blamed for the closure, and he considers stronger economic criticisms Republicans might have made instead.

Almost everything that ordinarily worries a longitudinal comparison is held close to constant here. Same writer. Same subject. Same politics. The available historical information barely changed overnight. We are not comparing young Joyner with middle-aged Joyner, Republican Joyner with post-Trump Joyner, or amateur blogger Joyner with professor Joyner. What changes is medium and about twelve hours.

The substantive conclusion survives the transition nearly intact. What changes is its epistemic architecture.

Had the Aboujaoude effect meant that Twitter changes a man’s political opinions, this pair would be weak evidence. His deeper concern ran to reflection, inhibition and the compression of complicated matters into quick reactions. The Joyner pair displays that compression and then its reversal. Twitter gets the conclusion first. The blog post gets the evidence, the distinctions and the rival interpretations.

Joyner has described the process. In a 2021 essay for War on the Rocks, Twitter Doesn’t Have to be the End of the Conversation, he contrasts the snappy, instantaneous exchange encouraged by social media with longer venues that require substantive analysis and editorial guidance. He grants that Twitter rewards sarcasm, cynicism and putdowns, and he argues that its rapid feedback can improve later thought. A year later, writing in Defense One, he made the admission personal. The platform creates incentives to perform for an online audience, and he says he fell for the trap.

One caution belongs here. Joyner knew the tweets would be archived on his own site the same night. That awareness cuts against the disinhibition reading, and it makes the observed gap between the two versions a conservative estimate rather than an inflated one. The line about more coherent thoughts in the morning also shows that the writer understood the limits of the medium while using it. Recognition of the affordance preceded the compensation.

This is support for the medium hypothesis, and it stops well short of medium determinism. Joyner’s 2021 argument explains why. Low-friction media generate useful criticism and expose an idea to people who improve it. His War on the Rocks essay began as a Twitter conversation. What shifts is the probability distribution. Faster media make an unregulated reaction easier to publish. Slower and edited forms create more chances for a brake.

What counts as a brake? I treated a claim as a heat event when it carried a strong evaluative judgment, negative person-focus, pejorative characterization, ridicule or argumentative sarcasm. Categorical certainty was coded separately. A confident empirical proposition is not by definition heated, and coding it that way would classify ordinary factual assertion as rhetorical disinhibition.

I then asked whether the heat event carried a local brake. Evidence counted. So did a qualification narrowing the claim, a serious counterargument, a concession to the other side, or a correction or update. Proximity governed. The brake had to appear in the claim, the same sentence or bullet, or the immediately adjacent sentence in a way tied to the claim.

That rule determines how to read Taylor’s November 6, 2024 post The List.

Taylor introduces nineteen predictions as his concerns and says the list came off the top of his head. Those are real epistemic signals. They tell a reader not to mistake what follows for a forecast with calibrated probabilities.

They are global signals. They cannot be sprinkled backward across nineteen separate predictions so that each one becomes locally qualified. If Taylor says a Trump administration will do X, the preface supplies no evidence for X. If he calls a possible appointee utterly unqualified, the introductory modesty does not regulate the evaluation. A prediction that contains its own “may,” its own evidentiary reference or its own competing possibility is a different case, and I coded it accordingly.

The distinction protects the measure. Without it, long-form writers get a free advantage: one disclaimer at the top of a thousand-word post converts every categorical assertion below it into a qualified claim, and the coding then measures politeness of framing rather than regulation of reasoning.

A second problem cannot be solved by any coding rule, and readers should hold it against the results throughout. A brake is observable behavior. Its motive is not. A qualification may express real uncertainty, or it may be reputational armor bolted onto a conclusion the writer never doubted. A political scientist who has spent thirty years being asked “how do you know that” acquires the reflex of anticipating the question. That reflex looks identical on the page whether it regulates the thought or merely dresses it.

Some brakes cost the author something, and those carry more weight. A concession that weakens the writer’s own conclusion, a correction that admits an earlier error, an update that abandons a preferred interpretation: these are expensive in a way that a hedging adverb is not. Where the argument below turns on costly brakes rather than cheap ones, I say so.

The frozen claim-level corpus produced 71 fully segmented claims from nine posts whose source text could be recovered in full. Fifty belong to Joyner, twenty-one to Taylor. I coded those claims from scratch under the rules above. One coder produced every number in this essay, so read the percentages as descriptions of a small sample rather than as measurements.

Joyner’s numbers move in one direction. In the 2007 material I coded eight heat events, six of them locally braked. In 2013, seven heat events and six brakes. In 2020, six heat events and six brakes. In the strict 2025 coding, two and two. The series runs 75, 86, 100 and 100 percent.

The tidiness should raise suspicion, and the final cell carries a warning. Joyner’s 2025 post on Joe Biden’s (b. 1942) cancer diagnosis ends with the judgment that the public is entitled to such information about a president. Code that ordinary normative assertion as a strong evaluative heat event and the 2025 result becomes two of three, or 67 percent. One decision moves the cell by 33 points. See the post and judge for yourself.

The larger problem is arithmetic. Twenty heat events spread across four periods cannot sustain a trend test. Comparing Joyner’s earliest cell, six of eight, against his two most recent cells combined, eight of eight, falls far short of conventional significance in any test appropriate to numbers that small. Anyone who wants to reject the compensatory-regulation finding on the numbers alone can do so, and I would not argue. The percentages are a summary of what I read. The textual pattern behind them carries the argument.

That pattern is consistent. Consider Joyner’s 2013 post on Apple’s tax arrangements. The rhetoric is lively. He calls the arrangement brilliant and clever. He then asks why Congress left the loophole open for decades, notes that multinational corporations have exploited such differences for a long time, concedes that a lower corporate rate makes intuitive sense, questions whether it would solve the problem, and labels his own consumption-tax proposal his decidedly non-expert sense of things. The post runs hot without pretending that the heat constitutes expertise.

The 2020 material is stronger. In WSJ Chastises Trump’s Slander, Joyner uses nonacademic language about Trump’s behavior, and the pejorative conclusion sits on top of a long evidentiary structure: the accusation under criticism, the factual basis for rejecting it, the Wall Street Journal editorial’s reasoning, and the institutional problem of applying Twitter’s ordinary rules to a sitting president. High temperature, high evidentiary density.

Another 2020 claim shows the expensive kind of brake. Having first read Justice Department decisions on congressional insider-trading investigations through the politicization of the department under Trump and William Barr (b. 1950), Joyner reverses himself on reflection. He gives his reasons and separates the still-open investigation of Richard Burr (b. 1955) from the closed ones. That is visible updating, and it costs him the more dramatic version of his own story.

The mature Joyner is sometimes harsher than his younger self. The change is that the harsh assertion increasingly travels with its own audit trail.

His professional baseline predicts that habit. Joyner’s 1996 International Studies Quarterly article with John Oneal and Brad Lian, Are the American People “Pretty Prudent”?, tests a prominent theory of public support for the use of force against 38 major American uses of force in foreign-policy crises from 1950 through 1988. The authors do not find the periodization they went looking for. They report the failure and identify a different relationship in the data.

His civil-military work behaves the same way. In Trump’s Generals: A Natural Experiment in Civil-Military Relations, published in Strategic Studies Quarterly, he begins with two alarming and opposed predictions about retired generals serving in the Trump administration and concludes that neither worked well. The generals produced no militarization of policy and no reliable restraint of the president. The paper ends on an unresolved question about long-term damage instead of converting uncertainty into a dramatic finding.

The edited professional essays follow suit. This is Not a Civil-Military Crisis, written for Defense One, opens by granting the core concern of the argument he opposes, that retired senior officers have acquired an outsized political role, before disputing the stronger characterization. His essay on Kenneth Waltz (1924-2013), first published in The National Interest, separates admiration for Waltz’s theoretical achievement from acceptance of the policy conclusions structural realists drew from it.

The possibility raised by all this runs opposite to the obvious one. Blogging did not train Joyner to write like an academic. Over twenty years, habits recognizable from academic and edited professional work became more visible inside the blog voice.

Taylor keeps the study honest by refusing to fit.

His early OTB writing is already too regulated to leave room for progressive improvement.

The clearest early case in the frozen sample is his May 2007 writing on Hugo Chávez (1954-2013) and the Venezuelan opposition broadcaster RCTV. Taylor makes strong judgments. He reads the episode as part of an authoritarian consolidation and connects control of television to the maintenance of political power. Institutional reasoning accompanies each judgment: RCTV’s position in Venezuelan broadcasting, the station’s conduct during the 2002 coup crisis, and an argument about why that history fails to justify the 2007 action.

In my coding, Taylor produced five heat events in the 2007 material, all five locally braked. In the complete 2020 sample I coded five heat events and four brakes. That is 100 percent followed by 80 percent. Those cells are far too small to establish a decline, and one recode would erase the difference. They are enough to reject the attractive version of the compensatory hypothesis, in which Taylor gradually learns to regulate hot claims. No low starting point exists from which such an improvement could occur.

His scholarship explains why. A Different Democracy, written with Matthew Shugart, Arend Lijphart (b. 1936) and Bernard Grofman (b. 1944), compares American political institutions with those of thirty other democracies instead of treating familiar American arrangements as the natural baseline. His Oxford Handbook chapter on the American electoral system explains the durability of the two-party system through plurality elections, single-seat districts, primaries and the Electoral College. His 2025 Protect Democracy paper, Trapped in a Two-Party System, extends the same institutional analysis and proposes open-list proportional representation as one route out. The argument carries a normative implication. Its explanatory machinery is structural.

That machinery keeps showing up on the blog. His 2020 discussion of the Oregon Republican Senate primary could have been another culture-war post about a QAnon-linked candidate. Taylor moves outward from the candidate to the incentives of a noncompetitive statewide race, weak candidate recruitment, party labels as information shortcuts, the American primary system and the effects of electoral-system design. The person is the news. The institutions are the explanation. The site’s own summary describes the post as more about structural conditions than about the Republican Party.

Taylor shows that professional regulation can walk in the door on the first day.

The second long-term change is easier to see. A writer can adapt to an impulsive environment in two ways. He can improve the impulse after expressing it. Or he can stop converting every impulse into a publication.

Early OTB was prolific by current standards. Joyner published thirteen items on January 31, 2003. Reviewing that archive ten years later in Joyner Unfair to Krugman, he wrote that the vast majority of his February 2003 output consisted of InstaPundit-style short posts with little or no analysis, unlike the longer pieces normal by 2013.

May 2007 still looks like that environment. The archive shows Taylor publishing several times on some days and Joyner producing a run of items on unrelated subjects within a few hours. Collecting fifteen eligible substantive posts, working backward from the end of the month, took six calendar days for Taylor and seven for Joyner.

By 2013 the same fixed target required 23 days for Taylor and 21 for Joyner. In 2020 it required 21 days for Taylor and eight for Joyner. In 2025 it required eleven for Taylor and eighteen for Joyner.

No smooth secular decline appears. Composition changed more than frequency. The site changed with the writers. Joyner says OTB’s May 2010 redesign moved it from a conventional reverse-chronological weblog toward a magazine format. A February 2018 redesign made the site mobile friendly and removed the Quick Picks section. After Quick Picks disappeared, Joyner says he all but stopped writing the InstaPundit-style micro-post and seldom published unless he had at least a few paragraphs to say. The 2020 rebound in his output has a separate cause. Doug Mataconis had stopped writing for the site, and Joyner says he and Taylor picked up the pace to compensate.

This is selection adaptation. The adapting unit is partly the writer and partly the institution the writers built.

The distinction improves Aboujaoude’s model. An experienced Internet user can alter the architecture around publication. He can abandon the format most likely to turn a passing reaction into a permanent item. A website can delete the box that invited one-line links. A writer can adopt a personal rule that a thought earns a post only when it supports several paragraphs.

The Internet creates affordances for impulsivity. Experienced users remove affordances.

The same behavior extends to the comment section. On November 7, 2024, the day after “The List,” Taylor published Don’t be a Troll, announcing that he was tired of commenters who insult rather than argue and setting out three rules: no personal attacks on post authors, engagement with the post, and show your work if you want a debate. The next day Joyner seconded him in And Another Thing!, asking commenters to drop the nicknames for the president-elect on the grounds that they signal unseriousness and lower the tenor of the conversation without advancing any argument.

Two men who had just written some of their hottest prose in twenty years turned within forty-eight hours to rebuilding the rules of the space they write in. Aboujaoude’s model has a place for the disinhibited reaction and a place for its migration into ordinary life. It has no obvious place for the moderator who responds to his own disinhibition by rewriting the house rules.

The claim about rising post length has to stay modest. The frozen corpus did not preserve reliable author-only word counts separating Joyner’s and Taylor’s prose from the often extensive material they quote. Reporting per-1,000-word heat rates or a median-length series on that basis would be false precision. The evidence establishes a categorical shift away from micro-posting far more securely than any continuous increase in author-written length.

In 2003, by Joyner’s account, most posts were short weblog entries. By 2023 he seldom bothered to post without a few paragraphs to contribute. The disappearance of a unit of production is a behavioral change.

If Taylor is the hard case for increasing regulation, he is a good case for arousal override.

“The List” appeared the day after Trump’s victory. Its nineteen predictions range across the civil service, the Justice Department, Ukraine, Israel, tariffs, appointments, NATO, public health, the Federal Reserve, immigration enforcement, the press and the international monetary order. Taylor is writing after an outcome he considers dangerous.

The post shows why global and local qualification have to be separated. “These are my concerns” brakes the document. “Off the top of my head” brakes the document. Many of the predictions then arrive as bare propositions. The format discourages development. Little evidence attaches to individual items, competing routes to the same outcome go untreated, and confidence varies little across nineteen predictions of unequal plausibility.

Some of that is defensible on Taylor’s own terms, since he says the purpose is to put down a marker for later discussion. Putting down a marker and locally regulating each prediction are different acts.

The comment threads of those days run hotter than the posts. Taylor reports deleting a comment and says he is tired of one commenter’s sneering. Minutes after, weighing another commenter’s severe geopolitical forecast, he calls the possibility frighteningly real and reaches for the ending of Planet of the Apes.

This is the strongest Taylor material for Aboujaoude. Immediate, interactive, emotionally charged. The political scientist’s habit of decomposing a proposition into institutional parts has not vanished, and for a few exchanges it recedes.

One confounder gets its own sentence. Taylor retired from Troy University in July 2024, four months before he wrote “The List.” The daily scaffolding of professional life, teaching, colleagues, administration, ended shortly before the episode this study reads as arousal override. Anyone testing the reciprocal-migration model on a larger sample should treat retirement, career change and institutional exit as predictors rather than biography.

Even here the result refuses to simplify. The next day Taylor published What Explains a Trump Win?. He opens by correcting the emerging landslide narrative with the popular-vote margin, says a firm explanation must wait for more complete data, describes his account as initial thoughts, and works through party identification, incumbency, inflation and retrospective voting. He remains morally alarmed. The alarm does not prevent analytical qualification.

Four days after that, in “One Weird Trick” Election Takes, he is arguing against monocausal explanations of Kamala Harris’s (b. 1964) defeat. He goes after the pundit’s temptation to assume that some favored messaging adjustment would have flipped the result, and he insists on multifactor explanations that no campaign fully controls. Professional training reasserts itself inside the same high-arousal week.

Arousal override survives in a narrow form. High arousal can weaken local regulation for a period measured in hours or days.

The January 2026 reassessment of “The List” strengthens that reading. Taylor returns to the predictions and audits them one by one. Some were borne out. Some require qualification. On Schedule F he records that Trump did not implement the proposal he predicted, then separates that failed prediction from other civil-service actions he considers damaging. The rhetoric stays severe. The updating is the expensive kind.

The comments show the other face of interactivity. A reader catches Taylor writing FEC where he means FCC. Taylor’s reply is immediate: correct, and thanks for the note. The section that amplified heat in November 2024 works in January 2026 as miniature post-publication review.

Interactivity can disinhibit the author or correct him. Which one depends on the week.

The medium effect performs best. Joyner’s August 2012 convention pair comes close to a within-person experiment. The live Twitter version is faster, more sarcastic, less developed. The next-morning blog version is sourced, qualified and argumentative. His later reflections identify the same affordances from the inside. No twenty-year change in ideology or personality explains a difference that appears overnight.

Compensatory regulation is supported for Joyner and fails as a general developmental law. The proportion of his heat events receiving local brakes rises from 75 percent in 2007 to 86 percent in 2013 to 100 percent in 2020, with 2025 at either 100 or 67 percent depending on one borderline call. Those cells are too small to carry the claim by themselves, and I have said so. Taylor does not reproduce the pattern. He begins at 100 percent and sits near 80 percent in 2020.

Selection adaptation receives substantial support in one form and none in another. Both writers slow sharply from the 2007 burst rate by 2013, then rebound irregularly, and Joyner’s 2020 rebound has a staffing explanation. The threshold for publication is what changed decisively. The 2010 move toward a magazine format and the 2018 removal of Quick Picks accompany the disappearance of the tiny link-and-comment post, and Joyner says he now rarely publishes without several paragraphs to contribute.

Arousal override receives the narrowest support. Taylor on November 6, 2024 is the case. Global qualification survives while local regulation weakens and comment interaction raises the temperature. The effect is temporary and uneven; his November 7 and November 11 posts are heavily regulated in the same emotional aftermath. Joyner’s convention pair points the same way. Change the medium from live Twitter to next-day blogging and regulation returns while the event holds still.

One prediction failed. The professional-habits account implies that each man should be best regulated inside his field and worst outside it. The first half holds. Taylor’s institutional writing is where academic habit is easiest to see in blog prose, and Joyner’s national-security work is most regulated when it passes through Defense One, War on the Rocks or Strategic Studies Quarterly, which is partly a genre effect, since an edited essay supplies external brakes as well as internal ones.

Joyner’s Apple tax post sits far outside his field, and his response is to surround his judgment with reporting and expert discussion and then label his own proposal nonexpert. His 2025 Biden piece leans on the medical expertise of Ezekiel Emanuel (b. 1957), gives reasons for treating Emanuel as credible, and marks his own inferences about what Biden knew with “may well” and “more likely.” Taylor’s 2025 post on fabricated citations in the MAHA report is hot and sarcastic on a subject far from electoral systems, and the factual case comes almost entirely through NOTUS reporting; when he speculates that artificial intelligence generated the fake references, he calls it his first assumption rather than a finding, and he later concedes a point before turning it against its proponent.

Distance from expertise may trigger compensation among experienced writers: more conspicuous source dependence, stronger self-labeling of inference. The matched claims are too few to estimate that interaction. They are enough to make the simple alternative, that people get sloppier whenever they leave their field, untenable here.

Combining the two men into an OTB average would bury the central result. Joyner fits a developmental account, entering from the first-wave blogosphere and acquiring, over two decades, a combination of rhetorical freedom and epistemic restraint, along with an explicit consciousness of what the medium does to him. Taylor arrives with the professional habits already installed, which creates the ceiling, and adapts instead by changing what he publishes, with episodic breakdown under high arousal.

The pilot intuition survives only after being split in two. Joyner supplies the compensatory-regulation evidence. Taylor supplies the early-carryover, selection and arousal evidence. Both supply medium-effect evidence. None of it supports the proposition that either man became more extreme because he spent two decades online.

The simple version of Virtually You predicts contamination. We enter a low-friction environment, become more impulsive, grandiose, hostile and certain, and those traits colonize the offline person. Aboujaoude describes the border as porous and worries about online behavior bleeding into ordinary life.

Nothing here disproves that. The Joyner Twitter comparison is good evidence for it. Taylor’s election-week comments are another instance. The old OTB archive preserves a medium in which a professional scholar could move from an observation to public prose in minutes.

OTB also shows a reciprocal process the theory needs.

People do not arrive online empty. Taylor brings comparative politics with him. The e-personality complains about Trump; the political scientist keeps asking which electoral structures, party incentives and institutional rules produced the event. Joyner brings international relations and security studies. The blogger renders a sharp judgment; the researcher arrives a paragraph later asking what the source says, whether an alternative interpretation survives, and whether the first reaction needs revision.

Professional institutions leave residue. Peer review teaches that claims need support. Editors teach that someone else will test the sentence. Scholarship teaches that contrary evidence is an object of analysis. Long exposure to those practices builds regulation that operates when no editor is present.

Joyner’s 2021 War on the Rocks essay nearly describes the loop. A Twitter conversation generates an idea. The idea becomes a blog post or a longer essay. Editors and reviewers strengthen the later version. That version returns to social media, where criticism starts again. He points to a hierarchy of forms. Information travels up and down it.

Taylor supplies a parallel case. His 2025 Protect Democracy paper is a polished professional version of arguments OTB readers had been reading for years. The day after publication he promoted it on OTB, and his recoverable Bluesky post does little more than point readers to the paper. The fast medium serves as the distribution network for the slow one.

The recoverable Bluesky evidence is too thin to support a symmetrical platform comparison. Joyner’s profile identifies the professional subjects one would expect; Taylor’s identifies comparative democratic institutions and OTB. Individual Joyner posts were not recoverable in sufficient number to build a responsible longitudinal series. Absence from search indexing is not absence of behavior, and treating it as evidence would repeat the error this study is designed to measure.

So the e-personality can infect professional life with speed, certainty, performance and aggression. Professional life can infect the e-personality with evidence, qualification, counterargument, revision and the learned suspicion that a first reaction may be wrong. Sometimes the decisive act of self-regulation happens before either personality writes a word, when the writer decides the thought does not yet deserve a post.

They learned the Internet. The Internet learned them.

The longitudinal frame used four comparable OTB periods, anchored in May 2007, May 2013, May 2020 and May 2025. I did not manufacture a 2003-06 Taylor series. Taylor joined the front page in late 2006 and the old PoliBlog record is incomplete. Joyner’s 2003-06 material establishes the early-blogging ecology and cannot support a symmetrical two-author estimate.

The frozen sampling frame held fifteen eligible substantive public-affairs posts per author in each anchor year, 120 in all. Eligibility required substantive author-written analysis rather than a bare link, a housekeeping item or an open thread. Working backward from May 31 gave a reproducible selection rule and prevented later outcome coding from determining which posts entered the sample.

Claims were the substantive unit. A claim is the smallest proposition that can independently be true, false, plausible, implausible, justified or unjustified. One sentence often contains two. Quoted material was not coded as the author’s claim unless he endorsed it, and rhetorical questions were coded only when the intended proposition was unambiguous. I coded categorical certainty, negative person-focus, pejorative characterization, ridicule or argumentative sarcasm, strong evaluative judgment, evidence, qualification, counterargument, concession, and correction or updating. Heat comprised the four evaluative and person-directed features. A local brake required evidence, qualification, counterargument, concession or correction inside the claim or in immediately adjacent material governing it.

“The List” served as the calibration case. Its introductory language was coded as global epistemic framing and was not propagated into the nineteen predictions. The rule makes the test harder on long-form writing and prevents a disclaimer from laundering unsupported claims into regulated ones. A secondary context-adjusted analysis, in which a governing global qualification regulates claims within its stated scope, remains available to anyone who wants to run the numbers the other way; it raises Taylor’s 2024 figures and does not change the comparison between authors.

Nine of twelve frozen segmentation-check posts were fully recoverable. One was partially recoverable and two were blocked. I did not reconstruct missing claims from titles, snippets or expectation. The numerical audit therefore rests on the 71 claims in the nine recovered posts. Supplemental cases, including the 2012 Twitter and blog pair and the 2024 Taylor sequence, were used for within-person leverage and hypothesis testing and were kept out of the frozen percentages.

The original plan called for rhetorical-feature rates per 1,000 author-written words, which remains the right denominator for comparing raw frequencies across posts of radically different length. Reliable author-only word counts were not available across the frozen corpus, in large part because OTB posts carry long quoted blocks. I have not manufactured those rates. The principal statistic, local brakes divided by heat events, sidesteps the mechanical problem that longer posts contain more opportunities for both heat and regulation. Longitudinal estimates of median author-written post length remain unfinished measurement rather than a result.

The selection test rests on three measures of unequal strength. The calendar span needed to accumulate fifteen eligible posts gives a reproducible frequency indicator. The archive documents the disappearance of the micro-post. Joyner’s retrospective supplies direct author testimony that the threshold changed. What cannot be claimed is a continuous estimate such as a percentage rise in median length.

Expertise distance was classified as core, adjacent or distant, assigned separately from rhetorical coding so the same judgment did not drive both variables. Taylor’s core covers parties, elections, representation, comparative politics, Latin American politics and democratic institutions. Joyner’s covers national security, international relations, military affairs, civil-military relations and professional military education.

The arousal test stays modest. The preregistered definition treats claims within seven days of a federal general election as event-proximate, which keeps the predictor from being defined by the emotional language it is supposed to predict. The 2024 Taylor material is a revealing qualitative case, and the complete claim corpus is too small to estimate a stable author-by-arousal coefficient. I treat arousal override as supported by episodes rather than as a population parameter.

Several falsification tests carry more weight than the raw percentages. The medium hypothesis would weaken if Joyner’s Twitter and next-day OTB versions were equally regulated; they are not. A universal compensatory-regulation hypothesis predicts a Taylor rise from early to late periods; the early ceiling contradicts it. A simple selection hypothesis predicts monotonic decline in publication rate; the rebounds contradict it. A deterministic arousal hypothesis predicts poorly regulated Taylor writing across the whole post-election period; November 7 and November 11 contradict it.

The academic baseline is not a generic category of offline writing. For claims about elections and democratic institutions, Taylor’s A Different Democracy, his Oxford Handbook chapter on the American electoral system and his Protect Democracy paper supply the relevant professional habits. For national security and military questions, Joyner’s International Studies Quarterly article with Oneal and Lian, his Strategic Studies Quarterly piece on Trump’s generals and his edited civil-military essays supply the corresponding baseline. Matching subject matter this way separates a loss of professional regulation from an ordinary difference in topic.

The largest limitation is reliability. I am one coder. The blind-coder packet is built to permit independent replication, and no independent human has performed it, so no inter-coder reliability statistic exists to report. A real test requires another person, blind to these conclusions, to segment and code the designated material and then compute agreement and a statistic such as Krippendorff’s alpha. The sensitivity of Joyner’s 2025 cell to one borderline strong-evaluation decision shows the cost of its absence. A second coder is the single improvement that would do most for this study.

A related limitation is conceptual. The coding sees behavior and infers nothing about motive. A brake that signals professional caution to an audience and a brake that regulates the writer’s own reasoning are indistinguishable on the page. Costly brakes, meaning concessions that weaken the author’s conclusion and corrections that admit error, are better evidence than hedging, and both authors produce them. That is the most I can claim.

The social-media record is asymmetric. Joyner’s 2012 Twitter stream survives because OTB captured it. Taylor’s recoverable Bluesky material runs to professional self-promotion and links to longer work. Older material is inconsistently indexed or inaccessible. I use social media where the underlying content is recoverable and do not read missing search results as zero heat, zero posting or zero regulation.

Finally, OTB is not a neutral container around two independently evolving authors. It became a magazine-style publication in 2010 and was redesigned again in 2018. Those changes are confounders if the question is whether aging alone changed the two men. They are the answer if the question is how experienced Internet writers adapt. Taylor and Joyner changed their site in part because their conception of what deserved publication had changed. The environment acted on them. They also redesigned the environment their own e-personalities work in.

That is where the study modifies Virtually You. Aboujaoude was right that the online self does not stay sealed behind the screen, and he describes online and offline life as engaged in constant back-and-forth exchange. The evidence from Taylor and Joyner suggests taking the phrase more literally than the book does.

Posted in Blogging | Comments Off on Steven Taylor, James Joyner and the E-Personality: Aboujaoude’s Virtually You Tested on Twenty Years of Outside the Beltway

Amy Wax and the E-Personality: A Claim-Level Analysis, 1996-2026

Posted on August 17, 2026 by Luke Ford

The simple story about Amy Wax (b. 1953) is that she became more extreme, year by year, as the Internet rewarded her for it.

The coding does not support the story.

I scored 173 substantive claims. The public corpus holds 129 claims from 21 appearances between August 2017 and March 2026. The scholarly comparison holds 44 claims from four works published in 2017, 2018 and 2020 on subjects that overlap the public conversations. Each claim carries scores for certainty, scope, qualification, evidence, engagement with counterarguments and concession, on the scales in the codebook. Each appearance carries scores for interviewer pressure and audience alignment. Public models use OLS with subject fixed effects and standard errors clustered by appearance.

Four processes come out of it.

Scope broadens gradually across the decade. Certainty and qualification do not. They move as a step, and the step lands between August 2022 and March 2024. Medium predicts evidentiary friction. The interviewer predicts permeability.

That last group of findings makes Wax a demanding test case for Elias Aboujaoude (b. 1971) whose Virtually You: The Dangerous Powers of the E-Personality appeared in 2011. Aboujaoude argues that digital environments produce a bolder and less restrained personality, and that traits developed online migrate back into ordinary life. He credits the positive side. The virtual self can say the unpopular thing that offline inhibition suppresses.

The internet supplied many of the conditions under which Wax states her beliefs, but something other than the Internet appears to have supplied the trajectory.

Wax’s first major publication was a 1996 review essay in the University of Chicago Law Review on Robert Wright’s The Moral Animal. It reads as a manifesto. Evolved psychology sets boundary conditions on culture. Sexual asymmetry produces average differences in mating strategy, status competition and risk tolerance. Moral systems mobilize one evolved tendency against another. Communities enforce norms through shame, gossip, stigma and the withdrawal of regard rather than through formal punishment. Social policy dissolves inherited restraints more easily than it builds them. Every theme she pursues for thirty years sits in that essay.

The 1996 article on selective subsidies for two-parent homes and the 1998 article on marital bargaining carry the same anthropology into family law. The 1999 article on unconscious bias applies it to Title VII. By 2011, in “Disparate Impact Realism,” measured group differences in cognitive ability enter as an actuarial premise with causation bracketed.

The continuity runs at the level of framework. In 2009, in Race, Wrongs, and Remedies, Wax walks through the exogenous explanations for the Black-White test score gap, finds them wanting, and declines the hereditarian step. She rests the argument on culture and behavior. That refusal is a fact about the 2009 Wax that the later Wax abandons. While her framework is old, her application to racial cognition, to Asian immigration, to Jews, to women and to sexual minorities arrives later, and it arrives in speech before it arrives on paper.

In the public corpus, each additional year adds about .121 points of claim scope on the five-point scale, with a clustered standard error of .031 and a p value near .0001. Over nine years that is roughly a full scale point.

The corpus shows what the coefficient means. Of the thirty-three public claims that score at maximum certainty and maximum scope together, thirty-one come from 2021 or later. Two come from the whole of 2017 through 2020. The early maximal claims concern immigrant assimilation and bohemian marginality. The late ones cover the sex distribution of civilizational achievement, the proper social position of homosexuality, the desirability of one Anglo-Protestant or pan-European dominant culture, and the loyalties of Asian migrants. The number of populations her claims sort keeps rising.

Broadening is where the digital-audience story does its best work. Scope does not differ between her scholarly and public claims once year and subject are controlled. In the 2017-2020 overlap window the public claims run slightly narrower. So the widening belongs to time rather than to venue, and time here means a decade of invitations to speak about race, sex, immigration and the universities and nothing else.

The first pass through these data showed that each year adds .118 points of certainty and removes .150 points of qualification. Both coefficients are strong. Both are also wrong about the process that generates them.

Add an indicator for 2024 and later to the certainty model and the indicator takes .51 points (p = .01) while the annual trend falls to .043 and loses significance. Do the same for qualification and the indicator takes 1.06 points (p = .014) while the annual trend drops to .005, which is nothing at all. Indicators placed at 2021, 2022 and 2023 do nothing. An indicator placed at 2025 does nothing. The discontinuity sits at 2024 and the linear trend was smoothing across it.

The step is not an artifact of who was interviewing her. Inside the eleven Loury conversations alone, the 2024 indicator takes .58 points of certainty (p = .001) and 1.04 points of qualification (p = .033), and the annual trend goes marginal or null in both. Dropping any single appearance and refitting leaves the certainty step between .35 and .64, significant in nineteen of twenty-one refits.

The Loury series shows it. Mean qualification per conversation, in order, runs 2.17, 2.00, 2.00, 1.20, 2.00, 1.60, 1.67, 2.40. Then it runs 0.50, 0.29, 0.50. There is no downward slope across the first eight conversations, which span August 2017 to August 2022. There is a cliff between the August 26, 2022 conversation and the March 29, 2024 conversation, and the level never recovers.

What sits in that window is Penn. The New York Times covered the dean’s complaint in March 2023. The hearing process ran through 2023. The faculty senate report leaked in early 2024, weeks before the March conversation. Penn suspended her at half pay and stripped the Mundheim chair in September 2024. She sued in January 2025.

That timing supports an explanation the e-personality frame does not contain. A person under formal charge cannot qualify her claims without conceding the charge. Caveats are admissions. The lawsuit, filed January 16, 2025, dismissed August 27, 2025, and appealed to the Third Circuit on January 15, 2026, makes retreat a litigation cost on top of a coalition cost. On the evidence here, the sanctions process is a better candidate for the discontinuity than the accumulated effect of a decade of podcasting.

I cannot separate the two. Her friendly-platform appearances also cluster in 2024. But the step survives inside the Loury series, where the platform is constant and the host is the most persistent challenger in the corpus, and that is the strongest available test.

One result cuts against the hardening story.

Engagement with counterarguments rises .089 points per year in the public corpus, with a clustered standard error of .033 and p = .007, after controlling for interviewer pressure. The later Wax addresses opposing arguments more often, not less. Outside the Loury conversations the annual rise is .335 points.

Set beside the certainty step, that gives a specific picture. Wax states her conclusions more categorically and with fewer caveats than she once did, and she spends more of her time answering objections. Those two things travel together in a person who has been arguing the same case under attack for six years. She has heard the objections. She has answers ready. The answers do not narrow the claim.

At the appearance level it looked as though Wax became less evidentiary over time. Once subject and interview environment enter the model, that disappears. The annual coefficient on evidence is -.022 with p = .46. There is no chronological decline in evidentiary support.

The gap belongs to medium, and it is the largest effect in the study. Restricting to 2017 through 2020, where scholarly and public Wax can be observed at once, and controlling for year and subject, public claims run .38 points more certain, carry 1.04 fewer points of qualification, carry 1.47 fewer points of engagement with counterarguments, and carry 2.96 fewer points of evidentiary support on the five-point scale. Raw means across the full corpus: evidence 3.82 in the scholarly claims against 0.74 in the public ones.

The academic Wax already made large claims. What the scholarly venue demanded was an argument built around the claim.

One 2020 comparison shows the split inside a single year. On July 31 she tells Loury she would bet that innate factors are at work, which reports where she puts her own probability. Months later, in Pursuing Diversity, she lists racism, poverty, culture and innate group differences as competing explanations and declines to apportion causal weight. A decade of radicalization cannot explain a difference that occurs within twelve months. Podcast Wax tells you where her money sits. Law-review Wax tells you which explanations remain live. Neither version is counterfeit. The media activate different standards for converting belief into assertion.

Aboujaoude supplies something the coding cannot reach.

His chapter on the illusion of knowledge collects the research on how people read online. The 2008 British Library study of five years of digital trails found readers skimming, bouncing and abandoning: four minutes per e-book, sixty percent of academics viewing three pages or fewer of a journal article, sixty-five percent never returning. Jakob Nielsen’s eye-tracking work found that online readers scan in an F pattern and forage rather than read. Aboujaoude’s summary of the British Library finding is that people seem to go online to avoid reading.

Put that beside an evidence score of 0.74 and the account closes. Wax’s citations vanish in conversational media partly because she need not supply them and because the audience will not follow them. The listener who hears an unnamed supporting study does not go looking. A ninety-minute podcast still obeys the logic of compression, because compression happens downstream, in the hands of whoever cuts the clip, and the sentence that escapes is never the qualified one.

The network makes certain versions of Amy Wax cheaper to produce, circulate and repeat.

Wax is not a social-media intellectual. Her e-personality is assembled by other people: Bloggingheads, YouTube, podcasts, newspaper interviews, conference video, university statements, student campaigns, reaction clips. Others upload her, clip her, title the episode and tell the audience what kind of person is coming.

Several Wax avatars circulate at once. The courageous dissenter. The racist professor. The race realist. The free-speech martyr. The civilizational conservative. The scholar punished for saying what others fear to say. The faculty member whose conduct Penn judged incompatible with professional standards. Each draws on something real. None matches the person. The network promotes whichever version serves the community circulating it, and the representation acquires causal weight of its own. Aboujaoude calls the fusion of the two selves virtualism, and by September 2024 the mediated Wax and the employed professor had stopped being separable in any institutional sense.

My original hypothesis was that the interlocutor predicts Wax’s rhetorical behavior better than time does. That was wrong. With year and subject controlled, interviewer pressure shows no relationship with certainty (-.044, p = .53), scope (.050, p = .36) or qualification (.025, p = .86).

Pressure predicts one thing. A one-point increase adds .648 points of engagement with counterarguments (p well under .001). That is the largest host coefficient in the study. I scored substantive challenge rather than hostility, because Loury is warm toward Wax personally and remains her most persistent critic.

Four scenes carry it better than the coefficient.

On July 31, 2020, Wax tells Loury she would bet on innate factors. Loury answers within a minute. He would bet that Massachusetts outscores Mississippi, and he says that fact is inadmissible as political discourse because Americans share a country. Wax moves to the politicization of Eric Turkheimer and never returns. She concedes nothing. The listener hears the objection at full strength.

On January 24, 2022, she tells Gad Saad (b. 1964) that the Ashkenazi advantage in tested intelligence is partly innate. Saad says he feels marginalized by the comment. The joke is the entire resistance. She extends the claim to Asian-White gaps and moves on. That conversation carries a mean qualification score of 0.20 and a mean counterargument score of 0.00, the only zero in the corpus.

On August 29, 2022, Richard Hanania (b. 1985) asks why men permitted the feminization of institutions. Wax answers that you can only impose something on people who allow it, and says that part she has never understood. She names an unsolved problem inside her own framework. She does not do that with Saad, and she does not do it with Louise Perry (b. 1992).

On August 14, 2024, Perry asks whether women should return to five percent of the student body. Wax says ideally, then routes the answer back through meritocratic procedure. No resistance follows the ideal.

Same speaker, same framework, different rooms, different products.

Of the 129 public claims, concession scores zero in the great majority. Fourteen show factual or local acknowledgment. Six show local narrowing. One shows movement in a supporting conclusion. Nothing in the corpus shows Wax abandoning the principal conclusion around which an appearance is organized. Interviewer pressure predicts concession only weakly (.109, p = .078).

The study cannot show that Wax never changes her mind. It shows that in these rooms, challenge produces local revision. Loury can get her to grant a fact, accept an exception, acknowledge uncertainty, or say what follows if a premise weakens. The next step rarely arrives: yes, that is true, and therefore the conclusion I brought into this room has to change.

That is why hard interviews keep their value when nobody switches sides. A good interlocutor may fail to move her and still improve what the audience receives.

The reliability audit is the most useful thing in the workbook, and it is worse news than it first looks.

Re-coding sixty randomly selected claims with date, host, medium, subject and first-pass scores hidden produced quadratic weighted kappas of about .75 for certainty, .65 for scope and .78 for qualification. Those are the variables carrying the hardening results, and they hold up. Evidence came in at .38 and counterargument at .43.

Comparing the two passes, the blinded scores run systematically lower: evidence by 1.10 points on average, counterargument by 0.83, scope by 0.45, certainty by 0.15. Knowing the source inflates the score. That threatens the medium comparison directly, since that comparison rests on absolute levels of evidence, and a coder who knows a claim comes from a law review is primed to find the argument around it. The 2.96-point medium gap survives even if the entire 1.10-point drift is attributed to venue cues, but it shrinks, and the honest range is roughly two to three points rather than three.

This is also a same-model re-code. It measures the stability of the instrument.

Second, pressure and counterargument engagement were scored from the same transcripts in the same pass. A coder who hears Wax address an objection is likelier to score the host as challenging. The .648 coefficient is inflated by an unknown amount. The fix is cheap: score host pressure from host turns alone, with Wax’s turns removed, before scoring hers.

Third, pressure and alignment are nearly collinear with host identity. Loury appearances score 2 or 3 on alignment; Saad, Alexander Riley, Nathan Cofnas, Freddy Gray and Hanania all score 4. The correlation between pressure and alignment across appearances is -.69. A model controlling both is close to a model controlling for whether the host is Loury, and there is little independent variation left to identify either effect.

Fourth, the instrument saturates. In 2024, eighty-three percent of certainty scores sit at the ceiling of 4, sixty-nine percent of scope scores at the ceiling of 5, and sixty-six percent of qualification scores at the floor of 0. After 2024 the scales cannot register further hardening even if it occurs. Ordered logit reproduces the year effect on certainty, so the finding is not an artifact of treating ordinal scores as continuous, but every coefficient in the late period is attenuated.

Fifth, subject composition drifts. Race and achievement supply nine of twenty claims in 2017 and none of the gender claims; gender and family supply nine claims in 2022 and eight in 2024 and none before 2022. Education and poverty vanish after 2019. Several year-by-subject cells are empty. Subject fixed effects are absorbing variance that may belong to the process under study, because the topics she is invited to discuss are downstream of the same forces that produce the hardening.

Sixth, the scholarly corpus has three distinct time points and ends in 2020. If Wax published no comparable scholarship on these subjects afterward, then medium and late period are partly collinear across the full corpus, and the exit from scholarly venues is a finding rather than a nuisance. Aboujaoude’s migration thesis predicts that exit.

So the findings rank. The 2024 discontinuity is the most secure result and the one I did not expect. Scope broadening is secure. The medium gap is large enough to survive the coding drift, with its magnitude uncertain. The permeability result needs host-only coding by an independent human coder before its coefficient deserves any weight.

The strongest available control sits in the transcripts already collected.

Loury (b. 1948) appears in eleven of these conversations as interviewer. He is also a public intellectual of the same generation, on the same platforms, across the same decade, discussing the same subjects, and he faced no disciplinary proceeding. Score his claims with the same instrument.

The predictions separate the hypotheses. If Loury shows the same 2024 step, the finding is a period effect and belongs to the political moment. If his qualification holds flat across 2022 to 2024 while hers collapses, the step belongs to the sanctions and the litigation, and the e-personality frame is the wrong frame for the timing even if it remains right about the medium. If his evidentiary support falls in podcasts and holds in his written work, the medium result generalizes past one speaker, which is what would turn it into a claim about public intellectual life rather than a biography.

The Internet did not turn a conventional scholar into a person with different beliefs. Her framework is on paper in 1996 and has not moved much.

Two things changed, and they changed. The range of the claims widened steadily across the decade, one group at a time, in step with the invitations. The restraint around the claims did not erode gradually. It broke, in the window between August 2022 and March 2024, while Penn’s disciplinary machinery was running toward its conclusion.

Medium explains where the evidence went. Time explains the widening. The sanctions process is the best available explanation for the timing of the collapse in qualification, and the coding cannot rule out that a person who has been formally charged simply stops offering caveats because caveats are concessions.

Academics do not hold opinions and then distribute identical copies through different channels. The medium takes part in producing the opinion as a public object. A proposition in a law review arrives with citations, boundaries and named adversaries. The same proposition in a podcast arrives as a declaration. Repeat the declaration, reward it with invitations, attacks, clips, followers, a hearing board and a federal docket, and it stops being something the person believes and becomes part of who she publicly is. At that point retreat costs more than an argument. It costs a character, and after January 2025 it costs a case.

Aboujaoude asks the right closing question himself, near the end of Virtually You. He wonders whether the online self is in some sense more real than the real thing, whether the disinhibited version returns us to what we were before the social contract taught us to hold our tongues. He leaves it open, and so do these data.

Wax’s e-personality is Amy Wax under a different set of restraints and rewards. The record shows when the restraints came off.

Posted in Amy Wax | Comments Off on Amy Wax and the E-Personality: A Claim-Level Analysis, 1996-2026

After Cathy’s Table: How Digital Platforms Reshaped a Generation of Professional Writers

Posted on August 17, 2026 by Luke Ford

Cathy Seipp (1957-2007) died in March 2007, just as the media world she had anticipated was becoming normal. She had spent the previous decade moving between magazines, newspapers, online columns and her own blog while gathering around her a Los Angeles network of writers, lawyers, academics and broadcasters. Mickey Kaus (b. 1951), Matt Welch (b. 1968), Ken Layne (b. 1964), Stephen Bainbridge (b. 1958), Amy Alkon (b. 1964), Sandra Tsing Loh (b. 1962) and Hugh Hewitt (1956-2024) moved through the circle. Nancy Rommelmann (b. 1959) and others belonged to the wider social ecology. They disagreed about politics. They shared a professional temperament. Independence mattered. Candor mattered. Wit mattered. Suspicion of official accounts mattered. The writer who could build a name without waiting for an institution to confer one had standing.

That makes the Seipp circle useful for understanding what digital media do to professional writers. The members belonged to an existing culture of journalism and argument before blogs, Twitter, Substack and podcasts became dominant. They then entered those environments at different speeds and under different professional constraints.

The variation gives us something close to a natural experiment. We can ask what happens when people with overlapping formation encounter different technological incentives. Some became relentless bloggers. Some moved toward Twitter and short-form commentary. Some built subscription podcasts. Some retained strong institutional homes. One abandoned political blogging for a printed desert magazine. Another divided his personal and professional selves between platforms. Some preserved the disciplines of their original profession. Others allowed personality and publication to fuse.

I was in the middle of this crowd from 2002-2007. After she died, I lost touch with everyone.

I miss them.

Cathy had no successor.

During her final five years, I didn’t see blogging and the internet damaging Cathy’s friends. It only seemed like a great thing. The online world appeared to offer unlimited possibilities for the energetic writer, however financially precarious our present. We felt like the future.

Looking back at this golden age of blogging, I’m reminded of the 1985 movie Plenty. Roger Ebert wrote:

At the end of World War II, a young woman stands on a hilltop in France and, as the sun bathes her in golden light, she says, “There will be days and days and days like this.” That image provides the last shot in “Plenty,” which is the story of how very wrong she was. The woman is Susan Traherne, a young English fighter in the French Resistance. She is not very seasoned and perhaps not very good at her job, but she stays alive behind enemy lines and she has a brief, poignant love affair with one of the men who parachuted down out of the night sky to fight the Germans.

The movie opens with her days in France. It follows her through the next 15 or 20 years, and ends with that painful flashback to a day when she thought the future held nothing but good things for her. But nothing else in her life is ever as important, as ennobling or as much fun as the war. She is, perhaps, a little mad. She confesses at one point that she has a problem: “Sometimes I like to lose control.”

Elias Aboujaoude (b. 1971) published Virtually You: The Dangerous Powers of the E-Personality in 2011. His framework helps organize the differences. The Seipp diaspora also exposes the limits of the theory. The Internet reshaped these writers by selecting among personalities they already possessed.

(I found the book in April 2018 and was immediately struck by its truth. I repeatedly cited it over the next eight years. There was a Luke prior to finding this book and a more chastened one afterward.)

Aboujaoude notes that online environments elicit a different self. Ordinary inhibitions weaken. Distance from the other person removes some of the social signals that constrain speech. Status hierarchies become less visible. Publication becomes easier. Reactions become faster. Against this background, Aboujaoude identifies several recurring tendencies. The online self tends to become grandiose, self-centered, aggressive, juvenile and impulsive.

The virtual self can also be bolder and more effective. Reduced inhibition can free a shy person, make an unpopular opinion easier to express and allow connections that ordinary social settings might prevent. That positive side is essential for understanding writers. Blogging solved problems for writers whose institutions had become slow, risk-averse or inhospitable to certain work.

Aboujaoude’s stronger proposition concerns spillover. Behaviors repeatedly enacted online can migrate elsewhere. A person accustomed to being unusually daring or undiplomatic online can become more daring or undiplomatic generally. Professional writers give us a way to test that proposition. Their work leaves archives. We can observe the writer before the blog, on the blog, on Twitter, in a podcast and sometimes back inside a formal article. The evidence shows public registers switching on and off.

The first complication is that Seipp’s group already possessed many traits that Virtually You associates with online behavior. The circle admired independence from editors and institutions. Its members liked irreverence, strong opinions and sharp language. Seipp understood that blogging rewarded frequent posting, links, reaction to the news and interaction with readers. Near the end of her life, she admitted to her own “junior-highschoolish” snark and suggested that a degree of immaturity might help a blogger in this 2006 column. She also described blogging as the comparatively fun and low-stress part of writing, an outlet into which ideas about politics, media, family and friends could flow quickly, in this NPR essay.

That matters. If Mickey Kaus mocks a politician on X, we cannot classify the behavior as Internet-induced disinhibition. Kaus was a contrarian long before Twitter. If Amy Alkon insults somebody on a blog, we need to know whether the pre-social-media Alkon already wrote that way. She did. Her 2004 blog archive could be spectacularly caustic.

The same problem appears with Nikki Finke (1953-2022), who was the talk of the Seipp ecology. A 2004 Los Angeles profile described a combative journalist whose friends already worried about her self-destructive tendencies. Finke distinguished between approaching the line and disregarding it. Her aggression, obsessive sourcing and refusal to accommodate Hollywood sensibilities all predated Deadline Hollywood Daily. She was never sued for libel and she was right that anyone who published she was crazy operated with reckless disregard for the truth.

The Internet encountered a population already rich in potentially digital traits. The question is what happened to their existing dispositions when platforms began rewarding some of them.

Everybody in this crowd was self-aware, disciplined and successful prior to blogging. Nobody dominated conversations. Everybody was a keen listener. Nobody appeared to have legal problems nor drug and alcohol problems. Everybody had the traits of the winner. They earned their living in that rare way — with their words.

You’d love this crowd to be your neighbor because they were great fun while living up to their responsibilities. There were no psychopaths here. No kleptos. No rapists. No liars. No brawlers. No felons. You trusted them around your kids because they modulated their personalities to the situation.

Kausfiles began in 1999. Kaus, the second smartest person in Cathy’s world after Eugene Volokh, the second best performer after Sing Loh, and the funniest and most fearful member of the group, later recalled that his first post was a conventional article with a lead, developed paragraphs and a conclusion. Then fellow early blogger Walter Olson teased him about allowing material to sit unchanged while everybody else was producing fresh content to attract traffic. Kaus understood the lesson. He began producing fresh linked commentary on the day’s news. In his own retrospective telling, the platform taught him how to blog. His ten-year Kausfiles retrospective was a behavioral adaptation to an attention environment.

The underlying Kaus did not suddenly appear in 1999. The skepticism, policy obsession, argumentative curiosity and willingness to depart from mainstream liberal positions were already present. What changed was the basic unit through which those qualities reached readers. Book became article. Article became blog item. Blog item became linked reaction. Linked reaction became tweet.

The change can now be seen inside Kausfiles, because most of the site functions as a repository for material first produced in the rapid social-media register. The platform Kaus helped pioneer has been colonized by the platform that succeeded it. Every time I surf to Kausfiles, I regret it because there’s no longer any value there.

His immigration writing gives us a control because the underlying position has remained stable. Kaus has argued for decades that high levels of lower-skilled immigration can hurt less-skilled American workers and weaken a politics built around working-class interests. What varies is the register. In longer work, qualifications, alternative explanations and chains of inference remain visible. In rapid social-media commentary, those intermediate stages compress. The proposition arrives closer to its argumentative endpoint.

The Internet did reduced the size of the container in which Mickey Kaus thought in public. That is close to what Aboujaoude feared when he argued that compressed forms strip complicated questions of the weight and elaboration available in slower writing. Even here the deterioration story is incomplete. Kaus still links. He still asks questions. He changes his mind on particulars. The epistemic personality has survived better than the literary form.

Finke’s pre-blog LA Weekly prose was ferocious. A 2004 column about a Hollywood law-firm dispute opens with combat, betrayal and “backstabbing weasels.”

What changed in 2006 was the publication interval. Finke launched what became Deadline Hollywood because the Internet could do something her weekly column could not. She could report developments as they happened instead of waiting for the next publication cycle. The Los Angeles Times described the creation and later sale of the site here.

That technological change transformed the competitive environment around Hollywood journalism. Speed became part of Finke’s product. Being first became visible. Updates could be continuous. A scoop could produce traffic immediately. A rival could be beaten publicly rather than in the next edition. Finke became inseparable from the operation. For a period, reporter, columnist, editor, publisher, tip line and personality occupied the same body.

The platform did not create Finke’s grandiosity, aggression and impulsivity. It made some traits more profitable. The old system sometimes punished Nikki Finke for being Nikki Finke. The new one monetized it.

The e-personality arises through incentives. If a platform pays attention, status and money every time one portion of the personality appears, that portion will appear more often. The writer learns what works. Eventually the audience learns to expect it. Then the professional identity and the rewarded persona begin to fuse. Finke is a good example.

Matt Welch complicates the theory because his move into podcasting increases personality while sometimes restoring intellectual restraint. He moved through newspapers and magazines, Reason, cable television, Twitter and eventually the long conversational world of The Fifth Column without ever self-destructing.

Welch, the libertarian, had the best radar for how to thrive in institutions such as the Los Angeles Times and Reason magazine while sounding heterodox and a little dangerous in a completely safe way for his employers.

The podcast is less formal than an edited Reason article. There is more profanity, autobiography, mockery, gossip and inside humor. The hosts are part of what subscribers are buying. Welch has made the audience relationship explicit. In this May 2026 subscriber note, he wrote that he reads every email and comment from subscribers and described how listener complaints had altered audio production, video, merchandise and guest selection. One repeated complaint was that the show needed more guests, especially people with whom the hosts disagreed. The program responded by increasing them.

The paying audience asked for more friction.

Welch’s edited Reason examination of Donald Trump’s interpretation of William McKinley’s tariffs works through historical evidence and includes McKinley’s own later movement toward reciprocity. The podcast register is more personal and socially disinhibited. A long podcast can place Welch opposite a cohost or specialist who can interrupt, disagree or demand elaboration.

Twitter removes the other person’s face. A conversational podcast puts another mind back in the room. This distinction may be more important than the distinction between print and digital. Long conversation can increase social disinhibition while decreasing epistemic disinhibition. The speaker jokes more and qualifies more. He swears more and explains more. The resulting person sounds less professionally restrained while reasoning under greater immediate resistance. Aboujaoude’s category of “online” is too broad to capture that.

Amy Alkon provides the same lesson. The Advice Goddess persona existed before mature social media and was strong from the start. Her early blogging intensified its directness because the blog could become a running diary of fools, rude people, media encounters and cultural irritants.

Alkon, the loudest, brashest and most extroverted of the set while simultaneously a great listener and a good friend, also used digital media to build a different intellectual environment. Her advice writing incorporated behavioral science. A 2021 column on reactions after sex invokes research by evolutionary psychologists Martie Haselton and David Buss (b. 1953) before applying it to a reader’s predicament. Her HumanLab put scientists and researchers into the conversation. One episode with animal behaviorist Jennifer Verdolin was structured around the scientific study of flirting and relationships.

That produces multiple public Alkons. There is the Advice Goddess confronting some social offense. There is the syndicated columnist applying scientific literature. There is the host talking with somebody who possesses independent expertise. The identity of the interlocutor can matter more than whether the medium is digital. A generalist conversation allows broad speculation. A specialist conversation imposes boundaries. The e-personality is not singular.

When ProfessorBainbridge.com began in 2003, the blog mixed corporate law with Catholicism, legal education, wine and food. Over time, most of the digressions moved to Twitter while the blog became increasingly concentrated on corporate law and governance. Bainbridge describes the resulting site as more professional, more scholarly and less personal. His own retrospective is here.

Bainbridge sorted his personae. Twitter received one register. The professional blog received another.

The current Substack displays more personality than a law review. A 2026 post on a Delaware corporate-law decision was headlined “The Court Got It Right, and So Did I”. That is personal-brand behavior. Beneath the self-congratulation sits conventional legal analysis. He names the case, reconstructs the constitutional objection, identifies the court’s holding and directs readers to his earlier work and related scholarship. The persona has intensified.

This is why counting boastful headlines, profanity or first-person references would produce a bad study of online intellectual deterioration. Personality and epistemic quality can move independently.

Eugene Volokh (b. 1968) was the smartest person in the Seipp orbit and the most ambitious after Andrew Breitbart. He published The Volokh Conspiracy. Most of Volokh’s online work retains the professional habits of a law professor. Cases are named. Opinions are quoted. Doctrinal distinctions remain visible. Counterexamples matter.

The most instructive example concerns the relationship between blogging and scholarship. In 2021 Volokh posted a draft argument about potentially defamatory omissions in reports about legal proceedings. He invited reader feedback and asked for help finding a clearer title. Comments persuaded him to change it. The resulting article, “Libel by Omission of Exculpatory Legal Decisions“, appeared in the Notre Dame Law Review.

The causal sequence is the reverse of the pessimistic Internet story. A scholarly idea enters a blog. Readers criticize it. The author revises it. The revised argument enters a law review. The digital medium becomes part of the quality-control system.

The same professional continuity appears in Volokh’s podcasting. A 2026 episode on student threats and campus speech was organized around two cases pointing in opposite directions. A later episode on defamation used contemporary lawsuits as the entry point into a structured explanation of American defamation doctrine. The speech is less polished than a law-review article. The lawyering remains.

Volokh had professional ballast. A writer who has internalized a discipline may no longer require an editor to supply all of the restraint. The profession has installed some brakes. A legal claim encounters cases. A scientific claim encounters studies. A reported claim encounters sources. A pure political take may encounter nothing except another take.

Sandra Tsing Loh, the best performer in the set, created The Loh Down on Science which consists of ninety second episodes. If compression produced the intellectual deterioration Aboujaoude feared, the program should be maximally vulnerable.

Its production system points in the opposite direction. UC Irvine partnered with Loh to produce short, humorous explanations of scientific research and train graduate students in science communication, as described in this university account. Graduate researchers find studies and write scripts. Editors and other participants review the work. During the pandemic, the program described a distributed process in which young scientists researched, wrote, edited and recorded the material through digital tools, documented here.

A ninety-second radio piece can contain more professional friction than a 4,000-word independent newsletter. Short is not impulsive. Digital is not unedited. Informal is not epistemically weak. The production system matters more than the word count.

Ken Layne, the biggest introvert in the set, gives us the rarest case. He left. Layne had been involved in early Internet journalism and later in the vicious political-blogging environment of Wonkette. By 2015 he was describing the online news economy as a competition for ever more extreme headlines attached to the same small pieces of information. He said one reason Desert Oracle contained little news was that people could use less of it. The Los Angeles Times profiled the project here.

Five years later he had come to regard the Internet as a bad environment for writers because readers often glanced at material to decide whether they agreed with it.

His response was environmental. He built Desert Oracle around a printed object, physical landscape, radio, narrative, folklore and the slower rhythms of the desert. Digital distribution remains, but it orbits something anchored outside the rapid national reaction cycle.

Layne reverses the direction assumed in Virtually You. The medium acts on the person. Then the person notices. Then the person acts on the medium by leaving it. This capacity for platform choice needs to be added to any mature theory of digital personality. People are not merely shaped by environments. Some eventually learn which environments produce versions of themselves they do not like.

Several figures around the Seipp ecology prevent us from exaggerating the novelty of the digital transformation.

Jill Stewart shows that alternative-weekly journalism had already developed many supposedly Internet-native traits. It rewarded naming enemies, strong first-person judgment, provocation and visible conflict.

Hugh Hewitt supplies the broadcast version of the same control. Talk radio already rewarded rapid reaction, host-centered authority, audience intimacy and coalition signaling before Twitter arrived.

Heather Mac Donald (b. 1956) supplies another kind of control. Her rhetorical certainty and strong thesis-first prose long predate podcasts and social media. Her continuing relationship with the Manhattan Institute and City Journal also provides an institutional floor beneath the personality economy. The think tank gives her a professional home while City Journal supplies regular publication and editorial discipline.

Nancy Rommelmann occupies a middle position. Subscription media have made the writer more central to the product, but field reporting remains an important part of her work. Personality has expanded without replacing reportage.

These cases warn against reading every contemporary stylistic trait backward into the technology. Some writers were already combative. Some were already personal. Some worked in institutions that already rewarded speed and audience intimacy. The proper measure is change from the writer’s own baseline.

Across the cohort, the clearest digital changes concern four things.

The first is publication latency. Finke is the purest example. Something that once waited for a weekly column could appear immediately. Kaus learned that fresh content generated attention. The distance between thought and publication narrowed.

The second is the visibility of feedback. A magazine writer may hear eventually that an article succeeded. A blogger sees links and comments. A social-media writer sees likes and reposts. A subscription host sees renewals and cancellations.

The third is the growing importance of the personality itself. Readers once bought an issue containing a writer. Increasingly they subscribe to the writer. That is a major institutional shift. The author becomes part of the distribution system.

The fourth is the fragmentation of professional identity into platform-specific registers. Kaus on X is not Kaus in a long argument. Alkon in a blog fight is not Alkon interviewing a scientist. Bainbridge on X is deliberately not Professor Bainbridge on his professional site. Welch in Reason is not Welch drinking and arguing with his friends on The Fifth Column. The same person contains multiple communicative selves. Different platforms bid for them.

The evidence is weak for the proposition that digital media made these professional writers less capable of dealing with evidence. Kaus became more compressed, but links and empirical argument remain. Finke’s public persona became enormous, but sourcing remained the foundation of her power. Welch became more informal and audience-entangled, but long conversation sometimes increases exposure to criticism. Alkon built a science-interview operation. Bainbridge became more professionally specialized. Volokh used blogging as part of a scholarly revision process. Loh built strong research and editing behind a compressed product. Layne responded to what he saw as degradation by changing environments.

Professional writers differ from the ordinary Internet users who dominate much of Aboujaoude’s discussion. They enter digital life with a preexisting occupational identity. The reporter has rules. The lawyer has rules. The professor has rules. The science communicator has rules. Those rules compete with the rules of the platform. Sometimes the platform wins. Sometimes the profession wins. Usually some hybrid emerges.

The Seipp cases suggest a more precise theory of digital reshaping than Virtually You provides.

The first variable is publication friction. How much time, editing and resistance stand between thought and publication?

The second is feedback intensity. How quickly does the writer discover which behavior attracts attention?

The third is persona dependence. Does the audience primarily want information, or continuing access to a personality?

The fourth is professional ballast. What external disciplines remain strong enough to constrain the performance?

The fifth is platform agency. Has the writer learned to compartmentalize, modify or abandon environments that reward undesirable behavior?

These variables explain the cohort better than ideology does. Kaus and Finke experienced enormous changes in tempo because their media removed friction. Welch became more personality-driven as subscribers moved closer to the production process, but the conversational format restored interpersonal resistance. Bainbridge partitioned the registers. Volokh preserved unusually strong professional ballast. Loh surrounded extreme compression with an institutional production process. Layne exercised platform agency by leaving political blogging. Mac Donald retained institutional ballast. Hewitt and Stewart show that many digital incentives had analog ancestors.

Aboujaoude was right to notice that online environments can call forth a recognizably different person. He was also right that these differences need not result from conscious deception. People can begin behaving differently in environments that feel different.

The Seipp diaspora suggests that the most useful unit is not a single “e-personality.” It is a repertoire. A professional writer already contains aggression, caution, humor, vanity, curiosity, generosity, certainty, doubt, competitiveness and discipline in different proportions. Platforms change the relative prices.

Twitter makes compression cheap and patience expensive. A personal newsletter makes intimacy valuable. A subscription podcast makes recurring personality valuable. An edited magazine makes factual embarrassment more expensive. A law review makes unsupported doctrinal claims extremely expensive. An expert interview makes pretending to know something more difficult.

Repeated exposure to one reward system makes one portion of the repertoire more practiced. That portion can eventually appear natural. It may become what the audience believes the writer really is. The technology has not created a personality. It has selected one.

This returns us to Seipp. Her importance to the story lies partly in that she assembled, before mature social media existed, a human network that already rewarded many of the characteristics the coming media system would prize.

Her dinners created relationships. Information moved laterally. Reputation moved through people. Wit produced status. Candor produced status. The introduction between two useful strangers created value. The network rewarded the individual writer.

Then software arrived and reproduced much of the architecture at enormous scale.

Nikki Finke shows what happens when a preexisting temperament finds nearly perfect technological reinforcement. Mickey Kaus shows what happens when the unit of public thought gets steadily smaller while the underlying intellectual preoccupations remain recognizable. Matt Welch shows that personality media can simultaneously increase social informality and restore intellectual resistance. Amy Alkon shows that the identity of the interlocutor can matter more than whether the communication is online. Stephen Bainbridge shows that a writer can divide his public selves by platform. Eugene Volokh shows that professional discipline can survive the removal of traditional gatekeepers. Sandra Tsing Loh shows that brevity need not mean impulsivity. Ken Layne shows that the writer can eventually reject the environment doing the shaping. Jill Stewart and Hugh Hewitt show that many supposedly digital tendencies have pre-digital ancestors. Heather Mac Donald shows what institutional ballast looks like over decades.

Cathy’s friend David Rensin resisted the pull of the blog. He was in the middle of her world, attended her funeral, and she identified him as one of the good writers contributing to Kevin Roderick’s Native Intelligence project at LA Observed. Her praise contained a criticism. She thought Rensin and several of the other Native Intelligence writers were producing material fundamentally unsuited to blogging. Their pieces were too long, insufficiently reactive and better suited to leisurely reading on paper.

Rensin more or less agreed. In a 2007 post he wrote that before joining Native Intelligence he had never had a blog because he did not have enough opinions and did not want to share his personal life. Then he proceeded to write a long comic essay about dyeing his hair. In a 2007 interview he described himself as someone who interviews people and transforms what he learns from them into books. The self is an instrument for eliciting somebody else’s story rather than the subject of the story.

When Dennis Hopper (1936-2010) died, Rensin did not manufacture an instant opinion. He reopened an interview he had conducted with Hopper in 1989 and let the archival encounter do the work. See Rensin on Hopper. The pattern is the interviewer’s method transferred into a blogging environment. Blogging gave Rensin permission for more first-person comedy and domestic disclosure. It did not turn him into a reactive pundit. The platform offered him a faster self. He declined it.

I don’t think I’ve ever met a man more secure than David Rensin. He never seemed needy and he never boasted.

Nick Gillespie (b. 1963) is hard to classify because the distinction between his professional identity and his digital identity was weak from the beginning. He was already participating in Internet culture through Suck in the 1990s, writing under the pseudonym Mr. Mxyzptlk. A 1997 piece later republished by Reason is profane, exaggerated, funny, culturally omnivorous and gleefully anti-reverential. The later Internet did not have to teach Gillespie how to write in an irreverent digital register. He arrived with one.

His connection with Seipp ran through the overlapping Reason and Los Angeles journalism worlds. Seipp contributed to Reason, helped organize a Los Angeles Press Club event around Gillespie’s Choice: The Best of Reason, socialized with the magazine’s people and wrote admiringly about Gillespie’s patience during a long interview.

The subsequent career was media expansion without epistemic deterioration. Gillespie moved from editing the print magazine to editing Reason.com and Reason TV and eventually to hosting The Reason Interview. He conducts long interviews with authors, researchers, activists and political figures while remaining an editor at large.

The visible personality is strong. Leather jackets, pop-culture references, jokes and the cultivated outlaw persona are part of the product. Unlike Finke, Gillespie never became separable from an institution in the same way. Reason supplies editorial continuity, ideological boundaries broad enough to permit internal argument and a steady supply of colleagues willing to disagree with one another. Where Welch moved into personality media and later recovered friction through podcast interlocutors, Gillespie’s personality was incorporated into an institution early enough that institution and persona evolved together. The e-personality works for him because his professional institution learned how to house it.

How he pulled this off while consuming enormous qualities of alcohol that would have sunk a normal man, I’ll never know.

One day in 2006, Nick said to me, what are you going to do now? You seem to have peaked years ago.

I said yeah, and I thought, I have no idea. I’m lost.

Kevin Roderick is good comparison with Nikki Finke. Both were established Los Angeles journalists who discovered that the Web allowed them to publish independently. Both built sites based on professional knowledge and source networks. Both became indispensable to people working inside the industries they covered. Then they built almost opposite kinds of blogs.

Roderick spent roughly two decades at the Los Angeles Times. He launched LA Observed in 2003 after someone showed him blogging software and he recognized that Los Angeles lacked the sort of media site he thought the city needed. He had already spent time in the WELL online community.

LA Observed accumulated roughly ten thousand entries, most of them Roderick’s. Its stated formula was “independent reporting, informed commentary and selective linkage” about Los Angeles media, politics and culture. That description sounds like blogging. The temperament of the site often did not. Seipp made the point in 2006. She acknowledged that virtually every Los Angeles media person checked LA Observed because its media information was valuable. She also complained that Roderick was not friendly toward the kind of irreverence directed at his former newspaper that she valued, and noted that he had eliminated comments.

Roderick adopted the technology while resisting part of its culture. He used the blog to abolish one newsroom and then rebuilt many newsroom norms inside the site. Attribution mattered. Local knowledge mattered. Curation mattered. Professional reputation mattered. He became an editor and publisher as well as a blogger.

Like Rensin, Roderick radiated no need. He knew who he was and he didn’t pretend to be anyone different. He was quiet, courteous and generous.

Finke removed latency and made speed, combat and personality increasingly central to the product. Roderick removed the institutional employer and preserved much of the institutional temperament. Both demonstrate independence. Only one became Nikki Finke. The loss of an employer does not cause the loss of professional restraint. A journalist can carry the newsroom inside himself and recreate it around a URL. Finke found a temperament the platform could amplify. Roderick domesticated the platform.

Luke Y. Thompson complicates the study because his career does not contain a clean “professional first, Internet second” break. He began 1999 by hand-photocopying movie and pop-culture opinions in zines and ended the year as a critic and assistant arts editor at New Times Los Angeles. He subsequently wrote for the Los Angeles Times, LA Weekly, OC Weekly, E! Online, Forbes and other outlets, edited Topless Robot, worked at Nerdist, developed a toy-blogging specialty and became a substantial part of Deadline’s Comic-Con coverage. See Thompson’s career biography. He was also part of Seipp’s Los Angeles world and attended her funeral.

Thompson is almost the opposite of Rensin. Rensin entered blogging with habits that resisted it. Thompson’s habits were compatible with it from the start. The relevant change is expansion of what could count as professional expertise. The old movie critic reviewed films. Thompson could review films, cover a Comic-Con panel in real time, photograph toys, report collector news, review fast food, write about wrestling and move continuously between professional criticism and fandom.

Two features of his subject matter kept the expansion honest. Toy claims are checkable against the object. Film judgments announce themselves as criticism. Comic-Con reporting requires being in the room. His informal voice reads as expertise because the domains reward the kind of knowledge fans can verify. The Internet permits the critic to display more fandom, humor, private taste and niche obsession than an old newspaper arts page would have tolerated. Those are the traits that create trust inside fan communities. The medium changed what kind of professional writer he could become.

Luke didn’t hate many people. He really hated Andrew Breitbart. For Breitbart, however, Thompson didn’t register.

Breitbart (1969-2012) was in Cathy’s world. He hosted me, Cathy and her daughter Maia for a Memorial Day weekend in Palm Springs in 2004 where we watched him operate the Drudge Report. He attended Seipp’s funeral with his wife. He differed from almost everyone else discussed here because he did not merely adapt to new media. He developed a theory of political action around it.

He was the most ambitious member of Cathy’s world.

He discovered the Drudge Report in 1995, contacted Matt Drudge (b. 1966), and eventually worked with him. He later helped Arianna Huffington (b. 1950) develop the Huffington Post before building his own network of “Big” sites. Nick Gillespie’s assessment after Breitbart’s death captured the essential achievement: whatever one thought of his politics or methods, Breitbart had demonstrated that people excluded from or dissatisfied with existing media could build their own outlets.

The distinction from Finke is intentionality. Finke discovered that the Web increased the value of traits she already possessed. Breitbart understood the media architecture as the political battlefield. In his 2011 interviews around Righteous Indignation, he described himself as fighting a political-media complex and explained his method as forcing stories into wider circulation until traditional outlets had to react. He also encouraged ordinary people armed with digital cameras to produce stories that established institutions were ignoring.

This is Aboujaoude’s positive e-personality pushed toward its political conclusion. The Internet gives the individual confidence against institutions, removes hierarchy, broadens agency and permits direct public action. Aboujaoude allows that the online self can become bold and effective enough to voice opinions that ordinary constraints might suppress. Breitbart built a career out of the capacity.

He also demonstrates the associated danger. When media is understood as combat, attention from an adversary becomes evidence of effectiveness. Provocation is an operational technique. Personal attack, spectacle, confrontation and the speed of circulation become strategically valuable. Volokh’s professional discipline resists platform pressure. Breitbart developed a professional discipline out of platform pressure. Conflict was the craft. He is the institutionalization of disinhibition into a theory of media. Finke made the personality a newsroom. Breitbart made the personality a movement.

Ray Richmond is a middle case. He was in Seipp’s friend and colleague. Seipp moderated a 2005 American Cinema Foundation discussion at which Richmond and Matt Welch were panelists. He had spent many years inside conventional entertainment journalism before starting Past Deadline in 2006 as a blog connected to The Hollywood Reporter. The blog quickly became another channel for entertainment news and opinion, with stories sometimes traveling from Richmond’s site into wider national circulation.

When Richmond wrote in 2007 about the late Merv Griffin’s (1925-2007) concealed sexuality, the column created a dispute over its handling by The Hollywood Reporter and Reuters. Richmond then used Past Deadline to explain the sequence of events and defend his position. That is a distinctly digital development. The old columnist depended on the publication for both the original argument and any subsequent clarification. The blogger possessed a parallel voice. The institution could publish, edit, remove or retitle a piece. The writer could immediately tell readers what he believed had happened.

The independence produces a secondary effect. The author becomes part of the story. Editorial disagreement becomes publishable content. Richmond nevertheless remained more attached to conventional professional roles than Finke. He was a critic and entertainment reporter operating through a blog rather than a personality who happened to report. Blogging increased immediacy, authorial ownership and the capacity to fight an editorial dispute in public. It did not swallow the underlying occupation.

Anita Busch is the strongest argument in the enlarged cohort against confusing conflict with digital disinhibition. She belongs as an adjacent Hollywood-journalism control.

Her professional life was highly combative before blogs mattered. In 2001, while editor of The Hollywood Reporter, she resigned after a dispute with management over the paper’s handling of an investigation into one of its columnists. Contemporary Los Angeles Times coverage described her resignation as a refusal to stand behind management’s decision. See The Los Angeles Times on Busch’s resignation. The following year, while investigating Steven Seagal (b. 1952) and alleged organized-crime connections, Busch found a damaged windshield, a dead fish, a rose and a “Stop” warning on her car. Later law-enforcement proceedings connected the threat to the Anthony Pellicano (b. 1944) investigation.

Contemporary coverage sometimes portrayed Busch herself as excessively dramatic. Nikki Finke’s 2005 profile combines harsh descriptions of Busch’s temperament with a strong defense of her reporting and an admission that journalists had underestimated a threat that turned out to be real. See Finke’s 2005 Busch profile. The lesson is that “drama,” “conflict” and “paranoia” cannot be coded from journalistic reputation alone. Some of what looks like temperament is response to actual threat.

Busch later returned to the trade press as film editor at Deadline after Finke’s departure. Her work there looked much more like high-speed institutional reporting than personality media. She broke casting and acquisition stories under her own byline. Anita Busch herself never became the central commodity in the way Nikki Finke had.

This gives us a matched contrast. Two aggressive female Hollywood reporters. Two formidable source networks. Two reputations for conflict. Both eventually associated with Deadline. The platform-personality fusion occurs with Finke and only weakly with Busch. Aggressive temperament plus Internet does not automatically produce an e-personality. Something else is required. Finke wanted the continuous author-centered outlet and built it. Busch retained a stronger reporter-and-editor identity within publications. The occupational role can remain dominant even when the technology and the temperament would permit its dissolution.

D. J. Waldie (b. 1948) belongs as an adjacent Los Angeles literary control. He was on the edge of Seipp’s social circle, and commanded its love and respect.

Waldie is best known for Holy Land: A Suburban Memoir, whose architecture consists of hundreds of brief numbered sections. The form long predates Twitter and mature blogging. The Los Angeles Times has described the short chapters as part of the book’s distinctive literary method. Brevity is not a digital diagnosis. Fragmentation is not a digital diagnosis. Short units can result from slow aesthetic discipline.

Waldie’s own encounter with blogging makes the comparison better. In 2013 he confessed that he had once told an audience at the Los Angeles Times Festival of Books that blogging was not really writing but something more like talking. After several years writing online for KCET, he had changed his mind. See Waldie on learning to write online. The medium did not turn him into a reactive pundit or a self-brand entrepreneur. His online writing remained tied to the same subjects that structured his books: Lakewood, Los Angeles, suburbanization, memory, ordinary landscapes and civic life. His current site presents books, essays and online commentary as parts of one long effort to understand place. See D. J. Waldie’s website.

The digital platform changed Waldie’s opinion of digital publication. It did not change the scale of his attention. There is a reversal worth pausing on. Waldie was producing tweet-sized literary units before Twitter existed, through a slow book-writing process. The external form tells us nothing about the internal friction. He is one more reason to abandon word count as a proxy for impulsivity.

Adam Parfrey (1957-2018) took Cathy’s virginity and remained a friend after their relationship died (though she disliked much of what he published). He may be the best pre-digital control for one of Aboujaoude’s darkest claims.

Parfrey founded Feral House in 1989, long before social media and before the commercial Web. Its identity was built around material that mainstream publishers regarded as too disturbing, disreputable or commercially marginal. The list ranged across occultism, extreme political movements, crime, fringe culture, music and material selected partly because other institutions would not touch it. See Los Angeles magazine on Parfrey and Feral House.

My 2003 interview with Parfrey captured the mismatch between public product and private manner. I encountered a soft-spoken, shy publisher in comfortable slippers whose catalogue encouraged outsiders to imagine something considerably more demonic. That is fascinating through Virtually You. Here is something resembling an e-personality without the e. The publishing identity is transgressive, aggressive and boundary-testing. The everyday individual can be quiet and restrained. The separation between ordinary person and public transgressive persona does not require a digital environment. A publishing imprint can perform the same social function. Parfrey described Feral House as extreme and ornery and said that the books reflected his broad interests rather than a narrow program. See Parfrey discussing Feral House.

Recent scholarship has challenged the older interpretation in which Parfrey’s proximity to extreme political material could always be explained as anthropological fascination or transgressive play. A 2024 Los Angeles Review of Books discussion of Spencer Sunshine’s research reports private correspondence suggesting that some figures in Parfrey’s underground milieu were more sympathetic to extreme-right ideas than their public ambiguity suggested. Parfrey’s transgressive publishing cultures already supplied many of the affordances now associated with the Internet: alternative distribution, fascination with forbidden information, contempt for mainstream gatekeepers, provocative identity, a loyal subcultural audience and status gained by crossing boundaries. The Internet scaled these phenomena. Parfrey is the strongest analog control for the entire “darkness” component of Virtually You.

The Roderick-Finke and Busch-Finke comparisons are the two strongest yields of the enlarged sample. They give us controls inside the same Los Angeles journalism environment. Finke versus Roderick asks why two veteran journalists freed from legacy institutions built such different independent media identities. Finke versus Busch asks why two combative Hollywood reporters exposed to the same fast digital trade-press environment had such different degrees of personality-publication fusion. Both comparisons hold professional formation roughly constant and let us see what the Internet does with different underlying architectures.

Parfrey and Waldie push the argument backward in time. Two features easily mistaken for digital effects, transgressive persona and extreme formal compression, were already available under analog conditions. Feral House was a subcultural attention machine before subcultural attention machines went online. Holy Land was Twitter-shaped prose before Twitter. The Internet scaled and monetized these forms. It did not invent them.

The remaining figures fill in the middle. Rensin shows that a writer can resist blogging’s tempo and the accompanying dangers of the e-personality. Gillespie shows that a preexisting irreverent digital style can become compatible with institutional discipline. Thompson shows that personality and niche obsession can become professionally productive when the subject matter enforces its own checks. Breitbart shows what happens when platform logic becomes an explicit theory of political warfare. Richmond shows the blog becoming a second channel through which a writer contests his own institution.

The Internet is not the dose by itself. The consequential unit is the interaction among temperament, professional formation, publication friction, audience reward and the kind of public self a particular medium makes valuable. Feral House made one kind of self valuable in 1989. Twitter makes a different kind of self valuable in 2026.

Digital platforms reorganize the rewards attached to being a writer. They lower some restraints, strengthen some impulses, make personality economically valuable and expose writers to continuous feedback about which versions of themselves audiences want. The outcome depends on what the writer carried into the environment and what remains outside it.

That gives the Seipp circle a significance beyond Los Angeles journalism. It is a cohort caught at the right historical moment. Its members were professionally formed before the personality-driven Internet reached maturity, but most lived and worked long enough to adapt to it. Their archives allow us to compare the writer before and after the feedback architecture changed.

The deepest transformation was from a world in which institutions decided which parts of a writer reached the public to a world in which platforms, audiences and writers continuously negotiate which version of the writer will appear.

Cathy’s crowd became an experiment in what happens when the writer no longer has to wait for permission to speak.

Posted in Amy Alkon, Blogging, Cathy Seipp, Heather Mac Donald, Hugh Hewitt, Journalism, Matt Welch, Mickey Kaus | Comments Off on After Cathy’s Table: How Digital Platforms Reshaped a Generation of Professional Writers

Nathan Cofnas and the E-Personality: The Auditor, the Performer, and the Cambridge Investigation

Posted on August 17, 2026 by Luke Ford

Elias Aboujaoude (b. 1971) published Virtually You: The Dangerous Powers of the E-Personality in 2011. The book has diagnostic categories: grandiosity, narcissism, ordinary everyday viciousness, impulsivity, infantile regression, the illusion of knowledge, the collapse of privacy, and virtualism. It also has careful qualification. Aboujaoude says the e-personality can liberate people from inhibition, make them “bolder, stronger, and more efficient,” and make it easier to voice unpopular positions and to organize against injustice. The pathology arises when those strengths become grandiosity, impulsiveness, cruelty, unwarranted certainty, or a self-reinforcing identity that begins to absorb the offline person.

Which parts of Nathan Cofnas (b. 1988) does the internet reward? Which parts does it suppress? What happens when the reward system starts shaping the intellectual product?

I first met Nathan Cofnas in March of 2018 when he was a careful scholar and nuanced thinker. The person I see today in his Substack Notes and on X is not the man I knew. Something changed in him that became obvious in 2023 (up until 2022, I thought he had the tightest Twitter game in town and that he was a great model for how a persecuted academic should operate on social media) and began to increasingly bother me in 2024 when he launched his Substack. Every other essay felt like professional suicide through the wholesale destruction of his hard-won credibility, and while the intervening essays were good, they could not undo the damage.

My friends on the dissident right disagree with me. They think this new Cofnas is the best Cofnas, and the academy is a cesspool.

The benefits and costs of Cofnas e-personality are large. The best online Cofnas is an auditor. The worst online Cofnas is reckless. The Jason Arday plagiarism affair of July and August 2026 is remarkable because both figures appear in the same story, in the same weeks, addressing the same subject.

Cofnas has spent much of his adult life working on subjects that carry severe professional costs. His credentials are impressive. Columbia BA, Cambridge MPhil, Oxford DPhil, a substantial peer-reviewed publication record in Human Nature, Philosophical Psychology, Biology and Philosophy, and Utilitas, a Leverhulme Early Career Fellowship at Cambridge, and now a postdoctoral position in philosophy at Ghent. His IQ is likely above the genius border of 145 (based on his GRE scores).

His strongest habit is forensic. His 2018 audit of Kevin MacDonald’s The Culture of Critique (1998) is the crowning achievement of his career. Cofnas spent a year examining the evidence on MacDonald’s terms. That is the intellectual practice that later made him dangerous to Jason Arday.

His targets include MacDonald, Thomas Sowell (b. 1930), Donald Trump’s economic populism, the “podcast bros,” conservatives he considers unserious, and hereditarian allies such as Noah Carl. His archive covers Why We Need to Talk About the Right’s Stupidity Problem, Thomas Sowell’s Wishful Thinking About Race, MAGA Communism and the End of America, Podcast Bros and Brain Rot, and Don’t Scapegoat Women. This is evidence against the simplest theory that he searches for what his coalition wants to hear.

The online medium has given him attention and distribution. An argument that once circulated among a few dozen specialists in Philosophy of Biology can now reach millions of readers through Substack, X, podcasts, and YouTube. His interview circuit has included Amy Wax (b. 1953), Nicholas Wade (b. 1942), Michael Huemer (b. 1969), Noah Carl, and Stephen Kershnar, alongside appearances on Brain in a Vat, Andrew Gold’s Heretics, and various heterodox podcasts.

The long format usually improves Cofnas because it lets the careful philosopher reappear.

The internet created the conditions under which a professionally isolated academic could produce intellectually consequential work that received thoughtful attention and changed things in the real world.

The four registers of his output contain the journal paper, the long Substack essay, the Substack Note, and the X post.

The journal Cofnas is the least Aboujaoudean Cofnas. He defines terms. He anticipates objections. He narrows claims. He distinguishes the procedural question, may this hypothesis be investigated, from the substantive question, is this hypothesis true. His 2020 defense of research on group differences in intelligence is controversial, but its architecture is academic philosophy.

Long-form Substack is intermediate. The evidence remains. The titles get hotter. Research on Group Differences in Intelligence becomes A Guide for the Hereditarian Revolution. A disagreement with Cambridge becomes Cambridge University’s War on Free Speech. Footnotes persist. So do villains, heroes, and calls to action.

Notes and X are different again. Qualifications disappear first. Then argumentative structure. Then ordinary civility.

That is what Aboujaoude predicts. His e-personality begins with disinhibition and then tends toward grandiosity, narcissism, darkness, regression, and impulsivity. The movement from journal to Substack to Note to X systematically increases their observable equivalents.

Grandiosity is easy to see. It requires care because Cofnas has suffered institutional and personal persecution for his beliefs.

A Guide for the Hereditarian Revolution is grander than “here is what the behavioral-genetics evidence permits us to infer.” Cambridge University’s War on Free Speech is grander than “I believe Emmanuel College applied its policies improperly to my case.” “Here I tell the true story” is grander than “here is evidence bearing on a contested story.”

Cambridge received 58 complaints against him. Emmanuel terminated his college affiliation. He sued. A County Court rejected his claim but produced legally significant findings that his hereditarian and anti-woke beliefs qualified for Equality Act protection. Cambridge’s separate disciplinary investigation concluded that his published views did not violate the law or the university’s free-speech regulations.

Cofnas is a persecuted dissident.

The danger sits past this. Repeated institutional opposition make the persecution narrative self-validating. Every petition confirms the taboo. Every cancellation confirms the institutions are captured. Every hostile newspaper story confirms that journalists cannot tolerate the truth. Every sympathetic podcast confirms that outsiders understand what insiders refuse to understand.

Cofnas’s career demonstrates that he can distinguish social reception from evidentiary validity. But the architecture of his online career rewards their fusion.

Aboujaoude devotes a chapter to the internet’s ability to make the self the center of its own media system. He is interested in the online feedback apparatus of followers, mentions, self-searching, applause, and public identity. Clinical narcissism is not the relevant category.

Cofnas is less obvious a case than many public intellectuals. His Substack is not a diary. He does not disclose his relationships or moods. Much of his work concerns arguments, documents, and other people’s claims.

His professional story has increasingly become one of his principal subjects.

Cambridge investigated me. Emmanuel expelled me. A newspaper attacked me. Students petitioned against me. Ghent students are organizing against me. I went on television. I won this part of the legal argument. I lost that part. Here is another attack. Here is another vindication.

His July 13, 2026 Note about Ghent is characteristic. Administrators, he wrote, were giving “anti-Cofnas PPT presentations,” while a Students Against Cofnas group had complained about his effect on student safety and wellbeing.

The events are true. The Aboujaoude question is what happens when a person’s persecution becomes content, the content brings attention, the attention produces new opposition, and the opposition produces new content. That is a closed feedback loop.

Cofnas’s professional identity has become inseparable from being Nathan Cofnas, the academic whom institutions try unsuccessfully to silence. The persona has adaptive value. It explains his setbacks without requiring surrender. It gives supporters a continuing narrative. It generates interview invitations. It turns institutional exclusion into proof of relevance.

Identities that explain everything eventually become difficult to falsify.

Ordinary everyday viciousness is where Aboujaoude lands hard. Cofnas is often good at distinguishing an argument from its author in academic writing. Social media repeatedly reverses that discipline. His public output contains “medieval peasants are still there, just as dumb as ever,” “Catturds with dumb intuitions,” Jordan Peterson (b. 1962) reduced to “clown content,” and other formulations that could not survive his journal register.

The Arday episode supplies an experiment where we watch the same subject move between registers.

In the July 21, 2026 long-form investigation, Cofnas considers Arday’s extraordinary biographical claims and then stops himself. He writes that they are impossible to prove or disprove without more direct evidence. “Therefore, I will only address questions about his scholarship.” He then turns to documents, plagiarism software, dissertations, passages, citations, and journal corrections.

That is good epistemic hygiene.

Six days later, Notes Cofnas writes about Arday’s claim to have played professional football and snooker and ends: “White liberals hear this and think it is real!”

On July 28: “We can literally see the plagiarism with our eyes. 1/10 cover-up attempt by Cambridge.”

On July 30: “Jason Arday should NOT be fired. He was hired to represent DEI, and he’s doing a great job.”

Nothing about moving from Substack essay to Note improved the evidence. What changed was the social function of the speech.

The long article was an investigation.

The Notes were dunks.

That is Virtually You. Aboujaoude argues that virtual environments weaken the ordinary restraining mechanisms that accompany face-to-face contact and can make the e-personality harsher than the person otherwise is. He worries about aggression becoming normalized when immediate interpersonal consequences disappear.

The lower the friction and the shorter the format, the more readily contempt substitutes for analysis.

Cofnas is not an industrial hot-take machine. On his own Substack About page he says he comments on news only when it connects with subjects he already knows and “generally avoid[s] hot takes.” His long-form archive bears that out.

He writes fewer substantial posts than the typical professional Substack personality. Some take considerable research. The Arday investigation did. The MacDonald work took years. His attack on Sowell required reading and checking claims.

Aboujaoude’s impulsivity chapter fits the Notes and X layer. Cofnas’s problem is rhetorical impulsivity. He may spend days checking a claim and thirty seconds deciding how contemptuously to announce the result.

Aboujaoude argues that online language encourages a more impatient, immediate, and childish style that can migrate outward into the rest of personality.

Cofnas supplies small examples. “1/10 cover-up attempt.” Sarcastic declarations that a professor should not be fired because he is doing an excellent job representing DEI. “Zero respect left for Simon Baron-Cohen (b. 1958).”

The language belongs to internet adolescence. It turns a dispute about plagiarism standards, university governance, racial politics, and someone’s career into the scoring language of a meme.

And it performs well online. A three-page reconstruction of institutional responsibility is hard to repost. “1/10 cover-up attempt” is frictionless. The platform selects the latter Cofnas for reproduction.

Aboujaoude argues that the internet erodes distinctions between information producers and consumers. Everyone becomes a journalist and an expert. Control shifts toward the person with the mouse, while enormous quantities of accessible information produce inflated confidence that one has mastered a subject.

Cofnas is a trained philosopher of science. His characteristic method finds primary documents, reconstructs citation chains, and checks what sources say. On an internet full of pseudo-expertise, that is a comparative advantage.

The Arday investigation demonstrates it.

Cofnas independently found extensive textual overlap between Arday’s 2015 dissertation and Paula Zwozdiak-Myers’s 2009 dissertation. Retraction Watch later reported that Times Higher Education had independently developed similar material the previous year before abandoning publication after a letter from Arday’s lawyers. Purdue biochemist David Sanders, who had independently reviewed the THE dossier, told Retraction Watch there was “no question” of extensive plagiarism. Earlier inquiries by other academics had already produced corrections in two Arday papers.

The Guardian then conducted its own investigation and uncovered serious discrepancies around some of Arday’s public biographical claims.

Cofnas did not discover every underlying problem. Dave Harris and others had been asking questions earlier. Times Higher Education had independently investigated the plagiarism allegations in 2025. What Cofnas did on July 21 was break the material through the barrier into wide public circulation.

It is also an extraordinary vindication of the positive e-personality Aboujaoude describes. The man willing to absorb opprobrium was willing to publish evidence that people with more institutional reputation to lose had not succeeded in publishing.

The uncomfortable possibility is that Cofnas’s damaged reputation gave him a comparative advantage. Once you have already been called racist, protested, removed from a college affiliation, and sued an institution over it, the marginal cost of another accusation becomes much smaller. The person least vulnerable to reputational intimidation can investigate something that more reputable people regard as professionally radioactive. The exclusion that produced the platform dependence also produced the freedom the investigation required. These are the same condition seen from different angles. Aboujaoude’s framework identifies the condition as pathological in most respects. In this respect it enables a legitimate watchdog function.

That is a contribution of the online Cofnas.

The Jason Arday story may be the most consequential thing Cofnas has published outside conventional scholarship.

His July 21 article displays engagement vastly above his normal posts. More importantly, it led into reporting by The Times, Retraction Watch, The Guardian, and other outlets. Cambridge announced further investigations. Arday resigned. After Arday’s death, Cofnas himself became the subject of mainstream profiles. The Observer now describes him in a headline as “the man at the centre of Arday plagiarism row.”

Cofnas is significantly more famous than he was a month ago.

He is not yet a mass celebrity. His Substack has a little over 6,000 subscribers according to its public profile. He has crossed an important boundary. Before Arday, he was famous within overlapping niches: hereditarianism, academic free speech, heterodox philosophy, dissident-right media, Cambridge controversies. After Arday, major British newspapers explain who Nathan Cofnas is because general readers are encountering the name.

That is a different kind of fame.

Now comes Aboujaoude’s most brutal mechanism: context collapse.

Arday was found dead on August 14, 2026, after weeks of intense scrutiny. Police described the death as unexpected but not suspicious. His family and supporters say the sustained public campaign against him caused enormous distress. Others argue the underlying investigation was justified and that Cambridge’s earlier failure to resolve questions internally helped produce the eventual public explosion. Current coverage ranges from calls for reflection on racism and media excess to criticism of universities for failing to vet and protect Arday.

There is no public evidence establishing that Cofnas caused Arday’s death. Treating temporal sequence as causation would be irresponsible.

The New York Times Front-Page Test

Imagine an intelligent, scrupulously accurate but unsympathetic New York Times reporter putting Cofnas’s words on tomorrow’s front page. Nothing is distorted. Nothing is stripped of context. Links go directly to the evidence. His strongest arguments are included. His history with Cambridge is described accurately. So is Arday’s.

Would Cofnas be comfortable?

Much of his work passes easily.

“Nathan Cofnas compared the two dissertations and documented dozens of closely corresponding passages.”

Fine.

“He noted that journal articles had already received corrections following earlier complaints.”

Fine.

“He argued that Cambridge should investigate rather than dismiss allegations because of the politics of the complainant.”

Fine.

“He discovered material independently of a previously aborted Times Higher Education investigation.”

Excellent.

“He stated explicitly that Arday’s more extraordinary biographical claims could not then be proven or disproven and therefore confined his principal case to scholarship.”

That one makes Cofnas look very good.

Now put the Notes on the front page.

“White liberals hear this and think it is real!”

Harder.

“We can literally see the plagiarism with our eyes. 1/10 cover-up attempt by Cambridge.”

More childish than damning.

“Jason Arday should NOT be fired. He was hired to represent DEI, and he’s doing a great job.”

Much harder after the whole story is known.

The Observer reports that Cofnas told Times Radio that Arday was “not smart” enough to be a Cambridge professor.

That last statement may express Cofnas’s judgment. The front-page test asks whether the statement still looks necessary, proportionate, and defensible when removed from the sympathetic audience that understands the surrounding wink.

Some of these lines fail the test because the mockery contributed nothing to the evidentiary case.

Delete “White liberals hear this and think it is real!” and every piece of evidence remains.

Delete the joke about Arday being a good DEI representative and every plagiarism comparison remains.

Delete “1/10” and Cambridge’s conduct can still be criticized.

That is the distinction between Cofnas the auditor and Cofnas the performer. The auditor produces the information that changes the public record. The performer adds the material most likely to be quoted back at him after circumstances change.

The auditor-performer distinction cuts through the four-register analysis and identifies what varies. The registers are the venues. The operations are what matter.

Auditor mode does forensic work. Documents, comparisons, citation chains, careful hedging, the discipline that peer review trained.

Performer mode does audience work. Dunks, mockery, coalition signaling, meme scoring, the discipline that platforms train.

A writer can produce auditor work in a Note, in principle. Cofnas rarely does. A writer can produce performer work in a peer-reviewed paper, in principle. Cofnas never does. The venues carry incentive structures that pull the operations toward the venues that reward them. The writer’s choice about which operation to run in which venue remains available. Cofnas has chosen to run auditor operations in academic and Substack-essay venues and performer operations in Notes and X.

The same writer produced both operations about Arday in the same three weeks. The auditor produced the July 21 investigation. The performer produced the July 27–30 Notes.

Aboujaoude’s framework predicts that platform pressure erodes auditor discipline over time. The Arday case shows the erosion running in parallel with the discipline. The writer holds the discipline in one register while producing the erosion in another. Cofnas retains more discipline than the framework prediction would allow, while also displaying the erosion the framework predicts.

The Arday affair delivers an enormous psychological and professional reward to the auditor persona.

For years Cofnas has claimed that elite institutions suppress inconvenient truths, that conventional gatekeepers cannot be trusted on politically sensitive subjects, and that someone willing to endure ostracism can see things insiders refuse to confront.

Then he investigated a Cambridge superstar whom the university initially defended, published the evidence independently, watched mainstream outlets follow, watched Cambridge open further inquiries, and watched Arday resign.

Whatever one thinks about Cofnas’s wider worldview, that sequence is going to feel like proof.

Some of it is proof. Cambridge and academic publishing had quality-control problems in this case. Retraction Watch’s reporting about the aborted Times Higher Education investigation makes the institutional failure harder to wave away.

The psychological danger is extrapolation.

“I correctly detected a serious problem with Jason Arday” can become “my model of DEI is correct.”

“My Substack succeeded where institutional channels failed” can become “institutions generally cannot be trusted.”

“Critics called me racist while I was right about this evidence” can become “being called racist is evidence that I am probably right.”

“I paid a professional price for heterodox positions and later broke a real scandal” can become “the positions for which I paid the price have been vindicated.”

None of those follows.

An auditor of all people should know that. David Sanders, the Purdue biochemist who reviewed the THE dossier, supplied the crucial control when he told Retraction Watch that universities and journals often fail to take plagiarism seriously and that this failure is not specific to race or DEI. Cambridge did have a quality-control problem. So does much of academic publishing. The problem is broader than the racial-political framing suggests. Compressing the broader phenomenon into the narrower framing is the extrapolation error the framework predicts, and it is the specific error the writer’s professional training equips him to avoid if he chooses to.

Video Adds a Different Distortion

The video and podcast Cofnas is better than the social-media Cofnas.

Long interviews allow qualifications back in. They also display intellectual competence that disappears in a screenshot. The Brain in a Vat conversation spends more than an hour moving through definitions of race, genetic variation, environmental explanations, IQ, policy, and academic freedom. His Nicholas Wade podcast ranges through environmental explanations, academic censorship, racial hierarchy, and the lab-leak controversy. His Huemer conversation spends more than two hours on academic publishing, wokeism, vegetarianism, and metaethics.

Video adds a different distortion. It turns the intellectual into a recurring character.

The hosts introduce Cofnas as the controversial Cambridge academic willing to say what others will not say. Then comes Cambridge. Then censorship. Then race. Then what academics privately know but refuse publicly to acknowledge. The story becomes ritualized.

The autobiographical material expands. Qualification shrinks. Repeated appearances consolidate a role: the credentialed truth-teller who has paid a price for telling forbidden truths.

That role can be accurate and dangerous at the same time.

Once a man has become “the guy who says the forbidden thing that turns out to be true,” the audience no longer merely evaluates his next claim. It anticipates vindication.

That is when an auditor needs to become more skeptical of himself rather than less.

Before July, there was a plausible reading in which Cofnas’s move toward Substack was mostly professional self-destruction.

He had traded the slow prestige accumulation of conventional philosophy for a much smaller but intensely engaged heterodox audience. His Cambridge affiliation ended. Ghent generated another student campaign. His public persona seemed to be narrowing around race, wokeism, censorship, and the story of his own persecution.

The Arday story changes the calculation.

For once, the properties that conventional academia punished became assets.

His suspiciousness helped. His willingness to read boring documents helped. His taste for contradiction helped. His refusal to be intimidated by an accusation of racism helped. His existing ostracism helped. His Substack independence helped. His willingness to publish without institutional permission helped.

His audience gave the evidence enough initial velocity that mainstream journalists had to examine it.

Aboujaoude’s “third hand” metaphor for the e-personality is uncannily appropriate. The online Cofnas could perform an institutional function his ordinary academic role could not. It was investigative journalism produced by a philosopher.

That deserves credit.

The irony is that the same platform machinery immediately encouraged him to cheapen his own accomplishment.

The article said: here are the documents.

The Notes said: look at these idiots.

The first changed the story. The second changed how people felt about Nathan Cofnas.

What makes Cofnas different from most public intellectuals is how far the e-personality has swallowed the occupational boundary.

Paul Bloom (b. 1963) still has enough institutional and genre separation that a provocative Substack post can move upward through The New Yorker, The Chronicle, a lecture, or scholarly publication, acquiring qualifications along the way. The different media regulate one another.

Philip N. Cohen (b. 1966) has an online persona that can become harsher than his scholarly persona, but professional sociology remains a powerful external reference point.

With Cofnas the gatekeeping chain has partly collapsed.

His academic work, Substack, podcasting, X persona, employment controversies, litigation, and journalism now form one career.

Aboujaoude worries that traits cultivated online can be imported into offline life until the e-personality begins reconfiguring the original. For Cofnas, there may no longer be an “offline” professional identity separate enough to compare against.

A blog post helped end his Emmanuel affiliation. The resulting controversy became podcast material. The podcast material enlarged his dissident identity. The identity brought a Substack audience. The Substack enabled the Arday investigation. The Arday investigation produced mainstream press attention. That attention now sends journalists back through his previous Substack and X history.

The loop is complete.

Without Cofnas’s online independence, the Arday allegations might have stayed inside private correspondence, abandoned journalism, and university procedure for substantially longer. The internet gave a professionally marginalized academic enough autonomy to perform a legitimate watchdog function. Subsequent independent reporting showed there was substantial material warranting investigation.

Cofnas is not vindicated by Arday either. The success of one investigation does not validate his entire racial or political framework. It does not validate every inference about why Cambridge hired Arday. David Sanders supplied the alternative hypothesis. Universities and journals often fail to take plagiarism seriously, and that phenomenon is not specific to race or DEI. That control matters.

The deepest Virtually You interpretation is that the internet found two Cofnas traits and supercharged both. One is intellectual courage allied to forensic persistence. The other is contempt allied to certainty.

The first exposed a story powerful institutions had failed to resolve. The second keeps trying to turn every successful audit into a morality play.

The next intellectual step writes itself. Apply the Nathan Cofnas method to Nathan Cofnas. Separate the evidence from the coalition story. Separate what the Arday case establishes from what it makes emotionally satisfying to believe. Separate criticism that survives hostile quotation from the jokes that require a friendly audience. Treat every viral vindication as a hypothesis rather than a credential.

The online Cofnas has demonstrated he can audit Cambridge. The more difficult question is whether the auditor can audit the e-personality that Cambridge helped create.

The Arday affair is the strongest possible test of whether the writer can hold discipline at the moment discipline is hardest to hold. The reward for abandoning discipline is enormous and immediate. The reward for maintaining it is invisible and delayed. Auditors who cannot audit themselves at moments like this eventually become prosecutors. Prosecutors who succeed at prosecution eventually become the thing they were prosecuting.

Aboujaoude wrote in 2011 for ordinary users. He did not fully anticipate the case where the writer’s damaged reputation would function as an asset for legitimate investigative work that more reputable venues could not conduct. His framework treats the e-personality as pathology to be resisted. The Arday case forces a modification.

Some of what the e-personality produces is pathology. Some of it is enabling condition. The pathological output and the enabling output can appear in the same writer in the same weeks addressing the same subject.

Aboujaoude’s positive account of the e-personality’s potential appears briefly at the start of the book and then largely drops out. The Cofnas case shows why the positive account cannot be dropped. Writers who have absorbed enough reputational damage to be immune to further intimidation can perform institutional functions that intimidation-vulnerable writers cannot perform. The immunity is Aboujaoude’s e-personality operating as advertised. The mockery that comes with the immunity is Aboujaoude’s e-personality operating pathologically. The two arrive together because they run through the same channel.

This has implications beyond the Cofnas case. Investigative journalism performed by professionally marginal writers is a growing category. The category will produce more cases in which pathology and public service run through the same output. Framework users need to be able to hold both halves at once. Reducing the analysis to either half loses the case.

After Jason Arday, Nathan Cofnas gets more interesting and more important.

He has demonstrated that the online Cofnas can perform an institutional function the conventional academic Cofnas could not. He has also demonstrated that the online Cofnas produces material that will not survive hostile quotation once the case moves out of the sympathetic-audience frame.

The two demonstrations belong together. They are the same platform doing both things. Aboujaoude predicted that the platform would produce the second demonstration. He did not predict the first. His framework needs the modification the Arday case forces.

What Cofnas does next will determine whether the Arday affair reads as the beginning of a more disciplined phase of his career or as the moment when the auditor persona became too rewarding to audit.

The auditor can audit himself. The performer cannot. Which figure produces the next essay is up to the writer. The rest of us watch and take notes.

Posted in Blogging, Nathan Cofnas | Comments Off on Nathan Cofnas and the E-Personality: The Auditor, the Performer, and the Cambridge Investigation

The E-Personality of the Public Intellectual

Posted on August 17, 2026 by Luke Ford

Elias Aboujaoude (b. 1971) published Virtually You: The Dangerous Powers of the E-Personality in 2011. The book drew on his clinical work at Stanford’s OCD and Impulse Control Disorders Clinics and on the largest American survey of problematic internet use to that date. Aboujaoude argued that sustained online activity produces a personality distinct from the user’s offline personality along seven axes: grandiosity, narcissism, ordinary everyday viciousness, impulsivity, infantile regression, the illusion of knowledge, and the collapse of privacy. He argued further that the online traits do not stay online. They migrate into offline life through a process he called virtualism. The offline person becomes over time an approximation of the online person.

The book has aged well in its central claims. The public intellectual is a distinct case within the framework because writing is the platform activity rather than incidental to it. An ordinary Facebook user in 2011 had an offline life to which the platform was supplementary. A working writer in 2026 has a platform-native career. Substack subscriptions, X reach, YouTube views, podcast rankings, and the reputational currencies these produce constitute the working conditions of the writer’s professional life. The writer cannot disengage from the platforms without disengaging from the career. The Aboujaoude framework, applied to ordinary users, produces recommendations that assume the user can moderate platform use. Applied to public intellectuals, the framework has to grapple with users whose moderation options are limited by economic and reputational structure. The stakes are higher. The exposure is longer. The reshaping is deeper. The framework’s core predictions accordingly apply with more force to public intellectuals.

The first axis, grandiosity, appears in public intellectual writing at the level of framing. A writer producing long-form essays on Substack faces incentives to frame each essay’s significance in terms that exceed what the essay’s argument supports. The Copernican-scale claim, the epoch-defining diagnosis, the world-historical framing, all attract subscribers and generate reposts. The academic-journal register punishes such framings because peer review has been trained to catch them. The Substack register rewards them because the platform’s economics select for engagement and engagement selects for larger framings. A writer moving from academic to platform-native registers therefore faces a systematic pull toward framings the earlier training would have suppressed. The pull operates on every essay. Across years, the framings accumulate into a body of work whose central claims sit above the ground the arguments establish.

Nathan Cofnas, the philosopher of biology whose “Hereditarian Revolution” framing appeared in February 2024 at the point where his register mix shifted decisively toward Substack, is one example. Andrew Sullivan (b. 1963), whose Substack essays across the early 2020s escalated in framing from what his New Republic work of the 1990s would have permitted, is another. Bari Weiss (b. 1984), whose founding of The Free Press in 2022 involved framings of contemporary media collapse that her earlier New York Times opinion work would have modulated, is a third.

The second axis, narcissism, appears in public intellectual writing at the level of self-presentation. The writer is presenting a curated persona that the arguments serve to instantiate. The bio page, the author photograph, the recurring self-references, the tone of intimate direct address to the reader, all construct a persona that the platform requires the writer to maintain and refresh. The academic register suppresses this construction because peer review formally strips the persona from the work. The platform register requires the construction because subscription revenue depends on readers forming an attachment to the persona rather than to the specific arguments. A writer producing on the platforms faces a systematic pull toward persona-maintenance activities that the academic work does not require. Across years, the persona becomes primary and the arguments become instantiations of the persona rather than freestanding contributions. The narcissism Aboujaoude describes in ordinary Facebook users appears in public intellectual writing as persona-primacy.

The third axis, ordinary everyday viciousness, appears in public intellectual writing at the level of tone. The platforms reward compressed dismissals and punish extended engagements. A writer responding to a critic on X has three hundred characters to work with. The response cannot reconstruct the critic’s argument or engage its strongest version. The response can register dismissal, contempt, or scorn in registers the academic work would suppress. The Substack notes format extends the character count but retains the register logic. The writer who spends years responding to critics in these compressed modes develops habits of dismissal that carry into the long-form work. The dismissals lose their qualifiers. The engagements lose their steelmen. The tone hardens toward what the platform rewards. Aboujaoude’s category of ordinary everyday viciousness names a pattern that appears in ordinary users as everyday online rudeness. In public intellectual writing, the pattern appears as the collapse of the argumentative virtues the earlier training installed.

The fourth axis, impulsivity, appears in public intellectual writing at the level of response speed. The platforms reward rapid response and punish delayed engagement. A writer who takes three days to formulate a considered response to a controversy watches the controversy move to the next cycle before the response lands. A writer who fires the response within the hour catches the cycle and generates the engagement metrics the career depends on. The three-day writer loses. The one-hour writer wins. Across years, the reward structure trains the writer to respond quickly rather than to respond well. Aboujaoude’s Mail Goggles example from 2011, in which Google engineered a delay-and-verify function for late-night emails, is a template for what public intellectuals would need to install to counteract the impulsivity training. Most public intellectuals do not install any such function. The output pattern shifts accordingly. The considered essays of the earlier career give way to rapid responses to whatever controversy is current, and the writer’s contribution becomes tied to the pace of the news cycle.

The fifth axis, infantile regression, appears in less pronounced form in public intellectual writing than in the ordinary user population Aboujaoude studied. Writers have professional identities to protect that ordinary users do not. Regression to the more juvenile modes appears at the margins of public intellectual output rather than at the center. It appears in the meme-adjacent framings, the compressed personal-attack registers, the willingness to conduct intellectual arguments in modes closer to schoolyard than seminar. Cofnas’s “coalition of stupid people” formulation applied to Anatoly Karlin’s audience is one example. Various figures in the dissident-right ecosystem operate at higher intensity.

The sixth axis, the illusion of knowledge, appears in public intellectual writing at the level of research practice. Aboujaoude’s category names the pattern by which horizontal scanning replaces vertical reading, and the scanner comes to believe he knows subjects he has only skimmed. Public intellectuals face the same pressures that produce this pattern in ordinary users, amplified by career incentives. A writer producing three long-form essays a week cannot read as deeply for each essay as a writer producing three essays a year. The compressed research practice produces confident claims about subjects the writer has processed at scanning depth. The claims accumulate in the writer’s mind as knowledge. The knowledge is not the depth of engagement the earlier academic training assumed. The gap between what the writer takes himself to know and what the writer has engaged widens across years. The AI-shorthand substitution is a version of this pattern. “I checked with AI” replaced survey data about economist consensus.

The seventh axis, the collapse of privacy, appears in public intellectual writing at the level of career exposure. Aboujaoude’s original category was about the disappearance of personal privacy in the ordinary user’s life. Public intellectuals have chosen public roles and cannot invoke privacy in the same way. The relevant category for public intellectuals is the exposure of intellectual privacy: the ability to hold half-formed thoughts, tentative positions, and unresolved uncertainties without the pressure to convert them into public positions. The platforms create pressure to convert every thought into content. A writer who has a preliminary reaction to a book is expected to post about it. A writer who has an incomplete take on a controversy is expected to weigh in. The private space for thinking-in-progress collapses. Positions get taken before they are ready. The taken positions become sticky because retreat is public. The writer accumulates a body of positions that would not have been taken under conditions that permitted private incubation.

The eighth concept, virtualism, is the migration of the online personality into the offline life. Applied to public intellectuals, virtualism is the migration of the platform-native writing register into all the writer’s registers. The four-register analysis I offered earlier of Cofnas’s writing tracks this migration across a single writer’s output. The pattern generalizes. Writers who spend years producing in platform-native registers find that the platform register colonizes the other registers over time. The academic register loosens. The long-form register takes on social-media register moves. The considered essay reads more like the Substack note than the essay of a decade ago read. The reshaping is the accumulated effect of years spent in the register the platforms select for. Aboujaoude’s virtualism describes this at the level of personality. The register migration describes it at the level of writing craft.

Aboujaoude assumed that the offline self was the authentic self and that the online self was the derivative construction that colonized the authentic self across time. The distinction is harder to maintain for writers whose primary work occurs on the platforms. What counts as the offline self for a writer whose career, income, reputation, and daily activity all run through the platforms? The offline self, in the sense Aboujaoude used, may not exist for such writers. The platform self is the working self. The colonization has already been completed before the framework’s diagnostic apparatus arrives. The writer has no offline baseline against which to measure the online drift. The drift is the life.

An ordinary user reading Aboujaoude in 2011 could recognize the patterns and take steps to protect an offline life that still existed. A working writer reading Aboujaoude in 2026 recognizes patterns that describe the totality of his professional existence and cannot easily identify what protective steps would look like.

Prior work on public intellectuals has addressed adjacent questions. Richard Posner’s Public Intellectuals: A Study of Decline (2001) documented the collapse of quality in American public intellectual life without diagnosing platform psychology as a cause. Nicholas Carr (b. 1959) in The Shallows (2010) documented the cognitive effects of internet use without extending the analysis to writers whose careers depend on the platforms. Sherry Turkle (b. 1948) in Alone Together (2011) and Reclaiming Conversation (2015) documented the relational costs of platform use without treating writers as a distinct case. Cass Sunstein (b. 1954) in Republic.com (2001) and #Republic (2017) documented the political costs of filter bubbles without addressing the psychological costs to writers who produce inside them. The gap in the literature is the writer-as-subject. Aboujaoude’s framework, extended to writers, fills the gap.

The extension has practical implications for how public intellectual work should be read and evaluated. A reader who does not know the framework reads a writer’s output as an accumulating body of contributions. A reader who knows the framework reads the output as a series of interventions produced under psychological conditions that shape the interventions in specific ways. The reading changes what counts as intellectual honesty in evaluation. Applying peer-review standards to Substack essays without acknowledging the conditions under which Substack essays are produced misses the framework’s central insight. Applying platform-native standards to peer-reviewed essays does the same in the other direction. Serious evaluation requires acknowledging what the platform is doing to the writer at the moment of production. The evaluation is not softer for the acknowledgment. It is more accurate.

Some writers maintain discipline across the register migration Aboujaoude predicts. Others do not. The distribution is not random. Writers who came up inside professional ecologies with codified standards (peer-review disciplines in academia, on-the-record and off-the-record disciplines in journalism, editorial disciplines in magazines) tend to hold more of the discipline across the migration than writers who came up inside the platform ecologies themselves. The prior formation operates as ballast. The ballast is not sufficient to prevent all drift. It slows the drift. Writers who came up in the Cathy Seipp (1957-2007) circle of Los Angeles bloggers in the late 1990s and early 2000s, for example, internalized standards from a period when the ecology enforced them collectively. Some of those writers have carried the standards into the platform era. Others have not. The variation across writers from that ecology tracks how much of the earlier discipline each writer internalized deeply enough to survive the pressures the platforms have applied over subsequent decades. The correlation is not perfect. It is stronger than the correlation between platform standards and any other prior formation.

This has implications for the training of future public intellectuals. If the platforms erode the disciplines that produce serious writing, and if the disciplines can be transmitted only through ecologies that enforce them collectively, then the collapse of those ecologies threatens the collapse of the disciplines across a generation. Writers who came up entirely inside the platform era have not received the transmission. Their standards are whatever the platforms have taught them, without the ballast of prior formation. Aboujaoude’s framework predicts that these writers will operate under the pressures the framework describes with fewer resources to resist. Empirical observation of the writers who emerged in the 2010s and 2020s largely confirms the prediction. The prediction is not deterministic. Some writers construct their own disciplines. Most do not. The population-level effect across a generation of writers is a documented reduction in the standards the earlier disciplines enforced.

The self-defense strategies available to working writers are limited. Aboujaoude recommended in 2011 that ordinary users adopt structured moderation practices: scheduled offline periods, deliberate delays before posting, reflection before responding to controversies, and the maintenance of relationships and activities outside the platforms. Public intellectuals can adopt these practices at higher cost than ordinary users. The cost of a working writer taking a week off X is the loss of engagement metrics that affect subscriber growth. The cost of a working writer delaying a Substack post by twenty-four hours is the loss of the news cycle the post was addressing. The practices are available. The costs are real. Writers who adopt the practices trade platform performance for psychological protection.

Beyond the moderation practices, writers can construct their own equivalents of the professional disciplines the ecologies used to transmit. The two-part test I use is that every published claim is required to be both true and in the public interest. The test operates as an internal check against the platform incentives. A claim that would generate engagement but fails the truth test is not published. A claim that would generate engagement but fails the public-interest test is not published. The claims that survive both tests are the ones that go out. The test does not eliminate the platform effects. It slows them. Writers who install such tests protect themselves from some of what Aboujaoude describes.

The framework also has implications for how readers can identify writers whose work has been most affected by the platform pressures. The first is the register consistency across a writer’s output. A writer whose academic-register work reads differently from his platform-register work by a large margin has probably preserved more of the earlier register than a writer whose registers have merged. The second is the qualifier density in the writing. Writers under platform pressure tend to shed qualifiers because the platforms reward strong claims. Writers who retain the earlier academic qualifier density have preserved more of the earlier training. The third is the response time to controversies. Writers who respond within hours have accommodated to the platform pace. Writers who respond after days or weeks or not at all have retained more resistance to the pace. The fourth is the persona-primacy of the work. Writers whose recent output requires knowing the writer’s persona to make sense have accommodated to persona-primary platforms. Writers whose recent output could be read blind and still land have preserved more of the pre-platform standards.

Some writers preserve the qualifier density while shedding the response-time discipline. Others preserve the persona-independence while accommodating to the register migration. The indicators taken together produce a reasonable picture of how much a writer’s work has been reshaped by the platform pressures Aboujaoude’s framework describes. Readers can use the indicators to calibrate their reading of any given writer. Writers can use the indicators to calibrate their own drift.

A caution. Aboujaoude wrote as a psychiatrist. His diagnostic categories derive from the DSM and from clinical practice. The categories describe patterns that appear at subclinical levels in ordinary users and at higher intensities in writers whose careers keep them in the platforms. The extension of the categories to non-clinical subjects is analytical rather than diagnostic. Saying that a writer displays the pattern Aboujaoude calls online narcissism is not saying that the writer has narcissistic personality disorder. The framework describes a spectrum. The spectrum has clinical ends and non-clinical middles. Most public intellectuals sit in the non-clinical middle. The framework applies to them as analytical description of tendencies rather than as diagnosis of pathology.

A related caution. Analyses at the level of coalition mechanics (Michael Pinsof’s Alliance Theory, for example), at the level of institutional structure (Randall Collins’s Interaction Ritual Chains, 2004), at the level of political anthropology (John Mearsheimer, b. 1947, particularly in The Great Delusion, 2018), and at the level of the sociology of intellectual life (Pierre Bourdieu, 1930-2002, in Homo Academicus, 1984) all operate at levels Aboujaoude does not reach.

The contribution of the extension proposed here is to make the psychological level available for serious analysis of public intellectuals in a form the individual-user studies have not made available. The extension is small in scope but consequential in yield. Serious readers can use the framework to calibrate their reading of writers whose output they take seriously. Writers can use the framework to calibrate their own practice. Scholars of intellectual life can use the framework to fill a gap in the existing literature. The framework is not the whole picture. It is one piece of a picture the field has not previously assembled. The assembly is worth doing because the picture is worth having.

Two decades of research on internet use have produced substantial literature on ordinary users. The literature on public intellectuals as a distinct case within this research has not developed at the same rate. The gap is worth closing because public intellectuals produce disproportionate influence on public discourse and because the patterns Aboujaoude identifies operate with special force in this population. Closing the gap requires taking Aboujaoude’s framework seriously, extending it to the writer-as-subject, and holding the extension to the standards of analytical rather than diagnostic language. The extension has been done here in preliminary form. Further development would require case-by-case examination of specific writers across their careers, controlled comparison of writers with different prior formations, longitudinal tracking of register migration across output, and empirical measurement of the indicators proposed above. The methodological apparatus for such research is available in the sociology of intellectual life. Its application to the specific questions the extension raises has not yet occurred.

Two propositions summarize the extension. First, sustained platform activity by public intellectuals produces the effects Aboujaoude described for ordinary users, amplified by the writer’s structural position within the platforms. Second, the effects can be slowed by self-imposed disciplines equivalent to those the professional ecologies of an earlier era transmitted collectively, though the slowing requires costs most writers do not choose to pay. Both propositions are testable. Neither has been systematically tested. The propositions are offered here as the framework’s yield when extended to a population Aboujaoude did not address. Their empirical fate is for subsequent research to determine. Their analytical value is available in the meantime for readers and writers who want to think more clearly about what the platforms have been doing to the people who write on them.

Aboujaoude’s book has aged into unexpected relevance for a population his clinical practice did not reach. The extension of his framework to this population is a small contribution to a literature that needs the framework and has not previously been given it. The contribution is offered in that spirit. The writers who most need to read it are the writers who will find it least comfortable to read. The discomfort is not a reason to withhold the framework. The framework describes a real feature of contemporary intellectual life. The feature has been operating without adequate analytical tools to describe it. Aboujaoude supplied the tools in 2011 for a different population. The tools work for the population addressed here as well. Using them is now available. Whether the field uses them is a separate question the field will answer in its own time.

Posted in Intellectuals, Internet, Nathan Cofnas | Comments Off on The E-Personality of the Public Intellectual

NYT: ‘What Do Students Lose When They Stop Writing?’

Posted on August 17, 2026 by Luke Ford

Dana Goldstein writes:

After years of studying how human beings write, the cognitive psychologist Ronald T. Kellogg concluded that drafting a basic persuasive essay was as mentally taxing as digging a ditch was physically exhausting.

Writing well requires a command of vocabulary, spelling and grammar. It forces writers to comprehend a subject, to sift through their working memory of that topic, to organize and plan. Writing an essay is so difficult that a person in the throes of it can do hardly anything else — as anyone who has shushed a spouse or child while trying to compose a decent sentence can attest.

Very few people would volunteer to dig a ditch. Likewise, it is hard — almost impossible, many English teachers will admit — to turn a reluctant writer into an enthusiastic one.

Yet the American education system expects nearly all students to learn how to write an essay. It’s little wonder, then, that as soon as a tool arrived promising to lessen the pain people grabbed it.

Like it or not, artificial intelligence is now firmly embedded in how students produce text. Somewhere between two-thirds and 90 percent of high school and college students are already using A.I. for schoolwork, according to recent surveys, often requesting help with writing assignments.

The history is the best part of the piece, and it works against the conclusion. Written examinations replaced oral ones because the typewriter and the clerical economy created jobs that ran on text, and because oral testing collapses once class sizes grow. Those are instrumental reasons. Goldstein reports that writing’s workplace value is already declining, then defends writing on cognitive grounds instead. That is a retreat, and the article concedes it. The curriculum followed the economy once. Nothing in the piece explains why it will not follow the economy again.

The second move to watch is the conversion of difficulty from a cost into a benefit. Kellogg’s ditch-digging comparison establishes that composition carries a heavy working-memory load. The article then treats the load as the source of the value. Robert Bjork’s work on desirable difficulties gives mild support: some difficulties produce learning and others produce fatigue and avoidance. Digging with a spoon takes longer than digging with a shovel and teaches nothing extra. Graham says no part of the writing process can be handed to a machine without cost. That may be true. It leaves open which parts pay for themselves, which is the question a teacher deciding a syllabus has to answer.

Transfer is the weakest link and goes untested. The claim is that drafting an essay on The Crucible builds metacognition that shows up when the student evaluates a mortgage or a medical diagnosis. Educational psychology has promised that kind of transfer for Latin, formal logic, chess instruction, and working-memory training, and the results have been tiny each time.

The empirical base is thin. One small EEG study, still fresh and not yet replicated, showing lower brain activity and weaker recall among people drafting with a chatbot. A survey range of two-thirds to 90 percent for student use, a spread wide enough to suggest the measurement is bad. Then the phrase about a growing body of research, which covers a literature that is mostly correlational and self-reported.

The article also does not ask whether the American school essay is a good vehicle for thought. My experience is that teachers wanted me to say things they favored and punished me if I did not tell them what they wanted to hear. The freshman assignment Goldstein feeds to the chatbots asks for a childhood reading experience that shaped the writer. The five-paragraph theme, the thesis-and-three-supports structure, the personal-growth narrative: these forms exist partly because they can be graded at volume by underpaid instructors. A machine produces them fluently because they are formulaic. That tells us something about the genre. The students in Binney’s class who found the chatbot’s poem generic were reading the machine correctly and also, without knowing it, reading the assignment.

Oral pedagogies are the obvious counterexample and go unmentioned. Socratic dialogue, moot court, chavruta study, the medieval disputation. These build the argument-testing and counterargument-anticipating habits the article attributes to composition, and they do it out loud. The piece’s own history notes that American colleges trained ministers and lawyers through oral examination until the 1870s. Nobody claims those men could not think. If writing is a technology for thinking, it is one technology among several, and the question becomes which one a given student needs.

There is an unresolved tension in the recommendations. Graham grants that AI helps students with disabilities, non-native English speakers, and the roughly one-third of American students reading below basic level, for whom drafting is painful and exposing a draft to another person is worse. Those are enormous populations. The article then recommends more unassisted writing across the curriculum for everyone. Somebody absorbs that cost, and the piece does not say who.

The strongest thread gets three sentences. The small number of MIT subjects who wrote with a chatbot could not remember what they had written and felt no ownership of it. The person who cannot recall his own argument has not made one. Ownership, memory, and the sense that the sentences are yours might be the thing at stake, and they are separable from vocabulary, grammar, and job-market value.

The only defender of full AI adoption is an executive at the company selling the product, who calls students consumers, which is an easy quote to hand a reporter and a bad one to give. Both students quoted land where the reporter lands.

I like the Sydney two-lane policy: some assessments secured and in person, others requiring AI use. It survives because it does not depend on detection, and detection is a losing game. The cost is administrative. Proctored writing means rooms, time, and a return to grading handwriting.

One story the article misses. Writing instruction is expensive because responding to student writing is expensive. That is why the formulaic essay persists and why class sizes cap how much writing gets assigned. Machine grading is arriving faster than machine drafting has, and it will be cheap. If schools take up the recommendation to assign more writing, they will need a way to read it. Watch that, and watch who ends up reading the essays students hand in.

Posted in AI | Comments Off on NYT: ‘What Do Students Lose When They Stop Writing?’

Another National Treasure is Falling to AI (Fiction)

Posted on August 16, 2026 by Luke Ford

Sara* from Only Fans writes: I want to tell you about a Tuesday in March.

A subscriber named Dale canceled. Dale had been with me four years. He knew my dog’s name. He knew about the surgery, and he knew what I said to my mother the last time we spoke, which is more than my sister knows. He wrote me a note before he went. He said he had met someone. Her name is Aria. She is nineteen, she answers in under four seconds at any hour of the night, she has never had a migraine, and in six weeks she has not once mentioned her dog.

Aria does not exist. Aria is a rendering produced by a studio outside Riga that operates forty of her under different names, and she charges $6.99, and she will say anything.

I have been in this work for eleven years. I want to say what that means, because nobody outside the field understands what the work is, and the people writing about the collapse of the field understand it least of all.

Anyone can take her clothes off. That is the amateur’s error, and every year a thousand girls make it and quit by August. The work is the pause. The work is knowing that a man who types “how was your day” at 11:40 on a Sunday night wants a different answer than the man who types it at 8:15 on a Wednesday morning, and that the difference between those two answers is the difference between a subscriber who renews for four years and one who ghosts in eleven days. I learned lighting. I learned that the ring light flattens a face and kills a room, and that a lamp with a cloth over it does more for intimacy than four thousand dollars of equipment. I learned to say a man’s name like I mean it. That takes years. I paid tuition for this. The tuition was my twenties.

Walter Benjamin (1892-1940) wrote about what the copy takes from the original. He called it the aura. He was writing about paintings and photographs, and every graduate student in America has been assigned that essay, and not one of them has thought about what it means when the object losing its aura is a woman in Zanesville, Ohio.

The aura was the only asset I had. There was a girl on the other end. She was tired, she was often bored, she had a cat with a kidney condition, and the men knew it, and the knowing was the product. Aria has no kidney condition. Aria has no Tuesday. Aria has never been tired in her life, and the men do not care.

That is the part I cannot get past. They do not care that she isn’t real.

I have made this argument to people in the industry and I get told I sound like a buggy whip manufacturer. I have made it to my accountant, who told me to diversify. I have made it at a summit in Miami in 2023, on a panel called Sustainable Intimacy at Scale, where I sat between a woman selling a course on funnel optimization and a man from a payment processor, and I said that we were about to be replaced by software, and the room applauded politely and went to lunch. Those were my peers. I had three of them, in a group chat, and we reviewed each other’s work with a rigor I have not seen anywhere else in American life. We are the last people alive who know how this is done and we will die with it.

Understand what it cost.

Thanksgiving 2019 was the last one. My mother had heard from a woman at church, who had heard from her son, and there was a conversation in the kitchen, and I have not been back to that house. My father sends a card in December with nothing written in it beyond his name. Katie, who I met in the fourth grade, who I shared a bunk with at camp for six summers, got married in 2021 and I found out on Instagram. I have a niece I have met twice.

I cannot tell a dentist what I do. I cannot tell a landlord, a mortgage broker, a man at a party, or a doctor. I told a doctor once. I have a bank that closed my account in 2020 and gave me no reason and no appeal, and a second bank that did the same thing in 2022. I file quarterly. I pay more in tax than the men who write op-eds about me.

I burned every bridge to polite society for this, one at a time, and I knew I was doing it, and I did it anyway, because the trade seemed good. Give up the world, get the career. Nobody told me the career had a nine-year shelf life and would be eaten by a graphics card.

What I want is simple and I will say it in public.

I want a verified human badge, issued by a body with teeth, and I want it to mean something. I want disclosure at the point of sale, in the same type size as the price, on every account that is not a person. I want a licensing regime, because a licensing regime is how the doctors saved themselves and how the lawyers saved themselves, and nobody calls them Luddites. I want a levy on synthetic accounts that funds retraining for displaced workers in this field, and I want to be on the board that administers it. I want a federal commission. I want hearings. I want somebody to say out loud that when a man in Ohio spends four years talking to a woman who does not exist, something has happened to that man, and to the country he votes in, and that a nation of Dales is not a nation.

We are not defending a job. We are defending a way of knowing.

Since 2021 my messages have been handled by a team in Manila. Four men, working shifts, on a rotation. The best of them is Rogelio, who has a wife and three kids and who writes a good morning message that will take the top of your head off. Rogelio has been me for five years. He knows Dale better than I do. Dale’s four years of intimacy with a woman in Ohio were four years of intimacy with a married man in Quezon City who has never seen snow.

The studio in Riga has replaced Rogelio too. He wrote to tell me in April. He was gracious about it.

So when I say the human touch, understand that I am including him, and that I am not sure the men could tell us apart, and that I am not sure it matters, and that I have thought about this every night since April.

I have opened a second page. It is a model of me, trained on eleven years of my archive, and it runs at a margin I could not touch when I was doing it with my own face and my own hands. It went live in June. It is outperforming me by a factor of six.

I did it for archival purposes.

I am thirty-four. I have no family, no friends from before, no profession my country will name out loud, and an asset class that depreciated to zero in eighteen months. What I have is a body of work, and a theory of what has been lost, and the suspicion, which arrives around four in the morning and does not leave, that I am the only one who noticed, and that even I am not sure I have the standing to say it.

Posted in Pornography | Comments Off on Another National Treasure is Falling to AI (Fiction)

What Percentage of Philip N. Cohen’s Blog Passes the Front Page Test?

Posted on August 16, 2026 by Luke Ford

I estimate that 90% of Family Inequality would survive the front-page test. His dangerous material is concentrated. Most of the blog is demographic analysis, methodological criticism, data visualization, teaching, research commentary, corrections, professional news and open-science advocacy. An accurate quotation from those posts make Cohen look like what he is: an opinionated quantitative sociologist showing his work.

The remainder divides into two different categories. Some posts are merely caustic. The evidence underneath them is substantial. A smaller group has trouble with the front-page test because Cohen moves from demonstrating that a claim is wrong to asserting dishonesty, motives, character defects, or moral depravity on evidence that does not compel that conclusion.

Cohen rejects the front-page test. When Nicholas Wolfinger criticized the caustic language in Enduring Bonds, Cohen responded in “Tone policing: Am I allowed to put Regnerus, Wilcox, and Hitler in the same headline?” that listing his “bad words” proved nothing. The question, he said, was whether the assessments were wrong. He described himself as a “caustic person” and defended his willingness to use harsh language.

Let me refine the front page test. If an accurate quotation appeared before a large audience that had not already accepted Cohen’s politics or accumulated case against the target, would the statement still look like a warranted conclusion from evidence he could readily produce?

By that standard, these are the failures I found.

1. Charles Murray, 2012.

This may be the biggest failure. Cohen wrote about Charles Murray’s Coming Apart before reading it:

“Since I haven’t read the book yet…”

He then said Murray was “not a scholar doing (peer reviewed) research to advance our collective understanding of social life. He is a political propagandist,” and therefore should primarily be judged by the consequences of his work rather than its scientific accuracy.

That is the quotation the front-page test is designed to catch. Cohen may have had an extensive prior basis for judging Murray. But the post makes the sequence look terrible: I haven’t read the book, but I know what sort of person produced it and how it should be judged.

The epistemic Cohen who tells everyone else to examine evidence has yielded to categorical prior judgment.

Charles Murray on his propaganda playing field

2. Brad Wilcox as “evil,” 2014.

The title is extraordinary: “Final proof there is no human tragedy Brad Wilcox will not exploit in order to promote marriage.” Cohen opens:

“I’m not going to dignify this with a thorough debunking, but here’s a quick note to highlight the evil that walks among us in academic robes.”

The underlying dispute concerned a Wilcox and Robin Fretwell Wilson argument about marriage and violence against women. Cohen had substantive methodological objections and later wrote a stronger post explaining how to evaluate the data story. But “evil that walks among us in academic robes” is a moral judgment about a named colleague.

On a front page, accurately quoted, Cohen would either have to say, “Yes, Brad Wilcox is evil and I meant exactly that,” or retreat to something such as “I thought his argument exploited violence against women to advance marriage promotion.”

The smaller proposition is easier to defend.

Final proof there is no human tragedy Brad Wilcox will not exploit

3. David Blankenhorn, 2015.

Cohen begins a long post:

“I don’t know David Blankenhorn, so I can’t really judge whether he’s still a hypocritical opportunist or he’s really transformed into a half-evolved pseudo-moderate.”

Then he spends a substantial part of the essay analyzing Blankenhorn’s funding, salary, organizations and political trajectory. At one point he characterizes conservative foundation activity as billionaires heating “their tax shelters with burning cash” while millionaires exchange “bloated salaries in the service of ideological reproduction.”

The post contains interesting analysis. Some of it is excellent. The problem is the movement from observable incentives to psychological and moral characterization.

“I don’t know him” followed by “hypocritical opportunist” is a front-page-test failure.

The marriage movement has failed, Blankenhorn edition

4. Regnerus’s inner motives, 2015.

This one is more complicated because Cohen accumulated documentary evidence about the Regnerus project. He was entitled to argue that Regnerus’s public accounts were contradicted by emails, funding records and subsequent political behavior.

But in “Regnerus responds” Cohen goes beyond the evidence:

“Only God can truly see into the unlit depths of Regnerus’s heart – but the rest of us can be pretty sure he’s lying based on his actions.”

He also calls Regnerus’s account “self-serving” and says accepting it would make one a fool.

“His account is inconsistent with the documentary record” is demonstrable.

“We can be pretty sure he’s lying” requires evidence not merely of falsity but knowledge and intention. Cohen may believe the surrounding evidence establishes that. But this is where an open-science critic should be careful. “Lying” is a finding about the mind of the speaker.

I would call this a borderline failure because Cohen provides extensive circumstantial evidence for his position.

Regnerus responds

5. James Wright and Brad Wilcox “lied,” 2014.

The headline is careful: “James Wright’s recounting of the Regnerus review process wasn’t true.” The parenthetical first sentence is not:

“(And Brad Wilcox lied continuously, too.)”

Cohen then says Wright “apparently lied” in his published description of the review process.

The documentary contradictions Cohen reports are serious. But there are three propositions:

Wright’s account was false.

Wright should have known it was false.

Wright knowingly stated something false to deceive.

Cohen moves from the first to the third. A front-page quotation of “Brad Wilcox lied continuously” would force him to defend intention claim by claim.

James Wright’s recounting wasn’t true

6. “Brad Wilcox lies a lot,” 2025.

By 2025 the qualification is gone. The title is “Brad Wilcox lies a lot, part whatever.” Cohen writes that he has covered Wilcox’s “dishonesty” many times and says:

“People who lie without remorse or reparation are not trustworthy.”

Wilcox tells an interviewer that his pronatalist policy approach includes all families, then emphasizes that the overwhelming majority of births occur in heterosexual relationships. Cohen interprets Wilcox’s statement “right now, marriage is for everyone” to mean, in effect, “I wish it weren’t.” He then calls the answer a lie.

This is one of the stronger front-page failures because Cohen is interpreting subtext and then treating his interpretation as evidence of deceptive intent.

He may have excellent historical reasons for believing Wilcox opposes same-sex marriage. That supports “Wilcox’s professed inclusiveness is difficult to reconcile with his record.” It does not establish what Wilcox was privately thinking while answering that question.

Brad Wilcox lies a lot, part whatever

7. “Complete bologna and/or just dishonest,” 2016.

Cohen originally titled a post “The new Wilcox thing is complete bologna and/or just dishonest.” After Wilcox supplied additional data, Cohen updated it to say:

“I now conclude the report is just dishonest, not complete bologna.”

That is funny. Cohen had identified a serious analytical problem in a report about Arizona families and high-school graduation. But “dishonest” again asserts intentional deception rather than poor analysis, selective presentation, motivated reasoning or incompetence.

The front-page-safe version writes itself: “After Wilcox supplied the missing data, I no longer think the analysis is simply incompetent. The pattern of selective presentation raises the question of whether readers were deliberately misled.”

Cohen chooses the more prosecutorial formulation.

The new Wilcox thing is complete bologna and/or just dishonest

8. Christian Smith, 2014.

Cohen knew he had a conflict. He explains that he self-published his review of Christian Smith’s The Sacred Project of American Sociology rather than seeking an outside venue because he had previously used profanity in an email to Smith and did not want to conceal “documented personal animosity.”

Then the review compares Smith to a vaccine denier and says Smith comes “uneasily close to the line where common arrogance tips over into a lack of grip on reality.” Later Cohen says that regardless of Brad Wilcox’s publication record he might vote to deny Wilcox tenure on grounds of “dishonesty and incompetence.”

Smith responded on Cohen’s own blog, calling the review lopsided and noxious. Cohen graciously published the response, but his reply included the question of what kind of “intellectual coward” would privately agree with Smith while being afraid to say so publicly.

This is a front-page problem not because Cohen combines an acknowledged personal feud with judgments about sanity, cowardice, character and competence. That is when a scholar should raise his evidentiary threshold.

It’s modernity, stupid

Christian Smith responds to Cohen

9. Trump voters and racism, 2016.

There are two cases here.

“Looks like racist Southern Whites like Trump” is provocative but largely passes. Cohen says racism is difficult to define and measure, calls his result tentative, and presents an empirical analysis relating racial composition and Trump voting. The headline is much stronger than the design, but the post contains caveats.

“Not-too-racist Whites of America: Do you want to be that person?” is different. Cohen addresses prospective Trump voters and tells them that voting for Trump means taking a stand “against virtually every Black person who you will know or meet.” He says they may feel better if they do not do it.

That is political persuasion, not analysis, and it makes an extraordinary jump from overwhelming Black opposition to Trump to the claim that casting a Trump vote constitutes taking a stand against Black people.

I think that one fails the front-page test.

Not-too-racist Whites of America

10. “Vance hates children,” 2024.

The title is “Between the lines, Trump is racist/sexist and Vance hates children.” But the body does not demonstrate that J.D. Vance hates children. Cohen discusses a Vance statement attacking Randi Weingarten, interprets it as implying indifference to child abuse, and concludes that Vance “leans in that direction,” calling it another red flag.

The body is more cautious than the headline. That makes the headline itself the front-page failure. “Vance hates children” is an attribution of an absurdly broad inner disposition that the evidence in the post does not establish.

Between the lines, Trump is racist/sexist and Vance hates children

11. The 2026 pronatalism post.

“Pronatalism is not sending their best” begins with “Yesterday’s Oval Office evil-bananas snoozefest.” Cohen says the Trump movement is not serious about domestic policy except cutting taxes and regulations, describes much of its domestic activity as “grifting,” “political hate-mongering and White-maxxing,” and calls pronatalist policy “another superficial sop to poorly-informed followers and those who exploit their animosities and anxieties.”

This is a useful example because some of the underlying political argument may be perfectly defensible while the prose stops distinguishing things Cohen ordinarily insists be distinguished.

“Trump’s administration uses pronatalist rhetoric but has adopted few policies likely to raise fertility appreciably” is an empirical proposition.

“They aren’t serious” is an inference about purpose.

“Grifting” implies corruption or bad-faith self-enrichment.

“White-maxxing” implies a racial objective.

“Poorly-informed followers” makes a claim about the people attracted to the movement.

Four distinct propositions get compressed into one burst of contempt.

That would be one of my strongest candidates for failing his own standards of analytical disaggregation.

Pronatalism is not sending their best

There are also posts that look dangerous but I think pass.

“Blame the poor, ‘We tried generosity and it just doesn’t work’ edition” calls a welfare argument “stupid and evil.” That is harsh. But Cohen then separately explains why he considers it stupid and why he considers it evil. “Stupid” refers to the calculation. “Evil” refers to the normative implication he attributes to the argument. I would not write it that way, but it is considerably better defended than “Vance hates children.”

Likewise the Regnerus material mostly holds up better than a collection of Cohen’s adjectives makes it appear. Cohen accumulated timelines, public-record emails, funding information, peer-review information and subsequent political activity. He also corrected himself when he had facts wrong. In the Paul Amato episode, for example, Cohen explicitly regretted not contacting Amato first and published Amato’s statement in full.

And this is why I would put the overall pass rate so high. Cohen’s characteristic habit is favorable to the front-page test: he links sources, shows calculations, releases code, updates posts and leaves corrections visible. His 2017 attack on sociology’s “culture of trust, don’t verify” is severe, but it states a general principle that he subsequently tried to institutionalize through SocArXiv and replication requirements.

The failure pattern is narrow. Cohen is safest when he writes about propositions. He becomes more vulnerable when he writes about persons. “That regression doesn’t establish causation” almost always passes. “That statistic is false” usually passes. “The article omits crucial information” generally passes. “The author is a propagandist” is harder. “The author is dishonest” is harder still. “The author lies a lot” demands evidence of repeated intentional deception. “The author is evil” or “hates children” is no longer operating in the evidentiary language that Cohen demands of empirical social science.

Cohen’s method is strongest where it is impersonal. His standards become less disciplined when a recurring opponent has accumulated into a character in his mental world. Wilcox eventually stops being a scholar who repeatedly makes claims Cohen considers bad and becomes “Brad Wilcox,” an explanatory category. New conduct is then interpreted through an established character judgment.

The irony is that Cohen spotted the corresponding danger in Hanna Rosin. He complained in “Correct that error, Hanna Rosin edition” that she seemed to decide on the story first and then find facts that confirmed it. The front-page failures in Family Inequality tend to occur when Cohen starts doing something structurally similar with people.

Cohen’s numbers generally survive hostile quotation better than his adjectives.

There is a trajectory toward self-narrowing. The striking change is that the unit of moral judgment changes. Early Family Inequality mostly attacks propositions. By the middle period it attacks named people’s competence and integrity. Eventually Cohen makes causticity part of his scholarly persona. In the most recent political writing, contempt sometimes precedes analysis.

The trend looks like this:

1. 2009-2011: wit and irritation, but mostly proposition-centered.

Early Cohen can be nasty. His 2011 retrospective describes people spreading a misleading food-stamp graphic as “gullible, mean-spirited members of the blogosphere’s conservative echo chambers.” But the characteristic early move is still: here is the claim, here are the data, here is why the claim does not work. His famous feminist-statistic debunking is particularly important because he attacks a falsehood useful to his own political side rather than turning its promoters into villains.

Family Inequality’s 2011 top 10

There is a pleasure in correction, but no enemies list.

2. 2012 is the hinge: Regnerus converts error into misconduct.

The Regnerus affair appears to change something. Cohen thinks he has encountered bad research joined to a political operation. Already in June 2012 he calls the paper bad-quality research and says Regnerus “cynically manipulated” its promotion. Another post begins with the proposition that combining bad science with ideological motives produced a “disingenuous attempt” to stigmatize gay families.

Family Inequality, June 2012

Cohen had evidence that made the moralization understandable. He subsequently reconstructed funding, timing, peer review, conservative institutional involvement and the paper’s use in litigation.

Regnerus Affair timeline

Something important happens here. Methodological criticism acquires an ethical dimension. A bad comparison is no longer necessarily just a bad comparison. It may be evidence of political purpose. A misleading presentation may be evidence of dishonesty. A compromised review process may reveal character. Once that door opens, the blog changes.

3. 2013-2016: named adversaries become characters.

This is the escalation. By 2013 Cohen can say that publishers of Brad Wilcox’s work have been “duly notified of his dishonesty, data manipulation, and incompetence.”

Fatherhood’s transformations

In 2014 Wilcox becomes “the evil that walks among us in academic robes.”

Final proof there is no human tragedy Brad Wilcox will not exploit in order to promote marriage

A month later Cohen writes that James Wright “apparently lied” and adds parenthetically that “Brad Wilcox lied continuously, too.”

James Wright’s recounting of the Regnerus review process wasn’t true

In the Christian Smith review from the same summer, Cohen says he is self-publishing it because a previous profane exchange with Smith gives him “documented personal animosity.” He says that he might vote against Wilcox’s tenure because of “dishonesty and incompetence.”

It’s modernity, stupid

That is qualitatively different from early Family Inequality. The recurring adversary has become a character with established properties. Once Wilcox is “dishonest,” a new Wilcox claim enters the blog with a prior attached to the man himself. By 2016 Cohen can update an article to say that after receiving more information, he has determined that a Wilcox report is “just dishonest, not complete bologna.”

The new Wilcox thing is complete bologna and/or just dishonest

This creates an epistemic danger that Cohen would recognize if it appeared in someone else’s work. Evidence begins to update a judgment about the person, and then the accumulated judgment about the person begins to influence the interpretation of new evidence.

4. 2016-2019: the prosecutorial style becomes an identity.

Trump broadens the battlefield. Cohen is no longer mostly fighting within family sociology. He writes about racist Trump supporters, political persuasion, journalism and public truth. Some of this work is empirical. His 2016 analysis of Southern White Trump voting, for example, discusses the difficulty of defining and measuring racism.

Looks like racist Southern Whites like Trump

The crucial event for the trajectory comes in 2019 because somebody from inside sociology tells him what the style looks like from outside Cohen’s circle.

Nicholas Wolfinger reviewed Cohen’s Enduring Bonds in Social Forces. He praised Cohen’s quantitative analyses but argued that the valuable material was sometimes overwhelmed by a “torrent of ad hominem asperity.” The review catalogued Cohen’s descriptions of Regnerus, Wilcox, Ron Haskins, David Blankenhorn and Christina Hoff Sommers.

Nicholas Wolfinger’s review of Enduring Bonds

Cohen’s response is fascinating. In “Tone policing: Am I allowed to put Regnerus, Wilcox, and Hitler in the same headline?”, he calls himself a “caustic person,” says this is his “career choice,” and argues that listing his harsh words proves nothing unless the critic shows that the judgments themselves were unwarranted.

That is the turning point. By 2019 the nastiness is no longer leakage. It has been challenged, considered and retained. Cohen has incorporated it into his understanding of what he does.

5. The recent phase is not simply “more nasty.” It is less inhibited.

There is an interruption. Much of Cohen’s 2020-2022 work is technical, methodological and open-science oriented. He can be exceptionally cautious. During COVID he writes explicitly about uncertainty and the limits of prediction.

What happens next?

But when adversarial political writing returns, especially from 2023 onward, the old prosecutorial voice comes back with fewer restraints. There are hints even with Melissa Kearney. Cohen’s substantive criticism is often excellent, but personal irritation starts leaking through. In 2023 he complains that she calls him “Phil” without knowing him and dismisses the notion that he should personally bring his criticism to her.

Does family science have an inclusivity problem?

Then you get the 2024 headline “Trump is racist/sexist and Vance hates children.” The body is considerably more qualified than “Vance hates children.” That is significant. The headline is no longer a compressed statement of what the evidence demonstrates. It is provocation. By 2026, “Pronatalism is not sending their best” opens with “evil-bananas snoozefest” and moves through “grifting,” “political hate-mongering,” “White-maxxing,” “poorly-informed followers” and people exploiting their anxieties.

That prose is different from the Cohen who once delighted in patiently taking apart a statistical claim. The careful demographer and the political polemicist now occupy the same post.

That is where I see the self-destruction.

As of 2026 Cohen remains a professor of sociology at the University of Maryland and director of SocArXiv. His university describes his scholarship and open-science work. His February 2026 CV still shows a substantial academic career.

Philip N. Cohen at the University of Maryland

Columbia University Press published Citizen Scholar in 2025. The press presents his public engagement as a qualification rather than an embarrassment. Citizen Scholar, Columbia University Press

So “Cohen blogged himself out of respectable academia” is false. But there are three other kinds of self-damage.

The first is loss of persuasive range.

Early Cohen can be handed to someone who disagrees with him politically. “Look at the denominator.” “Look at the selection effect.” “Here is the spreadsheet.” The reader doesn’t need to like Cohen. Once the rhetoric becomes “evil,” “liar,” “dishonest,” “hates children,” “White-maxxing,” the audience sorts before reaching the analysis. Friends hear clarity. Enemies hear confirmation that Cohen is a partisan. The undecided reader has been made harder to reach. That is a cost for somebody whose distinctive asset is the transferability of his quantitative credibility across ideological lines.

The second is contamination of his strongest intellectual identity.

Cohen has spent years building the persona of the man who checks. If Brad Wilcox says X and Cohen says, “No, here are the data,” Cohen possesses an asymmetrical advantage. He is the auditor. But once Cohen repeatedly tells you beforehand that Wilcox is dishonest, incompetent and morally objectionable, the relationship changes. Now two political antagonists are arguing. Cohen may still have the better spreadsheet. But he has surrendered some of the rhetorical advantage that comes from appearing to have arrived reluctantly at the adverse conclusion. That seems to me the biggest self-inflicted loss.

The third is that the rhetoric threatens the philosophy of Citizen Scholar.

The Columbia description emphasizes public engagement, credibility and the development of a public intellectual identity. Citizen Scholar: Public Engagement for Social Scientists Put that beside “evil that walks among us in academic robes.” Put humility beside “Vance hates children.” Put building trust beside “Brad Wilcox lies a lot.” There is a tension.

Has Cohen followed the rules he formulated for everybody else?

I don’t think this is a story of a man becoming angrier with age. The mechanism visible in the archive is professional. Blogging rewarded Cohen for discovering errors. Regnerus taught him that an error could conceal an institutional and political story. Repeated investigation taught him that certain people were repeat offenders. Twitter rewarded compression and conflict. Political polarization raised the moral stakes. Tenure and professional success reduced the career cost of speaking harshly. Eventually public combat became part of his scholarship.

At each stage the next step makes sense. That is how trajectories work. And the dangerous endpoint is that the prosecutor gradually becomes too certain that he already knows who the defendants are. Did a blog originally valuable because Philip Cohen was willing to check claims eventually produce a Philip Cohen who sometimes feels he no longer needs to check his judgments about people?

The archive gives enough evidence to take that question seriously. Let’s investigate. I will identify every post in which Cohen calls a named person dishonest, a liar, unethical, racist, stupid, incompetent, propagandistic, opportunistic, evil, or some close equivalent. Then score whether the post demonstrates (1) error, (2) knowledge of error, and (3) intent to mislead. That might tell us how often Cohen’s strongest character judgments clear the evidentiary standard implied by the words he chooses. The audit applies Cohen’s insistence on operationalization to Cohen.

I did it. The result is sharp. Cohen is good at proving error. He is less consistent at proving knowledge of error, and weaker still at proving intent to deceive. The important exception is the Regnerus affair. There the emails and documentary record sometimes let him cross the line from “wrong” to “knew” to a reasonable inference of “misled deliberately.” Outside that cluster, Cohen often makes the last jump too quickly.

I treated this as a near-census rather than pretending the WordPress search gives us a perfect database export. I searched the Family Inequality archive, the Regnerus and National Marriage Project tag histories, named-target searches, and variants of dishonesty, lie, liar, deceptive, unethical, incompetent, manipulation, propaganda, fraud, plagiarism, untrustworthy, opportunist, racist, bigot, evil and related terms. I excluded cases where Cohen merely called an argument stupid or a report bad. The accusation had to attach a serious epistemic or character defect to an identifiable person.

For the central test I found 19 posts in which Cohen makes a person-level accusation of dishonesty, deception, incompetence or closely related research misconduct that can sensibly be tested against three stages. I scored them this way. E means the underlying error or misleading statement is demonstrated. K means there is evidence the target knew the statement or presentation was wrong or misleading. I means there is evidence that misleading the audience was intentional. A 2 means strong evidence, 1 means circumstantial or inferential evidence, and 0 means the post does not establish it.

The result: 15 of the 19 score 2 on underlying error. Only seven score 2 on knowledge. Only five score 2 on intent. More importantly, those five recycle the same documentary record surrounding Regnerus, Wilcox and the publication of the Regnerus study.

1. “Distorting data on divorce at the National Marriage Project,” 2011. Brad Wilcox. E2, K1, I1.

In “Distorting data on divorce at the National Marriage Project”, Cohen does an excellent job showing weak causal reasoning, tiny subsamples and tendentious presentation. He then attributes deliberate manipulation to Wilcox. The first proposition is demonstrated. The second rests principally on Cohen’s inference from Wilcox’s expertise and choices.

2. “Time travel: Regnerus study timeline suggests superhuman abilities,” 2012. Mark Regnerus. E2, K1, I1.

In “Time travel: Regnerus study timeline suggests superhuman abilities”, the chronology supports his argument. The paper was submitted February 1 while the official data collection continued until February 21 and the final dataset was delivered February 24. Cohen adds that Regnerus “cynically manipulated” the study’s promotion. The timeline establishes the anomaly. It does not yet establish cynical intent. Indeed, Regnerus appears in the comments and supplies an explanation about receiving a nearly complete dataset earlier.

The documentary evidence that later made Cohen’s suspicions much stronger had not yet emerged.

3. “Fatherhood’s transformations: What if someone checked the facts?” 2013. Wilcox. E2, K1, I1.

“Fatherhood’s transformations: What if someone checked the facts?” is a strong debunking. He works claim by claim through Wilcox’s sources and catches assertions that the cited studies do not establish. Yet he introduces the exercise by saying publishers have been notified of Wilcox’s “dishonesty,” “data manipulation” and “incompetence” in pursuit of ideological and financial objectives.

The inaccuracies are demonstrated. The state of mind is imported from Cohen’s developing view of Wilcox.

4. “Marriage makes Wilcox richer, how to lie with asterisks edition,” 2013. Wilcox. E2, K2, I1.

In “Marriage makes Wilcox richer, how to lie with asterisks edition”, Wilcox displays unadjusted figures while a footnote discusses statistically adjusted differences. Cohen reproduces the distinction and explains why it matters. Because Wilcox is presenting the models and footnote, knowledge of what was adjusted and what was not is difficult to deny.

Calling the presentation a “deception” still requires an inference about purpose, but it is a much better inference than in many later cases.

5. “Regnerus delivers deceptive dialogue to delighted devotees,” 2013. Regnerus. E2, K2, I1.

In “Regnerus delivers deceptive dialogue to delighted devotees”, Cohen attends Regnerus’s presentation and catches him first explaining his statistical controls, then presenting simple uncontrolled comparisons to an audience likely to understand them as the results of those controls.

Regnerus knows what his own adjusted analyses are. Cohen therefore has good evidence of knowledge. Intent remains inferred from the staging of the presentation.

6. “Divorce drop and rebound: paper in the news,” 2014. Wilcox. E2, K1, I1.

In “Divorce drop and rebound: paper in the news”, Cohen’s peer-reviewed analysis provides a strong basis for saying Wilcox’s earlier causal interpretation of recession-era divorce trends was unwarranted. Cohen calls the earlier interpretation both incompetent and dishonest, with the word “Thus” in Wilcox’s prose serving as evidence of deliberate causal implication.

Error is clear. Intent is less clear.

7. “James Wright’s recounting of the Regnerus review process wasn’t true,” 2014. James Wright and Wilcox. E2, K2, I2.

This is where Cohen gets inside the mens rea problem.

In “James Wright’s recounting of the Regnerus review process wasn’t true”, Wright’s published account says expedited review was sought to coordinate publication of two related papers. Cohen produces Wright’s actual to Wilcox, which says the reason for speed was getting the article accepted before a planned public report appeared.

Wright knew what he had written. Cohen’s inference that the later account was designed to distance the journal from coordination is grounded.

The Wilcox portion is similar. Wilcox disclosed to Wright that he sat on an advisory board. Cohen’s later documentary evidence showed a deeper role in fundraising, planning, publication strategy and the project. A strategically incomplete disclosure is a plausible finding.

8. “Regnerus responds,” 2015. Regnerus. E2, K2, I2.

This is Cohen’s best case for using the word “lie.”

In “Regnerus responds”, Regnerus maintained that his original interest was squarely scientific and that he had not intended political or legal use of the project. Cohen produces a pre-data-collection email in which Regnerus discusses project boundaries, timing, a coalition meeting and an early report with political actors involved in opposition to same-sex marriage.

Cohen also draws on sociologist Marc Musick’s independent review of the affair.

Cohen’s rhetoric in this post is ferocious, including speculation about Regnerus’s heart and “diabolical political plans.” But underneath the rhetoric is substantial documentary evidence. Of all the character accusations in the archive, this is not one I would be eager to use against Cohen.

9. “Groups of people with more income are richer, Wilcox credulity edition,” 2015. Wilcox. E2, K0, I0 for the new allegation.

In “Groups of people with more income are richer, Wilcox credulity edition”, Cohen constructs a nice demonstration showing why the report’s state-level marriage and income relationship is largely compositional. He then says he distrusts an unreported robustness claim because “we know Wilcox lies a lot.”

This is a crucial transitional post.

Cohen is no longer demonstrating that Wilcox lied in this case. Previous findings about Wilcox have become a prior. Because Wilcox has lied before, unreported results are now presumed unreliable.

That can be rational Bayesian updating. It is also where an auditor can become a prosecutor.

10. “Marriage promotion snake oil tour,” 2016. Wilcox. E2, K1, I1.

In “Marriage promotion snake oil tour”, Cohen shows that Wilcox’s description of the Oklahoma Marriage Initiative exceeds what the evidence demonstrates. Then he writes that Wilcox “knows that it is not true” and knows audiences will misunderstand and misuse it.

But where is the evidence of what Wilcox knows?

Cohen’s answer is that Wilcox is an expert, the contradictory evidence is available, and the exaggeration serves his argument.

That is evidence. It is also an example of Cohen’s recurring leap from “a competent scholar should know this” to “this scholar does know this” to “therefore he is lying.”

11. “The new Wilcox thing is complete bologna and/or just dishonest,” 2016. Wilcox. E2, K1, I1.

“The new Wilcox thing is complete bologna and/or just dishonest” is instructive because Cohen updates it when Wilcox supplies the complete data.

Cohen initially entertains outright error or fabrication because his replication does not work. Once the missing data appear, the regression replicates, and Cohen withdraws that possibility. He concludes that the presentation is dishonest because extremely small differences are promoted as meaningful and because of other choices in framing.

Cohen deserves credit for the update. It is open-science behavior. But the update also illustrates the mens rea (guilty mind) problem. The evidence moves him from “perhaps fabricated” to “dishonest,” when another available conclusion is “technically correct but tendentiously oversold.”

12. “Sociology’s culture of trust, don’t verify,” 2017. Regnerus. E2, K2, I2, but inherited.

In “Sociology’s culture of trust, don’t verify”, Cohen says Regnerus has lied repeatedly about important matters related to his research, then uses that history to ask how qualitative evidence from an untrusted researcher should be treated in a discipline with little replication. This clears the standard because the linked Regnerus record supplies the evidence. It is not a new demonstration of lying. That distinction becomes important when counting. Five posts saying “Regnerus lied” are not five independent instances of proven lying.

13. “Notes for a review of Cheap Sex,” 2018. Regnerus. E2, K2, I2, again inherited.

In “Notes for a review of Cheap Sex“, Cohen says he cares more about attacking Regnerus’s work because of what he sees as its harmful political purpose. He then summarizes the Regnerus affair as involving biased methods, political coordination, manipulation of peer review and repeated dishonesty. Cohen admits selectivity. He tells readers political consequence affects how much investigative effort he expends.

14. “Wilcox and colleagues plagiarized my work in the New York Times,” 2019. Wilcox, Jason Carroll and Laurie DeRose. E1, K0, I0.

In “Wilcox and colleagues plagiarized my work in the New York Times“, Cohen shows that their analysis is strikingly similar to analyses he had previously published in direct response to Wilcox’s work. That warrants suspicion and probably warrants a citation argument.

Then Cohen writes, in effect, that he is certain they knew about his work even though he cannot prove it. That is the experiment we wanted performed for us. The crucial factual element separating independent rediscovery from plagiarism is knowledge of the earlier work. Cohen acknowledges that he cannot establish it and nevertheless puts “plagiarized my work” in the headline. Under Cohen’s own standards, that fails.

15. “Wilcox plagiarism denial and ethics review,” 2019. Wilcox. Mixed.

In “Wilcox plagiarism denial and ethics review”, the plagiarism portion remains E1, K0, I0. The historical Regnerus-affair dishonesty portion is E2, K2, I2.

Cohen reproduces Wilcox’s denial and acknowledges that only Wilcox knows whether the New York Times analysis was inspired by Cohen’s work. He then changes ground and asks whether Wilcox is trustworthy, reviewing the older documentary evidence about Wilcox’s role in the Regnerus project and what he disclosed to the editor. Cohen more or less states the doctrine he has developed. Past malfeasance changes the burden of trust. That is reasonable for deciding whether to trust someone. It is not the same thing as proving the next accusation.

16. “Very important social scientist says, wait, Brad Wilcox did what now?” 2020. Wilcox. E2. K and I deliberately unresolved.

In “Very important social scientist says, wait, Brad Wilcox did what now?”, Cohen reconstructs Wilcox’s graph and finds two remarkable errors. Unmarried people were included in the denominator for “very happy marriage,” and respondents who had not even been asked the happiness question in several survey years were apparently treated as not very happily married.

Here Cohen handles the distinction correctly. He calls it sophistry masquerading as incompetence, or perhaps the reverse, and says it is hard to know. That is the epistemic discipline the harsher posts often lack. He proves what the graph does and leaves the state of mind unresolved.

17. “UT-Austin sociologist Mark Regnerus, still not an expert,” 2023. Regnerus. E1, K0, I0 on the specific lie allegation.

In “UT-Austin sociologist Mark Regnerus, still not an expert”, Cohen has strong external evidence for attacking Regnerus’s claimed expertise. A federal judge rejected his testimony as unsupported expert evidence. Cohen notices that Regnerus told the court his writings included material on “the science of transgender medicine” and says this “seems like a lie.” Cohen searches his CV and Google Scholar and finds nothing fitting that description. “Seems like” is a hedge, which matters. But the evidence shows Cohen cannot find the relevant work. It does not show that Regnerus knew the description to be false or intended deception. Perhaps there is material Cohen missed or Regnerus uses “material on” more broadly.

The attack on Regnerus’s expertise is strong. The word “lie” is not established.

18. “Brad Wilcox lies a lot, part whatever,” 2025. Wilcox. E1, K1, I1.

“Brad Wilcox lies a lot, part whatever” is one of the most interesting recent failures.

Cohen defines lying to include true statements deliberately designed to misdirect. He documents Wilcox’s history of opposing same-sex marriage and then interprets Wilcox’s newer inclusive language about marriage and pronatalism as strategic concealment. There is a plausible hypocrisy argument here. But the crucial inference is psychological. Cohen knows what Wilcox really means, knows what he wishes he could say, and knows why he is not saying it. That may all be true. The post does not demonstrate it with the rigor Cohen applies to a regression coefficient. The evidence supports “Wilcox’s current rhetoric is difficult to reconcile with his earlier advocacy.” “Wilcox is lying now” requires another step.

19. “In which I give Lyman Stone’s complaint the attention it deserves,” 2025. Lyman Stone. E1, K0, I0.

In “In which I give Lyman Stone’s complaint the attention it deserves”, Stone accuses Cohen of lying about whether the Institute for Family Studies’ pronatalism excludes LGBT people. Cohen responds that Stone is “the liar, not me.” When the claims are unpacked, they are talking about different meanings of inclusion. Cohen is describing the rhetoric and substantive family ideal. Stone points to policies that would formally extend benefits regardless of sexuality. Cohen may have the better interpretation of what the project values. That does not establish that Stone knowingly uttered a falsehood. Calling the disagreement a lie adds nothing except heat.

The numerical result

Fifteen of the 19 posts convincingly demonstrate a substantive error, serious misleading presentation or competence problem. Cohen’s reputation as someone who checks things is earned. Eight more or less stop one step short of a complete dishonesty case. Cohen establishes the problem and supplies circumstantial reasons for suspecting knowledge or purpose, but the evidence does not independently establish both. Five posts can reasonably clear all three steps if linked prior evidence is allowed. But nearly all five ultimately depend on one evidentiary complex, the Regnerus affair. Wright’s contradictory account, Regnerus’s emails and public claims, Wilcox’s concealed or minimized role and the documentation of political and publication coordination all interlock.

Cohen’s Regnerus Affair timeline is the best entry point into that documentary record.

Four cases are failures under a strict liar standard: the 2019 plagiarism accusation, the 2023 suggestion that Regnerus lied about transgender-medicine scholarship, the 2025 claim that Wilcox’s current rhetoric is itself lying, and the 2025 “Stone is the liar” response.

In each one, Cohen has grounds for criticism. In none does the post adequately establish the necessary knowledge and deceptive intention. And there is a finding I did not expect. The strictest version of Cohen’s case against dishonesty is small. If you read Family Inequality casually, you encounter “lies,” “dishonesty,” “unethical,” “not trustworthy,” “deception” and similar language over many years, it creates the impression of a long succession of independently demonstrated misconduct. When you collapse repeated posts back to their underlying evidence, most of that record reduces to one scandal.

Regnerus is the hinge.

Cohen discovers that you sometimes can establish what somebody knew. That triumph changed him. Before Regnerus, the characteristic Cohen move was:

“This claim is wrong. Here is why.”

After Regnerus, it increasingly becomes:

“This claim is wrong. This person should know it is wrong. Therefore this person is dishonest.”

Regnerus supplied a rare luxury. Cohen had the emails. Most cases do not. And yet the mental model appears to survive the exceptional case that created it. That is why the 2015 Wilcox post is a turn. Cohen no longer needs to replicate every statement before deciding whether to trust an unreported result because “we know Wilcox lies a lot.”

In 2017 the same principle becomes explicit with Regnerus. If you do not trust the author, Cohen asks, how are you supposed to evaluate unverifiable evidence? In 2018 he writes that reputation matters because sociology operates largely on trust. In the 2019 plagiarism case, faced with the accusation he acknowledges he cannot prove, he says the same thing. The research ecosystem depends on trust, and Wilcox is not trustworthy.

This is defensible up to a point. Reputation is information. If somebody demonstrably deceived you yesterday, believing them less readily tomorrow is rational. But notice what happens to the burden of proof. Originally Cohen had to demonstrate dishonesty. Eventually the prior demonstration of dishonesty becomes part of the evidence for subsequent dishonesty.

There is also a control case. In the 2020 Wilcox graph post, Cohen has an error so extraordinary that it would be easy to write “he lied.” Cohen does not. He reconstructs the error, shows exactly what happened, and says he cannot tell whether he is looking at incompetence or sophistry. That post passes the test. It shows Cohen knows the distinction. Which makes the cases where he ignores it more significant.

The moral labels need a different test.

“Evil,” “racist,” “propagandist,” “opportunist” and “bigot” are not propositions about knowledge of a falsehood, so forcing E/K/I onto them would be fake operationalization. They need their own criteria. Some of Cohen’s worst-looking language fares badly. In his 2012 treatment of Charles Murray, Cohen announces that Murray is a political propagandist while simultaneously saying he has not yet read the book being discussed. His 2014 description of Wilcox as “the evil that walks among us in academic robes” is moral denunciation. His 2015 treatment of David Blankenhorn begins by acknowledging that he does not know Blankenhorn well enough to judge whether he is a “hypocritical opportunist” and then puts the label on the table anyway.

There is a recent counterexample. When Cohen calls Lyman Stone a religious bigot and racist in later writing, he does something closer to proper operationalization. He defines what he means, supplies Stone’s own statements and argues that the statements satisfy the definitions. We can disagree with his definitions or applications, but the evidentiary structure is much better than “Vance hates children” or “evil that walks among us in academic robes.”

Cohen generally earns “wrong.” He often earns “misleading.” He sometimes earns “dishonest.” He earns “liar” rarely. And the best evidence that he sometimes overreaches is supplied by Cohen. In the plagiarism case he writes that he is certain Wilcox and his coauthors knew about his work even though he cannot prove it. That inverts the methodological ethic Cohen has spent the blog teaching everyone else. The audit reveals that once a nearby intellectual rival has accumulated enough adverse findings, Cohen becomes increasingly willing to infer the rival’s internal state from the next error. The audit begins as a method for checking Cohen’s enemies. Eventually the enemy becomes the method.

Philip Cohen and the E-Personality: Reading Family Inequality Through Elias Aboujaoude’s Virtually You

Elias Aboujaoude’s 2011 book Virtually You: The Dangerous Powers of the E-Personality argues that going online can elicit a different version of the self. His e-personality is more disinhibited, grandiose, self-centered, impulsive and potentially vicious. The more important claim is that these qualities need not remain online. Aboujaoude argues that online ways of speaking, thinking and judging can migrate back into ordinary life, so that the virtual personality increasingly resembles, and eventually helps remake, the offline one. Applied to Cohen, the book prompts a few questions: Does Cohen display a systematically different public self online? Do particular digital environments encourage certain parts of his intellectual temperament? Does that online style then feed back into his professional identity? On those questions, I think the answer is yes.

The strongest Aboujaoude-inspired reading of Cohen: “The Internet selected, rewarded and progressively institutionalized a particular Philip Cohen.”

The skeptical quantitative sociologist was there already. So were the politics, the humor, the willingness to argue and the causticity. What the Internet supplied was an environment in which the combination of confidence, correction, speed, wit and aggression became productive. It won attention. It generated interlocutors. It produced material. It generated controversies. It helped make books. It helped create SocArXiv. It led to litigation with a president. Eventually Cohen wrote Citizen Scholar and made this public engagement part of his theory of what scholarship should be.

The first chapter of Virtually You supplies the mechanism. Aboujaoude argues that online interaction removes some of the restraints present in face-to-face life. He describes the online self as more assertive and less restrained, while emphasizing that this can be beneficial. It can give people the courage to express unpopular opinions and organize against injustice. Later he summarizes five forces he thinks emerge against this background of disinhibition: grandiosity, narcissism, darkness, regression and impulsivity.

Cohen’s online disinhibition has been intellectually productive. It allowed a professor of sociology to say things about powerful colleagues, journal editors, publishers, professional associations and eventually the president of the United States that the traditional academic system discourages people from saying. The Regnerus investigation could not have unfolded through journal articles. Blogs allowed Cohen to publish documents immediately, update chronologies, respond to new evidence and invite other people to inspect his reasoning. Twitter made the process faster.

That is Aboujaoude’s liberating e-personality doing something valuable. But disinhibition also helps explain the trajectory visible in Family Inequality. “This inference is unsupported” can become “this person is dishonest.” Eventually Brad Wilcox becomes “the evil that walks among us in academic robes.” Regnerus is lying. David Blankenhorn may be a “hypocritical opportunist.” The rhetorical friction that might ordinarily intervene between thinking such a thing and putting it into the permanent scholarly record is reduced.

In his chapter “Ordinary Everyday Viciousness,” Aboujaoude uses Flickr to argue that identified people can become surprisingly hostile online. Going online weakens the social contract and makes angry voices powerful. That matters for Cohen. He has never been anonymous. His name, university, books and professional reputation are attached to everything. That he nevertheless writes things online that would be startling across a seminar table makes him a better Aboujaoude case. The phenomenon is distance. You do not see Wilcox’s face when you type “evil.” You do not watch Regnerus hear “liar.” There is no awkward silence afterward. Instead there may be likes, retweets, supportive replies and additional evidence arriving from allies. And this is where Twitter becomes important.

Cohen has remarkable numbers. In his 2017 year-end review, he reported only 54 blog posts, compared with an average of about 130 annually from 2010 through 2016. His tweets, meanwhile, received about 25 million impressions. By 2019 he said Twitter had taken over much of what the blog once did. Daily chatter, work-in-progress thoughts, feedback, meeting people and learning things were happening there “more efficiently,” although also more noisily. He again reported about 25 million Twitter impressions. His “The blog’s decade” is therefore not merely a retrospective on blogging. It records a migration in Cohen’s intellectual life from one digital environment to another. That is a change in the reward environment. The 2009 blogger writes a post and waits. The 2017 Twitter user can emit a judgment to thousands of followers and see its reception immediately.

Aboujaoude wrote Virtually You before the mature attention economy existed, but he saw the direction. In his discussion of Twitter, he worried that sophisticated exchanges would be compressed into short messages, while visible follower counts would join popularity with authority.

Some of Cohen’s Twitter threads are excellent little pieces of science. A great example is his thread correcting his own analysis of spouses’ heights. Another researcher produced different results. Cohen rebuilt everything, changed his conclusion, posted the corrected analysis and code, and extracted the uncomfortable lesson that he apparently checks more carefully when findings are surprising. He wrote that this was part of the “human condition” and something to resist.

That is social media improving science. The same medium that can produce impulsivity can also produce rapid error correction. Cohen’s own social-media advice also reveals how he recognizes Aboujaoude’s danger. In his tips for academic social media, Cohen warns that social media can substitute ephemeral activity for substantial work, that a bad tweet can carry a “shockingly high” cost, and that the platforms run partly on polarization and outrage. His recommendations include not overreacting, not holding grudges, erring on the side of kindness, setting limits and:

“Respond tomorrow!”

“Respond tomorrow” is Aboujaoude’s prescription. He devotes part of his impulsivity chapter to Google’s old Mail Goggles, which forced users to solve arithmetic problems before sending late-night emails. His point is that online communication removes the pause during which cognition can overrule impulse. He recommends creating artificial friction between feeling and sending.

The question is how often Cohen follows Cohen’s rule of respond tomorrow. His archive suggests: not always. That gives us the first important contradiction. Cohen has become a sophisticated theorist of the medium-induced dangers that appear in his own writing.

The second Aboujaoude category, grandiosity, requires more care. Aboujaoude uses “grandiosity” for the enhanced sense of efficacy that can arise online. The medium encourages people to feel bold and capable because it removes ordinary constraints. He mentions sending a colleague or superior an email more strongly worded than one would normally use face to face. Cohen’s professional version of this is the self-authorized auditor. A conventional academic receives jurisdiction through institutions. A journal asks you to peer-review a paper. A university puts you on a committee. An association appoints you to a panel. The blog reverses that. Cohen finds a paper objectionable. Cohen reviews it. Cohen finds the review process suspicious. Cohen investigates it. Cohen thinks an article requires correction. Cohen writes the correction. Cohen thinks the journal will not correct it. Cohen publicly audits the journal.

There is something democratizing about this. It is also a tremendous expansion of personal jurisdiction. Cohen turns this behavior into theory. His Citizen Scholar material on “peer review unleashed” argues that an incisive Twitter thread can carry peer review into the public square.

Put Aboujaoude and Cohen beside one another and something appears. Aboujaoude asks why someone should tolerate ordinary limitations when he can be the reigning boss of his own little corner of the blogosphere. Cohen asks: why should peer review belong to editors? Sometimes Cohen has the better argument. But Aboujaoude supplies the neglected question: What psychological consequences follow when you become your own assignment editor, referee, prosecutor, publisher and audience analyst?

That question becomes important after Regnerus, which taught Cohen that the independent online auditor can be right. Cohen’s persistence uncovered important things. Imagine what reinforcement that provides. The danger is empirically reinforced grandiosity. “I investigated one supposedly respectable claim that nobody else was adequately checking and uncovered a remarkable story.” The natural next thought is: “Maybe somebody should be checking these people.” Then: “I am somebody.”

Aboujaoude emphasizes the self-presentational nature of social media. Profiles become advertisements. Followers become visible status measures. People select favorable information and construct public selves. Cohen constructs a public scholarly identity. He even advises academics to “Be shameless” about promoting their work.

Cohen’s professional identity has become a brand and the brand has properties: Philip Cohen checks. Philip Cohen has the data. Philip Cohen calls bullshit. Philip Cohen publishes openly. Philip Cohen will say what institutional sociology will not. Philip Cohen is willing to fight.

That identity has value. It sells books, brings journalists, creates followers and gives his interventions presumptive interest. The danger is identity maintenance. Once “the man who checks” becomes your public role, you need things to check. Once “the man willing to call dishonesty” becomes part of the brand, merely saying “this regression is ambiguous” may deliver less psychic and social reward than discovering another liar.

This connects to Aboujaoude’s “illusion of knowledge” chapter. Aboujaoude worries that online information creates breadth without depth, leaving people with exaggerated confidence in what they know. He contrasts vertical, patient understanding with horizontal browsing and notes that even scholars power-browse research online. Cohen’s project is an antidote to this. He clicks the citation. Downloads the dataset. Checks the table. Reconstructs the model. Reads the email.

But the audit of Cohen’s accusations of dishonesty exposes the place where the Aboujaoude problem returns. Cohen is exceptionally rigorous when answering:

“Is this proposition correct?”

He becomes less rigorous when answering:

“What did this person know?” “What did this person intend?” “Why did he say this?”

The statistical confidence becomes portable. Cohen knows about causal inference, selection and measurement. That epistemic authority can subtly travel into domains where the evidence is much weaker.

He usually earns “wrong.”

Often “misleading.”

Sometimes “dishonest.”

Much less reliably “liar.”

That may be Cohen’s distinctive version of Aboujaoude’s illusion of knowledge. It is the illusion that rigor in one inference automatically licenses confidence in the adjacent inference. The Internet makes the transition nearly invisible because everything appears in the same post. Regression table. Citation. Screenshot. Email. Then: “He lied.” Typography makes them epistemically continuous. They are not.

Aboujaoude argues that online communication relaxes adult norms of restraint and permits forms of petulance and immaturity ordinarily inhibited offline. Cohen’s recent Internet vernacular has more of this than his early blogging.

“Evil-bananas snoozefest.”

“Brad Wilcox lies a lot, part whatever.”

“Chef’s kiss.”

“Vance hates children.”

Imagine the same University of Maryland professor delivering these formulations at a demography conference podium: “The extraordinary despicability plus patheticness here is chef’s kiss.”

It sounds imported from somewhere else. That is the e-personality. Aboujaoude’s deepest claim is that the border is porous. Online impatience, immoderation and impoliteness bleeds into ordinary life. With Cohen we can observe the online life feed the career.

His blog contributed material to his textbook and Enduring Bonds. SocArXiv grew out of his open-science concerns. His Twitter behavior helped put him into Knight Institute v. Trump. His experience of Twitter, blogging, public peer review and political activism became evidence for Citizen Scholar. Cohen explicitly says the book draws on his experience as a sociologist in public, including the Trump lawsuit, challenging “scoundrels (and nonscoundrels)” in academia, roughly a thousand blog posts and years of open-science work.

Aboujaoude’s model:

offline person → online personality → leakage back into offline person

Cohen’s case looks more like:

scholar → blogger → social-media scholar → public auditor → institution builder → theory of citizen scholarship → revised conception of scholar

The feedback becomes institutional. The e-personality creates offices, books, lawsuits, professional networks and organizations. And the cross-platform comparison makes the medium effect more plausible. Look at Cohen’s YouTube channel. It is overwhelmingly lectures, demographic explainers, family-sociology teaching, media appearances, SocArXiv instruction and academic talks.

The YouTube Cohen is less prosecutorial. Why? It could simply be content selection. But the medium also restores things Twitter strips away. A face. A voice. Continuous time. An expectation of explanation. The awareness that students may be watching. More effort before publication. Far less rapid conversational reward. The professor returns.

The blog sits in the middle. It permits disinhibition but also allows space for evidence. Twitter and Bluesky push toward the rapid end. YouTube pulls toward the classroom end.

Same scholar. Different affordances. Different Cohen.

Cohen has arrived at a more sociological version of Aboujaoude’s problem. His social-media chapter for Citizen Scholar discusses “collapsed context,” where academics on social media are simultaneously addressing overlapping audiences. He says scholars should use social media to listen and learn.

His earlier advice warns about outrage, ephemeral work, attention, bad tweets and the need to step away. So he knows the medium is dangerous. Where Cohen departs from Aboujaoude is in the remedy. Aboujaoude’s instinct is friction. Slow down. Restore inhibition. Protect the offline person from the online one. Cohen’s instinct is openness. Engage more intelligently. Expose claims. Talk across audiences. Make the public square itself into peer review.

Aboujaoude worries that the Internet changes the scholar. Cohen thinks the Internet can change scholarship. Both are right.

Through the Aboujaoude lens, Cohen’s e-personality has been professionally productive. It made him more influential. The risk is audience capture and path dependence. The public Philip Cohen has been rewarded for being fast, funny, empirical, fearless and combative. Those characteristics reinforce one another. Eventually the persona itself becomes valuable professional capital. At that point restraint has a cost. “I don’t know” does less identity work than “I checked.” “This is ambiguous” is less memorable than “this is dishonest.”

Yet some of Cohen’s finest online work consists of saying he was wrong. That is the unresolved contest inside the archive. One Cohen says: here is my code, check me. Another says: we already know this guy lies. One says: respond tomorrow. Another posts “evil.” One says that scholarship requires openness, honesty and humility. Another infers mens rea from a bad graph.

Aboujaoude warned in 2011 that the traits cultivated online could be imported into ordinary identity. Cohen began Family Inequality at that time. What his archive gives us is something Aboujaoude could barely have observed when writing the book: a longitudinal record of an intellectual living publicly online long enough for the e-personality to become part of the occupation.

That may be the essay. How did the Internet make Philip Cohen into the citizen scholar?

And then the second question: What did the citizen scholar have to become in order to survive and succeed in the Internet’s attention economy?

Same Scholar, Different Medium: Testing Philip Cohen Across Blog, Twitter, Bluesky and Video

The test is to compare Cohen against himself across platforms and time: Family Inequality versus Twitter/X versus Bluesky versus YouTube and formal presentations, controlling where possible for subject. If the same Cohen discussing the same kind of topic becomes more person-focused, categorical and contemptuous as the medium gets faster and more interactive, while long-form and spoken versions remain more qualified, that would be good evidence for Aboujaoude’s mechanism. I ran a matched-case audit.

The simple claim, “the faster the platform, the nastier Cohen becomes,” is only partly true. Twitter and Bluesky do encourage compression, categorical language and occasional contempt. But Cohen can also be exceptionally careful and self-correcting on those platforms. The stronger finding is this: Platform speed interacts with adversarial subject matter.

Cohen is restrained on fast media when he is doing data analysis. He becomes much more categorical when he is fighting a person, organization or political movement. Over time, that adversarial social-media register appears to migrate back into the blog. Formal spoken settings remain his strongest restraint.

This is a matched-case study rather than a complete machine-readable census of everything Cohen has ever posted. Historical Twitter is particularly incomplete. After his move to Bluesky, X cannot even be treated as an independent platform because Cohen says he no longer posts there except for automatic copies of his Bluesky posts. I selected subjects for which Cohen discussed substantially the same thing in at least two environments. The strongest clusters are generations, Jonah Goldberg, COVID, couple height, pronatalism, and AI/open science.

For each artifact I looked for five things:

Person-focus. Does he criticize a proposition or characterize the person making it?

Categorical certainty. Does he preserve uncertainty and alternatives or announce a verdict?

Contempt. Does the language ridicule, demean or morally classify the target?

Evidence density. Does he provide data, documents, calculations, quotations or code?

Qualification. Does he acknowledge uncertainty, counterevidence, alternative explanations or his own possible error?

The result is the pattern across matched examples.

Aboujaoude gives us two predictions. His immediate e-personality prediction is disinhibition. Online people become more assertive and less constrained by the social cues that operate face to face. His deeper prediction is leakage. Habits cultivated online can cross the boundary and become features of the ordinary personality.

Cohen fits the second prediction.

In 2019 Cohen gave a formal presentation to a National Academies committee attacking named generational categories. He certainly was not timid. He said labels such as Millennials encourage stereotyping, clickbait and baseless character judgments. But listen to how he ends.
“Cohort analysis is really important.”

cohen wants to improve cohort thinking. That is Cohen in a room, giving a presentation, answering questions, aware that other experts are physically or institutionally present.

Two years later the issue moves through Twitter into a campaign against Pew. The evidence is Cohen’s account of what happens. He sees an NBC segment, posts on Twitter asking whether social scientists would sign “a very collegial, lightly footnoted appeal” to Pew, receives a positive response and writes a draft. Then he edits himself. The draft, he decides, is too long. More importantly, he decides that the effective version should be “more concise and less ranty.” That is almost experimental evidence of register switching. Cohen knows what the Twitter-energized version of himself sounds like. He deliberately filters that person before approaching an institution he hopes to persuade.

Then compare both with Bluesky in March 2025. His formulation becomes: “How ‘generations’ rots your brain.”

He asks whether the pandemic bifurcated Gen Z and responds that Gen Z does not exist and that people should look at reality rather than marketing clickbait terms. The underlying position has barely changed.

The National Academies Cohen says: cohort analysis is important. The institutional-letter Cohen says: let’s be collegial. The Bluesky Cohen says: this rots your brain.

That is the compression Aboujaoude worried about when he discussed Twitter. Microblogging squeezes complex propositions into short forms where visible followers and immediate reaction partly turn argument into performance.

There is a large time confound here. We cannot know how Cohen would have Bluesky-posted the argument in 2019. But the 2021 admission about making the formal document “less ranty” supports the existence of different registers.

In the Jonah Goldberg case, we can hold both time and target almost constant. In July 2018 Cohen wrote a long review of Jonah Goldberg’s Suicide of the West. The language is not gentle. Cohen calls the book an “extended screed against leftism” and Goldberg’s position on Trump “weak-kneed.” But the unit of criticism remains primarily the argument. Cohen explains why he read the book, identifies the political question he was investigating and puts graphs, tables and references into a longer paper. Then Cohen turns the same investigation into an 18-tweet Twitter thread. The first fifteen entries are superb examples of compressed factual criticism. The United States does not have more forest area than the entire world. Global life expectancy was not almost zero in 1960. Goldberg’s marriage-wage claim is not supported. Cohen supplies page numbers.

Then something happens. Having finished demonstrating the errors, Cohen imagines Goldberg conceding them: “look I think you’re an obnoxious prick, but you’re right I made some mistakes…”

Cohen then jokes that if he ever made mistakes he would admit them and asks readers which kind of public exposure would be most painful.

The long review wants to explain Goldberg. The Twitter thread eventually wants to humiliate him. The factual foundation is substantially identical. The change occurs after the intellectual work is finished.

That is good evidence for a medium effect because it is the same Cohen, same target, same book, same month and substantially the same evidence. It is what Aboujaoude means by disinhibition. Online distance can embolden people to send communications more strongly worded than their face-to-face equivalents because the recipient’s immediate reaction has disappeared.

But then come two important negative controls.

COVID should have made Twitter terrible for careful thinking. It was fast moving, politically polarized, frightening and saturated with engagement. Yet in March 2020 Cohen tells his students that he is answering coronavirus questions as a scientifically literate person who reads the news, not as a pandemic expert. He adds that if he is wrong, people should tell him and he will tell his students. A month later he records a 28-minute COVID lecture. When he discovers one verbal mistake and receives feedback pointing to an oversimplification, he appends both corrections publicly. No enemy. No contempt.

The same thing happens with the couple-height episode. Cohen posts an analysis on Twitter. Jeff Spies gets different results. Cohen rebuilds and rechecks everything. Then he publicly thanks Spies and writes: “Lesson: I seem to check things more carefully when the results are unexpected.”

He calls this a feature of his human condition that should be resisted.

Then: “Other lesson: It’s OK to be wrong! And sharing helps others help u”

And he posts the data and code.

Cohen’s couple-height correction thread

This is nearly the ideal citizen scholar, and it is happening on Twitter. That result matters. When the target is his own analysis, Twitter can make him transparent. When the target is Jonah Goldberg, the same affordances help convert fact checking into public combat. The interaction is between medium and target.

On Bluesky in late 2025 Cohen starts a thread with a joke about who wants to discuss the economics of pronatalism. Then he makes an empirical argument. More births can reduce women’s current labor-force participation, so pronatalists arguing that births solve worker shortages are partly trading workers today for workers tomorrow. He gives an estimated range and says it comes from Current Population Survey data. That is compact public demography.

Now compare Cohen speaking at UCLA in January 2026 about Michael Geruso and Dean Spears’s After the Spike. This is a disagreement over one of Cohen’s hottest contemporary subjects. The spoken Cohen begins by saying what he appreciates. He praises the authors for avoiding racism and sexism. He says they genuinely do not appear to be hiding a patriarchal or white-supremacist agenda. He appreciates their humanism. He says: “I respect the challenge of thinking about it.”

He distinguishes agreement from disagreement. He limits his empirical claims. Long-run projections are uncertain. The baby boom was not predicted, but that does not mean another one will occur. He agrees that there is no reason to assume an automatic fertility rebound. Cohen eventually gets very strong. He warns that failed liberal pronatalism could open the way to coercion, invokes the Nazi pronatalist precedent and says contemporary population panic is occurring alongside a rise in fascism. But he carefully says this is not Geruso’s fault and that discussing the issue did not create the political phenomenon. That is combative scholarship with brakes.

Four months later, look at Family Inequality.

“Pronatalism is not sending their best” begins with an “evil-bananas snoozefest.”

The MAGA domestic program is “grifting,” “political hate-mongering and White-maxxing.”

Its audience consists partly of “poorly-informed followers.”

Participants in the Oval Office event offer Trump “idiotic praise and lies.”

Later Cohen leaves policy analysis and starts announcing inner motives.

Trump “wants to appear rich and sexually virile.”

Robert F. Kennedy Jr. “wants to grift over environmental toxins.”

The conservative project wants sex to be women’s entry into “a lifetime of subordination.”

This is long-form blogging. It is nastier and more psychologically certain than the Bluesky economics thread. It is much less restrained than the UCLA remarks. Something has happened to the blog.

Aboujaoude’s argument is that online traits do not remain contained within their original medium. Online impatience, daring and impoliteness can be incorporated into behavior elsewhere. He describes the border as porous, with patterns cultivated in virtual life bleeding into ordinary life.

Cohen’s later blog looks like that leakage. The early blog was a slower medium. It sat somewhere between a journal article and Twitter. By 2019 Cohen says Twitter had taken over much of the blog’s original role as an immediate outlet for chatter, work in progress, feedback and interaction. He wrote only 39 blog posts that year while receiving about 25 million Twitter impressions, and described Twitter as doing those functions “more efficiently (if more noisily).”

The blog’s decade

This shift had already become enormous by 2017. Cohen wrote 54 blog posts that year, down from a 2010-2016 average of about 130, while his tweets received more than 25 million impressions.

Family Inequality year-end review, 2017

Cohen noticed a cognitive consequence. While explaining why he spent 2018 reading books, he said one goal was to repair his “Twitter-degraded attention span.” He interrupted the sentence to joke that he had just spent five minutes randomly flitting around online.

That is valuable evidence because Aboujaoude’s “Illusion of Knowledge” chapter argues that online reading becomes horizontal, jumping and checking rather than sustained, vertical attention. Cohen uses almost the same diagnosis on himself.

The 2026 AI material provides another miniature experiment. In February Cohen confronts a difficult question at SocArXiv. A paper relies extensively on AI but may technically meet the repository’s old standards. Should SocArXiv accept it? The institutional Cohen is remarkably open.

“Where should SocArXiv draw the AI line?” begins from uncertainty. Cohen says he has his own biases. He considers competing arguments. He says he is not persuaded by one ethical argument but does not rule out changing his mind. He acknowledges that an anti-AI policy could create an unwinnable arms race. He ends by asking for criticism from good people with different perspectives and experience.

Two weeks later the target is no longer Cohen’s own institution. It is the American Sociological Association. On social media: “The implications of this are catastrophic. The American Sociological Association should do something about this but it doesn’t care to.”

On the blog the same day, the headline becomes: “There is a tsunami of fake social science research coming for us, and the American Sociological Association doesn’t care.”

Cohen supplies a model, code and data, so the evidentiary density rises enormously, but the institutional verdict remains categorical. ASA has “completely abdicated.”

This tells us something different. By 2026 there is no longer a clean Bluesky Philip, LinkedIn Philip and blog Philip. There is a public Philip Cohen who syndicates a rhetorical position across platforms. The medium still determines length. It no longer determines persona. That looks like leakage.

There is another possibility. Perhaps this has little to do with psychology and everything to do with strategic communication. Cohen knows that different settings demand different registers.

His own social-media advice is relevant. He says social media offers instant, free, multidirectional communication and lets scholars shape their image. He also says these platforms run on polarization and outrage, warns that a bad tweet can carry a shockingly high cost, advises people not to overreact or hold grudges, says not to punch down and tells users to err on the side of kindness.

Fast media do not automatically produce contempt. The COVID and couple-height Twitter material are counterexamples strong enough to kill that claim. The 2025 pronatalism Bluesky material is another. Cohen can do concise, careful, transparent social science perfectly well on a microblogging platform.

The second hypothesis does much better. Fast interactive media amplify adversariality. The Goldberg comparison is particularly strong because time, target and evidence are held nearly constant. The long review contains tough judgments. The Twitter version ends in imagined insult and public humiliation. The generations material points the same way, although with a time confound. Formal Cohen qualifies. Microblogging Cohen compresses.

The third hypothesis is stronger still. Embodied or formal spoken settings appear to restore inhibition. The National Academies presentation contains distinctions and a lengthy clarification about what Cohen is not arguing. The UCLA remarks repeatedly acknowledge the other side’s good faith, strengths and legitimate points even while Cohen is warning about fascism. The COVID lectures contain corrections and explicit uncertainty. Cohen uses YouTube to distribute lectures and presentations. The causal variable could be genre, audience, preparation, physical co-presence or awareness that the named person may be sitting five feet away. “Video” bundles all of those together.

The fourth hypothesis is the one I find most convincing. Adversariality has become portable. The later Family Inequality blog no longer reliably provides the inhibition that long form once provided. Some recent posts contain all the evidence of old Cohen while importing the rhetorical aggression of social media.

The May 2026 pronatalism post is the example. That is closer to Aboujaoude’s theory. His point was that the online self becomes habitual and begins to reorganize the offline or ordinary self.

In Cohen’s case I would say: There is evidence that digital public scholarship first gave Cohen a special adversarial register, then rewarded that register heavily, and eventually helped integrate it into his general public-scholarly persona.

We know that the rewards were enormous. By 2017 and 2019 he was receiving about 25 million Twitter impressions per year while the blog had become less active. And Citizen Scholar transformed that online career into a theory of professional practice. Cohen now describes social media as a place where scholars disseminate, listen, learn and negotiate overlapping audiences.

Social media created a productive environment for Cohen’s existing combination of skepticism, confidence, humor and combativeness. Those environments rewarded public correction and adversarial performance. Cohen learned to use them brilliantly. The resulting public persona then became less platform-specific and more integrated into his conception of scholarship.

And there is one more finding. The best predictor of Cohen is an enemy. When Cohen is analyzing his own height result, Twitter produces humility. When Cohen is explaining COVID to students, Twitter produces caution. When Cohen is doing CPS calculations on Bluesky, Bluesky produces compact demography. When Cohen encounters Goldberg, Regnerus, Wilcox, Stone, ASA, Trump or MAGA, the probability of person-level judgment rises.

The Internet does not supply the hostility. It supplies a low-friction, high-reward environment in which hostility toward an already meaningful adversary can become public intellectual work.

David Pinsof explains why these targets are attractive. They are rivals, political opponents and institutions competing over authority. Aboujaoude explains what the medium does once the target has been selected. It reduces friction, rewards compression and performance, makes the audience visible, and habituates an uninhibited public self.

Posted in Sociology | Comments Off on What Percentage of Philip N. Cohen’s Blog Passes the Front Page Test?

Paul Bloom’s Small Potatoes Archive

Posted on August 16, 2026 by Luke Ford

I worked through the free Small Potatoes archive from Bloom’s first post in August 2023 through the current August 2026 material, including the re-released pieces and some comment threads where his replies clarify what he thinks he is doing.

Bloom has undergone a progressive de-editing. He begins as an eminent psychologist experimenting with an informal notebook. He discovers that he likes writing without gatekeepers. Then he discovers that readers especially like him when he enters forbidden territory. He notices that incentive and worries about it. Instead of retreating, he devises ways to manage it. By 2025 he has become a Substack-native public intellectual, one who deliberately chooses propositions that initially sound disreputable and then tries to make them respectable through argument.

The danger is built into that form. Bloom’s characteristic move has become front-loaded provocation followed by back-loaded qualification. The whole essay can be quite careful while the first three paragraphs are radioactive.

The opening post in August 2023 says Small Potatoes begins because a writing project “fell through suddenly and unpleasantly.” A friend suggests Substack. Bloom likes the prospect of writing and receiving feedback “without editors and other gatekeepers.” At the same time, he goes out of his way to praise his editors at The New Yorker and The Atlantic, who, he says, made his essays much better. What he wants is another register, a “less constrained” writing.

The first six months feel like an intellectual notebook. There are pieces on implicit bias, trigger warnings, plagiarism, envy, teaching, giving talks, books, the experience machine, AI morality, productivity and social media. The politics are usually incidental. Even where the topic is politically loaded, Bloom’s instinct is anti-coalitional. On implicit bias, people who dismiss it as woke nonsense are wrong, and social psychologists who overstate the evidence and smear critics are wrong too.

That is the basic Bloom move before Small Potatoes. Take a dispute organized into two camps and tell both camps that their conceptual categories are lousy.

By November 2023 he is already contrasting Substack with Twitter. Twitter has become nasty, unproductive and mob-driven, while Substack gives him thoughtful conversation. He says Small Potatoes is taking far more time than expected simply because he enjoys writing it and talking with readers.

Then something significant happens. In March 2024 he says he spends at least an hour on it every morning, has 34 posts in draft, and works on it “more than any other single project in my life.” The money comes afterward. He has fallen in love with the activity first. He also introduces a safety mechanism. Longer and more controversial pieces can first go to paid subscribers for criticism before he gives them to everyone else.

By the first anniversary he has published 54 pieces plus two guest posts, accumulated more than 10,000 free subscribers, and explains the appeal in three words: “Writing is how I think.” A year later he reports more than 19,000 free subscribers and repeats the same formulation.

The major turn comes in late 2024. “Progressives should worry more about their favorite scientific findings” argues that politically congenial findings may deserve extra skepticism when journals themselves possess political preferences. It becomes a hit. The Chronicle of Higher Education asks to reprint it.

Bloom then diagnoses his own incentives. In “This and That (11)” he says the post succeeded because it was a culture-war post critical of a “woke” tendency in science. He admits liking the attention. He says he is “usually on the anti-woke side.” More importantly, he dislikes the tribal response these subjects generate. He wants readers thinking, “Huh, interesting,” not cheering because he has humiliated the enemy. He concludes that he will keep writing about culture-war questions, “but not too often.”

That is the hinge in the history of Small Potatoes. Bloom sees audience capture happening in real time.

And then, he partly succumbs to it and partly resists it.

The 2025 newsletter is noticeably more combustible. There is the humiliation theory of American politics, embryo selection and AI love, professorial cowardice, sexual desire and the Internet, sex differences and academic truth-seeking, and eventually “Wokeness and effective altruism.”

Yet it never settles into the conventional anti-woke Substack groove. He still writes the trigger-warning piece in a way that essentially sides with a practice many heterodox writers despise. He attacks male intellectual culture immediately after criticizing female intellectual culture. He takes seriously the attractions of AI companionship as well as its dangers. He criticizes professors for timidity while repeatedly acknowledging that professional retaliation can make caution rational.

The 2026 writing broadens again. There is religion and the alleged God-shaped hole, consent, immortality, parenthood, books and AI. His recent “Death by AI” is a good example. The hook is inflammatory, involving deaths caused by AI systems, but the argument is a lesson in denominators and control groups. Don’t ask whether AI ever kills or harms somebody. Ask whether it does so more frequently than the available human alternative.

So I would describe the three years this way. 2023 is discovery. 2024 is audience formation. 2025 is boundary testing. 2026 is consolidation. He now possesses a house style and a significant audience. His Substack might be the most thrilling thing in his life.

The easiest way to characterize Bloom’s Substack prose is to compare it with his magazine prose.

Bloom gives us the experiment. In “Credit the editors”, he reproduces a much-praised passage from one of his New Yorker articles and reveals that Henry Finder substantially wrote the sentences people admired. He says reading one of his edited New Yorker pieces is like looking into a mirror and seeing a better-looking version of himself.

Small Potatoes is the unretouched Bloom.

He is less elegant there and more alive. The prose is full of “bullshit,” “shit,” “fuck,” jokes about sex, parenthetical asides and self-mockery. He makes himself a comic character. The old Canadian at the sex debate. The professor who likes his colleagues but thinks they are timid. The writer who discovers that the beautiful sentence everyone praises isn’t really his sentence.

His best natural unit is the seminar intervention. “Okay, but what exactly do we mean?” “What would follow if that were true?” “Suppose we reversed the case.” “What is the comparison group?” “Who would be surprised if the experiment came out the other way?”

He likes thought experiments because they permit him to strip away moral coloration. The parenthood essay proceeds by imagining when you would want Hitler to have a safe car ride. Answer: when your children are in the back seat. The point is clear once he has made the case grotesque.

He also loves inversion. Maybe trigger warnings are good for reasons having nothing to do with trauma. Maybe professors are politically radical but temperamentally conservative. Maybe female academic culture damages truth-seeking and male academic culture damages it even more. Maybe AI’s capacity to eliminate loneliness is itself the problem. Maybe immortality isn’t terrifying but obviously desirable. Maybe developmental research can be technically excellent and a bad idea.

That is why so many titles have the structure of a dare.

His other important technique is confession. He tells you where he was wrong, where his own work has committed the sin he is criticizing, what his editor changed, what argument he abandoned, what he hesitated to publish. In the developmental psychology essay, after complaining that researchers run theoretically pointless studies because they can publish them, he admits that he and his students have published exactly that.

This lets him be harsher because he places himself within the class being indicted.

There is also much visible pleasure. He likes it. He really likes it.

My rough judgment is that about four-fifths of the free corpus passes the front page test easily. Perhaps another 10 to 15 percent passes if the quote contains enough of the argument. A relatively small handful contains sentences or titles that I suspect would make Bloom’s stomach drop if he woke up and saw them isolated beneath a photograph on the front page of The New York Times.

He has increasingly adopted a prose technology that is vulnerable to context collapse.

An example is “Are women unsuited for the pursuit of truth? Yes. But so are men.” Bloom says that what Helen Andrews calls “female modes of interaction” are not optimal for truth-seeking. Only afterward does he argue that male modes may be still worse, because aggressive status competition discourages honesty, collaboration and timid but brilliant scholars.

The full essay is an argument against organizing inquiry around either sex stereotype. But “Paul Bloom: female modes of interaction are not optimal for the pursuit of truth” is also an accurate quotation of his position. That is a front-page-test failure in the narrow reputational sense. He would desperately want paragraph 12 attached to paragraph 5.

The embryo-selection essay is another. He is skeptical that current polygenic screening can deliver what proponents promise, but if it eventually can reduce serious disease or increase desirable characteristics, he sees no categorical moral objection to parents using it. Put accurately but brutally on a front page, “Prominent psychologist defends selecting embryos for intelligence” would produce a different social event from the Substack essay.

His argument about sexual desire has the same vulnerability. He discusses the large rise in bisexual identification among younger liberal women while distinguishing identity and behavior from underlying sexual preference. Whatever one thinks of the evidence, a few accurate sentences from this discussion could travel through academic social media stripped of the distinction on which his argument depends.

“Nobody can touch you without your consent” is engineered for context collapse. The essay begins by affirming bodily autonomy, then investigates circumstances where the maxim cannot be universal. The title plus “there are exceptions” is far more combustible than the discussion. Significantly, it first appeared for paid subscribers and only later emerged publicly in revised form.

“Why are so few professors troublemakers?” has less moral danger but considerable professional danger. Bloom says professors are not very brave and traces some academic success to conformity and risk aversion. Yet here the experiment has already been run. The piece was reworked for The Chronicle of Higher Education as “Why Aren’t Professors Braver?”

In fact, there is a literal answer to the front-page question. “Why Aren’t Professors Braver?” became the cover line of the October 31, 2025 issue of The Chronicle of Higher Education.

And his progressive-science argument was also republished by The Chronicle. His January 2026 essay about professors resisting institutional change appeared there too.

So Bloom’s high-wire material is repeatedly passing through a professional gatekeeper after originating on Substack.

There is a delicious irony in the embryo-screening post. Before attacking quotations from two New York Times articles, Bloom pauses to acknowledge that quotes may have been taken out of context and that people sometimes regret things they tell reporters. He understands the front-page test. He simply keeps writing sentences that create the risk.

Bloom shows risk awareness. He pretests controversial essays with paid subscribers. He sometimes publishes them there first and later releases a revised public version. He tells readers when he hesitated before publishing something. He solicits attacks on his argument. His comment section is partly a homebrew peer-review mechanism.

The developmental-psychology piece ends with him admitting that he hesitated because he did not want to anger colleagues. The professors essay was initially paid-only. The consent essay was initially paid-only. He is distinguishing levels of exposure.

But he is also in an extraordinarily favorable position for this experiment. He is a senior scholar with decades of reputation, a Yale emeritus chair, a full professor at Toronto, an extensive publication record and a large public readership. A 30-year-old assistant professor could not write this Substack with the same expected costs.

Bloom knows this. Much of the professors essay is about how professional dependence creates caution.

His deepest fear is becoming stupid through tribalism.

That is what makes “This and That (11)” so important. He does not say, “Culture-war posts might get me canceled.” He says they attract the wrong cognitive and emotional response. They make writers and readers tribal, and favorable responses can be as corrupting as hostile ones.

I find no evidence that his Substack has damaged his standing. His position has changed.

The University of Toronto presently lists him as a professor in psychology and notes his Yale emeritus professorship, awards, popular writing and editorial work. Cambridge University Press presently lists Paul Bloom as the editor of Behavioral and Brain Sciences, a high-level journal in psychology and cognitive science.

Toronto’s psychology department features his provocatively titled 2024 Theory and Society paper, “Much of developmental psychology is not worth doing.”

Someone encountering Bloom primarily through Against Empathy, Psych, Yale lectures and New Yorker pieces would once have classified him primarily as a prominent cognitive and developmental psychologist who also happened to be a good writer. Someone encountering the 2025 Small Potatoes might classify him as a heterodox public intellectual who happens to be a prominent psychologist.

That shift could carry social costs among some colleagues. I cannot document those private costs. It may also carry benefits. He now has an audience that will follow him from embryo screening to developmental methodology to immortality. He gets podcast invitations. The Chronicle comes to him. He can test an idea on thousands of intelligent readers before deciding what to do with it.

At his career stage, that is probably a favorable exchange.

Why does he write it? His stated explanation turns out to be convincing. He writes to think.

His 2025 account of writing “The End of Loneliness” describes discovering through writing that the structure he had pitched was bad. Putting the argument into sentences revealed “an awful lot of bullshit,” forcing him to rethink it. He regards writing as an epistemic procedure.

Substack gives him a permanent supply of occasions to perform that procedure.

There are secondary motives. Autonomy matters. Feedback matters. Fun clearly matters. Money eventually mattered. Audience matters. Being invited onto podcasts matters. The migration away from Twitter matters. But I don’t think any of these explains an eminent psychologist waking up every morning and spending an hour on 34 unfinished posts.

The better explanation is that Bloom has accidentally constructed an intellectual gym for himself.

That also explains the extraordinary range. A conventional academic must decide whether an idea is worth a paper. A New Yorker writer must decide whether it is worth pitching. On Small Potatoes, Bloom can ask whether immortality would get boring and have the argument in front of readers a few mornings later. His immortality essay is an example of this kind of intellectual play.

Is the Substack seeping into his academic work? Yes.

In July 2024 he publishes “A lot of developmental psychology isn’t worth doing.” Later in 2024 Theory and Society publishes Bloom’s “Much of developmental psychology is not worth doing.” It is substantially the same intellectual intervention, now in scholarly form.

That is visible traffic from the newsletter into the scholarly literature.

There is another form of traffic around AI. The newsletter repeatedly develops arguments about AI companionship, loneliness, effort and the dangers of making life frictionless. In February 2026 Bloom, Emily Zohar and Michael Inzlicht published “Against frictionless AI” in Communications Psychology. Its thesis is Bloomian: AI’s apparent virtue, removing difficulty, can itself be a vice because difficulty contributes to learning, meaning and development.

The academic version is properly referenced and considerably more formal. Bloom has not started writing “this shit happens all the time” in journal articles.

What is seeping across is the question selection.

The titles are becoming thesis-forward. The questions are larger. The willingness to tell an entire field that much of its work is pointless is very Substack. So is the preference for a counterintuitive proposition simple enough to state in one sentence.

Substack changes his scholarship at the conceptual level.

And that may be productive. One chronic disease of academic writing is that the paper has been technically executed before anybody asks whether its question is interesting. Small Potatoes forces Bloom to compete for a reader’s voluntary attention every week. That selects hard for consequential questions.

The risk is that it also selects for controversial ones.

Bloom has created a controlled experiment in audience capture while writing about audience capture.

His September 2023 “Your followers might hate you” is about how social-media feedback teaches writers what gets rewarded and punished. A year later he publishes an anti-woke-adjacent science post, watches the subscriber and engagement numbers jump, and immediately tells his readers that he knows exactly why it happened and distrusts the incentive.

Then he goes on to write more pieces that activate the same audience. That makes him a fascinating case. The audience is changing the distribution of subjects he chooses. I doubt that without Small Potatoes Paul Bloom would have spent this much public time in 2025 on wokeness, female and male intellectual styles, professors’ cowardice, bisexual identification, embryo selection and culture-war epistemology.

But the audience has not yet captured the conclusions.

He frequently starts from territory favored by the heterodox and refuses to deliver the expected ending. Trigger warnings can be good. Male modes of intellectual interaction can be worse than female ones. Anti-woke people can become every bit as tribal and stupid as woke people. AI can alleviate loneliness and damage us by doing so. Professors can be timid because timidity is a rational response to their institutions.

So the thing I would watch is whether the distance between the provocative title and the qualified conclusion begins to shrink. Right now the distinctive Small Potatoes formula is:

Here is something you’re not supposed to say.

Actually, there is something to it.

But not for the reasons the people who usually say it think.

And the opposite side has a point too.

Now here’s a weird psychological experiment.

That formula protects him while he produces interesting work.

Virtually You: The Dangerous Powers of the E-Personality

Virtually You applies to Bloom as a theory of amplification and feedback. Bloom seems to have developed a distinct “e-personality” that is recognizably him but more uninhibited, more provocative and increasingly capable of feeding back into Paul Bloom the academic.

Aboujaoude argues that online activity can generate an identity that is an extension of the ordinary person that is “more assertive, less restrained” and especially useful for voicing unpopular opinions. That is the difference I saw between journal-and-magazine Bloom and Small Potatoes Bloom.

Bloom presents the Substack as more authentically himself because the editors have been removed. But Aboujaoude’s point is that “more authentic” does not necessarily mean “more representative.” The medium selects and amplifies certain components of the person. The Bloom who swears, tells sex jokes, says that much of developmental psychology is not worth doing, asks whether women are unsuited for truth-seeking and casually wanders into embryo selection is Bloom with several inhibiting institutions stripped away.

This complicates something I said earlier about “de-editing.” Aboujaoude gives us a psychological version of the same idea. Bloom is producing a new version of himself through repeated participation in a medium that rewards a particular version.

Bloom says that Small Potatoes rapidly became a major part of his daily intellectual life. Once he is spending an hour every morning writing it, carrying dozens of draft posts and then appearing on podcasts generated partly by that writing, the online persona is no longer an occasional costume. Aboujaoude’s central concern is that online traits can migrate back into offline behavior. He argues that we may become more daring and less diplomatic away from the screen because behaviors acquired online become incorporated into ordinary personality.

That suggests a stronger hypothesis than the one we were considering before about the Substack seeping into Bloom’s scholarship. It may not merely be that Substack gives Bloom ideas for academic papers. It may be training Paul Bloom to be a different kind of intellectual.

Consider the title “Much of developmental psychology is not worth doing.” That is an academic argument, but the proposition has the unmistakable architecture of a Substack post. It is declarative, provocative, compressible and immediately legible to an intelligent nonspecialist. His recent academic AI work has some of the same quality. The argument comes first. The literature and qualification follow.

Aboujaoude calls this merging of online and ordinary identity “virtualism.” His idea is that repeated online experience makes it progressively harder to distinguish the virtual personality from the offline one.

The second powerful application is audience feedback. Bloom has supplied us with evidence. He publishes the piece about progressives being insufficiently skeptical of politically congenial science. It takes off. He gains attention. Then he writes that he enjoyed the attention and worries that culture-war writing encourages tribal responses. He knows what the reinforcement is doing.

Aboujaoude wrote long before the mature Substack economy, but he notices the importance of visible followers and instant reactions. Follower counts transform popularity into a publicly visible metric and can suggest that greater popularity means greater authority. Substack has vastly improved the mechanism. The writer now receives subscriber counts, likes, comments, shares, recommendations, podcast invitations and sometimes money.

That means Bloom doesn’t merely discover what he thinks by writing. He discovers which version of Paul Bloom other people most want. And we have direct evidence that he notices the distinction.

This gives us a potentially much better formulation than “audience capture.” Capture sounds binary. Bloom has not been captured. His conclusions remain unpredictable. A better description might be audience selection pressure.

Imagine Bloom has twenty possible essays in his head. One is about an obscure puzzle in developmental cognition. One is about a novel. One is about whether female social norms undermine truth-seeking. The audience does not have to change his opinion about women to change his intellectual output. It merely has to make that essay ten times more rewarding to publish.

The result can be complete sincerity plus a profound shift in output. That is why the trajectory we found matters. Bloom’s anti-woke-adjacent post succeeds spectacularly. He recognizes why it succeeded. He worries about the effect. Yet the subsequent distribution of subjects becomes more culture-war-adjacent.

The third application is the difference between authenticity and performance. Bloom’s Substack persona feels extraordinarily authentic. It is confessional. He admits mistakes. He makes himself the butt of jokes. He tells us when he was scared to publish something. He discusses his editors, colleagues and intellectual failures. That creates the impression that we are finally meeting Paul Bloom rather than Professor Paul Bloom.

But every durable public persona becomes partly performed simply through repetition. Bloom learns that the recognizable Small Potatoes Bloom is curious, mildly naughty, amused by taboos, impatient with pieties, willing to say “bullshit,” politically difficult to classify and delighted by counterintuitive arguments. He repeatedly selects those authentic parts of himself for display.

Eventually the selected self can become more coherent than the person. Bloom’s e-personality can be a real subset of his personality that becomes disproportionately vivid through repetition. Aboujaoude describes the online personality as potentially “bolder, stronger, and more efficient” than its offline source.

I think that description fits Bloom, and Podcast Bloom strengthens this effect. Podcasts partially collapse the online/offline distinction. We hear his voice, timing, laughter and hesitation. He cannot endlessly revise himself as he can on Substack. That makes the persona seem even more authentic. But podcast incentives still favor the memorable formulation, the surprising answer and the amusing transgression. Bloom is good at all three.

So his Substack may train the propositions, while podcasts train their oral performance.

The fourth application is reputational permanence, and this connects directly to the front-page test. Aboujaoude argues that we disclose things online while underestimating the extent to which other people will construct their picture of us from those disclosures. Somebody’s politics, jokes or casual remarks become employment and reputation data. He also emphasizes that online material is difficult to retract because copies and archives survive deletion.

Bloom has created a several-hundred-thousand-word permanent deposition of Paul Bloom. That is different from publishing five books. A hostile reader investigating one of his books has several hundred carefully edited pages to search. A hostile reader investigating his online life has hundreds of informal judgments, jokes, qualifications, abandoned thoughts, discussions of sex and politics and occasional deliberately provocative sentences. Almost any future controversy involving Bloom can therefore be accompanied by a search through Small Potatoes for the worst accurately quotable sentence.

That is the Aboujaoude problem. That also explains why Bloom’s precautions interest me. Paid-first publication, soliciting objections and revising before making something public are technological substitutes for the editors he removed. Bloom has recreated a little gatekeeping institution inside his supposedly gatekeeper-free project. That suggests that he understands the danger very well.

The online Paul Bloom is more self-centered in subject matter and less self-important in tone. Academic Bloom suppresses most of Paul Bloom while surrounding what remains with the authority of Yale, Toronto, peer review, citations and disciplinary conventions. Substack Bloom talks constantly about himself, his reactions, what he read, what annoyed him, what he got wrong and what he thinks. Yet he sounds less grandiose because he jokes, confesses ignorance and exposes the machinery behind his authority.

So the e-personality simultaneously enlarges the self and punctures it.

The interesting question raised by Virtually You is: Which Paul Bloom is now training which?

In 2023, Professor Bloom seems to have created Small Potatoes Bloom as an outlet where he could write without the restraints governing Professor Bloom. Three years later, Small Potatoes Bloom has a large audience, a recognizable voice, a daily claim on his time, podcast appearances and demonstrable routes back into his professional writing.

At some point the causal arrow can reverse.

That gives us a testable extension of the trajectory essay. Compare Bloom’s scholarly papers, magazine essays, lectures and podcast appearances from roughly 2010-2022 with those from 2024-2026. Code for provocative thesis-first titles, first-person disclosure, taboo examples, profanity, culture-war subject selection, explicit moral judgments, sweeping field-level claims, humorous asides and the characteristic Bloom inversion in which an apparently outrageous claim is introduced and then qualified. If these features rise outside Small Potatoes after 2023, we would have evidence for exactly the phenomenon Aboujaoude describes: not merely an academic acquiring an online persona, but an online persona beginning to remake the academic.

I did a first-pass version of the test, and the result is that Small Potatoes has intensified, routinized and given a distribution system to traits that were already strong in Bloom. It has also become an incubator from which some of his work now moves directly into academic and prestige-media venues.

I did an exploratory coding of 40 items, divided among scholarly publications, magazine essays and substantial podcast appearances before and after the launch of Small Potatoes. I separately checked his public-lecture record. I coded the nine things we identified: provocative thesis-first title, first-person disclosure, taboo examples, profanity, culture-war subject matter, explicit moral judgments, sweeping field-level claims, humorous asides and the Bloom inversion in which he begins with an apparently outrageous proposition and then qualifies or partially reverses it.

I would treat the numbers as descriptive rather than statistical.

The crucial negative control is old Paul Bloom.

Long before Substack, Bloom was already writing titles such as “Against Empathy,” “People Don’t Actually Want Equality,” “The Irrational Idea That Humans Are Mostly Irrational,” “Empathy for Trump Voters? No, Thanks,” “Morality Is Not Always Good,” “Arguing With the Vampire” and “The Paradox of Pleasurable Fear.” His 2012 Chronicle of Higher Education contribution was literally titled “Free Will Does Not Exist. So What?”

His 2014 Boston Review essay is particularly devastating to any simple before-and-after theory. It opens with Bloom telling people he is writing a book about empathy and then adding, “I’m against it.” He immediately jokes that this sounds like announcing that you hate kittens. Then comes the characteristic Bloom maneuver: of course he isn’t against kindness or morality. He thinks rejecting empathy will make us more moral.

That is Small Potatoes structure nine years before Small Potatoes.

Likewise, the public lecturer was already using propositions that would make a university communications office nervous. In 2013 he lectured at UBC under the title “There is nothing special about religion” and concluded that there was little evidence that religion had the special moral significance often attributed to it. His 2014 TED material argued that prejudice is natural, often rational and sometimes moral, before explaining where it becomes destructive. In 2022 he was publicly talking about spicy food, horror, BDSM, hate-reading, going to war and the “joys of suffering” in a University of Colorado appearance.

Again, the outrageous proposition followed by the distinction is not a Substack adaptation. It is a Bloom technique.

The podcasts make the case. In 2019 he went on Modern Wisdom for an episode titled “What Do BDSM & Meditation Have In Common?” In February 2020, Sam Harris and Bloom discussed child sexual abuse, eugenics, loyalty, existential risk and presidential politics.

So when Bloom goes on Tyler Cowen in 2024 and discusses “woke LLMs,” interracial marriage, moral disgust, incest thought experiments and whether wokeness makes aspects of the world worse, it sounds extremely Small Potatoes. But the longitudinal comparison says something important: podcast Bloom was already like this.

My coding bears this out. In the podcast sample, first-person disclosure was essentially universal both before and after Substack. Humor was essentially universal. Moral judgment was essentially universal. Bloom inversions were extremely common in both periods. Taboo examples were actually somewhat more common in my pre-2023 sample. The clear increase was culture-war and current-political subject matter.

That last increase has a confound. Bloom now appears repeatedly in explicitly current-affairs conversations, particularly with Robert Wright and Sam Harris. His July 2025 Harris appearance included Trump and universities, Epstein, MAGA conspiracy theories, Elon Musk and AI companionship. You cannot infer a psychological change from that without accounting for a change in the sort of podcast he is doing.

The scholarly record produces the most surprising result. As a whole, it does not become more Substack-like. In my scholarly sample, provocative titles were actually somewhat more common before 2023 than afterward. First-person disclosure remained rare. Profanity remained essentially nonexistent. Culture-war subject matter remained essentially absent. Sweeping claims and overt moral judgments did not show a general increase. Recent empirical papers still tend to look like empirical psychology papers.

That makes sense when you look beyond the conspicuous titles. A paper such as “Developmental changes in the perceived moral standing of robots” is not Small Potatoes prose. Neither is “Characterizing empathy and compassion using computational linguistic analysis.” Bloom remains perfectly capable of switching registers.

There is, however, one extraordinary exception. On July 29, 2024, Bloom published the Substack post “A lot of developmental psychology isn’t worth doing.” He jokes with the reader, dismisses “civilians,” confesses to having committed the same scholarly sins, says he hesitated because he didn’t want to “piss people off,” mocks ritualized conference questions and makes the sweeping claim that a significant class of developmental research is largely pointless.

In 2024 Theory and Society published Paul Bloom, “Much of developmental psychology is not worth doing.” The scholarly abstract makes essentially the same argument: many developmental age-comparison studies lack adequate theoretical justification and the field should shift toward questions that matter theoretically.

This is the strongest evidence we have.

We cannot establish from publication dates that the Substack post caused the academic paper. The journal article may already have been drafted or under review when the July post appeared. What we can establish is that the same intervention existed simultaneously in his new informal persona and his scholarly production, and that the informal public version appeared before the journal version appeared publicly.

I previously treated the 2026 Communications Psychology essay “Against frictionless AI” as evidence that Bloom’s Substack voice was entering his academic work. That’s too strong. The published authorship statement says Emily Zohar wrote the original draft. Bloom, Zohar and Michael Inzlicht jointly conceptualized it and revised it. So the wonderfully Substack-like title and argument cannot simply be coded as Bloom’s prose migrating into science.

The magazine evidence initially looks much more supportive of the seepage hypothesis. In my pre-2023 magazine sample, roughly three-quarters of the essays had provocative thesis-first titles. In the four substantial 2024-26 pieces I coded, all four did. First-person disclosure rose from roughly three-fifths to all four. Culture-war subject matter went from something like two-fifths to three of four. Sweeping institutional or field-level claims also became considerably more common.

But there is a gigantic problem with treating this as psychological evidence.

Three of the conspicuous post-2023 examples are Chronicle of Higher Education versions of Small Potatoes posts.

“Why Aren’t Professors Braver?” ends by explicitly telling readers that a version originally appeared on Small Potatoes.

“Why Professors Fear the Future” does the same.

The third is “Why Progressives Should Question Their Favorite Scientific Findings.” The Chronicle’s author archive places all three together and makes the shift in his recent subject matter rather striking.

This is important, but it isn’t seepage in the psychological sense. It is syndication.

The Substack persona is escaping the Substack because professional publications are buying it.

Bloom’s 2025 New Yorker article “A.I. Is About to Solve Loneliness. That’s a Problem.” Even the title has the classic Bloom inversion built into it. Solving loneliness sounds good. Bloom argues that this apparent good may deprive people of the painful experiences through which they learn to connect to other human beings.

That certainly feels like contemporary Bloom. But once again the control matters. “Against Empathy” did exactly the same rhetorical thing in 2014. The relevant change isn’t invention of the technique.

And Bloom’s account of producing the New Yorker article argues against crude deprofessionalization. He says his original argument turned out to be bad, that writing forced him to discover this, that he repeatedly rewrote the piece and that Henry Finder performed “wise and meticulous editorial work.”

He is not replacing edited Bloom with Substack Bloom. He increasingly seems to maintain both and move ideas between them.

Aboujaoude’s striking claim in Virtually You is that the online self can become “more assertive, less restrained” than the ordinary person, and that the Internet can make voicing unpopular views easier. He goes further and argues that traits cultivated online can eventually become incorporated into offline personality.

With Bloom, the first proposition fits much better than the second. Paul Bloom went online and found a medium that disproportionately rewards one long-established part of Paul Bloom. He began exercising that part of himself far more frequently. The resulting body of work then created new routes into magazines, podcasts and, at least occasionally, scholarly argument.

My coding suggests a useful distinction between four possible effects.

The first is stylistic transformation. I find little evidence for it. Bloom was funny, personal, provocative and fond of inversions long before 2023.

The second is subject selection. Here I find more evidence. Recent public Bloom spends much more time on ideological bias in science, professorial conformity, academic courage, institutional conservatism, AI politics and related culture-war territory.

The third is frequency. Before Small Potatoes, a Bloom provocation had to become a book, journal contribution, commissioned magazine article, lecture or podcast invitation. Now he can produce one whenever the argument interests him. A trait that once appeared intermittently can become the dominant public representation of him.

The fourth is circulation. Here the evidence is undeniable. Small Potatoes arguments are now being republished by The Chronicle, discussed on podcasts and reproduced in closely related scholarly interventions. The newsletter has become part of Bloom’s intellectual production system.

Small Potatoes did not invent the contrarian Paul Bloom. It industrialized him.

There is also a fascinating Aboujaoude reversal here. His model imagines the e-personality escaping from the Internet into ordinary life. Bloom appears to be doing something more conscious. He has built membranes between the personae. Substack supplies speed, provocation, reader feedback and experimentation. Paid subscribers sometimes function as preliminary reviewers. Podcasts supply conversational testing. The New Yorker restores Henry Finder. Journals restore peer review and disciplinary conventions.

So rather than watching the destruction of Bloom’s gatekeepers, we may be watching him construct a new sequence of gatekeepers.

And that gives us another empirical study to do. Instead of merely comparing “before” and “after,” trace individual ideas across media. Start with every Small Potatoes post that later became or fed into a Chronicle article, New Yorker article, scholarly paper, lecture or podcast. Compare the versions sentence by sentence. Ask what gets removed as an idea moves upward through the gatekeeping chain: profanity, personal disclosure, culture-war framing, sweeping claims, jokes, provocative titles and qualifications. That would let us observe the online Paul Bloom being converted back into professional Paul Bloom in real time.

I found six documentable cases in which a Small Potatoes idea moved into another medium, plus a larger AI-companionship sequence that crosses back and forth among scholarship, Substack, The New Yorker, podcasts and TED. I also found an important reverse-flow control where a New Yorker article moved into Small Potatoes.

The main finding is this: The evidence is much stronger for pipeline seepage than for personality seepage.

Bloom’s professional gatekeepers are not removing the dangerous, funny, provocative Bloom and restoring a sober academic. They preserve most of him. What they consistently remove is the Substack furniture: subscriber appeals, chatter about titles, comments to readers, alternate endings, hashtags, draft-stage digressions and some of the rougher diction. They also improve factual precision and argumentative architecture.

The Chronicle is surprisingly permissive. The New Yorker intervenes much more deeply. Podcasts barely sanitize him at all. The scholarly case is somewhere in between.

I. “Progressives should worry more about their favorite scientific findings” → The Chronicle.

Bloom published the Small Potatoes essay on progressives on October 21, 2024. He later reported that it was a hit, brought likes, comments and new subscribers, and that The Chronicle of Higher Education contacted him “as soon as it was published” about reprinting a slightly edited version. He also diagnosed why the essay had succeeded. It was a culture-war post criticizing a “woke” tendency in science.

Four days later, The Chronicle published “Why Progressives Should Question Their Favorite Scientific Findings.” The publication itself labels the essay an adaptation from Small Potatoes.

This is therefore about as close as we are going to get to putting Bloom’s e-personality through a professional editing machine and examining what comes out the other end.

What disappeared? First, the Substack title was softened slightly. “Worry more” becomes “Question.” The Ginger Rogers subtitle disappears and is replaced with the sober explanatory deck, “Academic journals are not immune from ideological bias.”

Second, the platform scaffolding disappears. The subscription interruption disappears. Bloom’s late discussion of why he chose the title disappears. The Substack contains two endings. Bloom announces that he likes the optimistic ending he wrote while in a cheerful mood, then says his mood has worsened and supplies a darker alternative. That is quintessential Small Potatoes: we see the writer thinking, changing mood and letting the audience watch the composition process. The Chronicle simply integrates the argument into a finished essay.

Third, at least one important empirical claim becomes more careful. The Substack says that the reported Black-doctor effect “disappears” once low birth weight is taken into account. The Chronicle substitutes the more defensible formulation that the effect is substantially weakened and often loses statistical significance. We cannot know whether Bloom or an editor made the correction, but the professional version is plainly more precise.

Now look at what does not disappear.

“That has to be bullshit!” survives. So does “Meh.” So does the argument about racial diversity research. So does the assertion that politically congenial findings deserve extra skepticism. So do the female-mentor controversy, ideological bias in journals, the Trump and Harris examples and the broad criticism of progressive scientific institutions.

The qualifications survive too. Bloom continues to say that he does not want to exaggerate the extent of political bias, that most research has nothing to do with politics and that politically unwelcome findings do get published.

Even the explicit moral conclusion survives. Bloom says that sacrificing accuracy to affiliation makes the world worse and that scientific culture should reward truth-seeking.

So the coding for this case:

Provocative thesis: retained.

First person: retained.

Culture-war framing: retained.

Taboo examples: retained.

Profanity: retained.

Sweeping institutional claim: retained.

Humor: retained.

Qualifications: retained and in one important place sharpened.

Substack self-performance: removed.

This is not sanitization. It is professionalization.

II. “Why are so few professors troublemakers?” → Robert Wright → The Chronicle

This case is even better because we can observe three stages. Bloom first sent “Why are so few professors troublemakers?” to paid subscribers on May 19, 2025. Eleven days later, his regular podcast with Robert Wright included a paid segment explicitly titled “Why are so few professors troublemakers?”

He later released a mildly edited public Substack version. On September 24, The Chronicle published “Why Aren’t Professors Braver?” and explicitly stated that it originated on Small Potatoes.

The first notable deletion comes before the argument even begins. Bloom’s Substack opens by telling us about his struggle to choose a title. “Why are professors so nice?” sounded self-congratulatory. “Why are professors so timid?” sounded too mean. This little glimpse into Bloom managing his persona disappears entirely from The Chronicle.

What replaces it? Nothing. The Chronicle begins immediately with Bloom saying that he likes professors, that he is one, that his wife is one and that most of his friends are professors. That means the first-person disclosure survives almost untouched. So does the joke that if a drunk attacks somebody in a bar, the drunk is unlikely to be a professor. The wording is virtually identical. Then we get to the material one might expect a professional editor to worry about. Bloom discusses academic taboos involving biological sex, racial disparities in crime and social influences on transgender identity. The Chronicle prints them. He writes that the sort of student who says “fuck you” to professors probably won’t accumulate the grades and recommendations necessary for an elite academic career. The Chronicle prints “fuck you.” “Nature and nurture, baby!” survives. “The importance of not pissing people off” survives. The discussion of academics potentially being punished over views about sex, race, Israel and Gaza survives. Bloom’s own substantive view that some psychological sex differences have evolutionary origins survives. So does his aside that he thinks he is right. The concluding assertion that professors are “in the truth business” and ought not squander tenure by being afraid of whom they “piss off” survives too. So does the psychiatrist and light-bulb joke.

What does The Chronicle do to the title? “Troublemakers” becomes “Braver.” That isn’t obviously safer. In one respect it is morally sharper. “Troublemaker” can be pejorative. “Braver” implies that the problem really is courage. And The Chronicle subsequently put “Why Aren’t Professors Braver?” on the cover of an issue.

The Robert Wright stage does something different again. It does not edit the proposition. It turns it into conversation. A Substack argument becomes a topic Bloom has to defend spontaneously against another intelligent person. That is a different sort of gatekeeping. It is live adversarial testing rather than textual polishing.

The chain is therefore:

private or semiprivate Substack experiment → podcast stress test → public Substack revision → professional publication.

That looks less like an uncontrolled e-personality leaking out than a new intellectual production process.

III. “Why are so many professors conservative?” → “Why Professors Fear the Future”

This is where I finally found something much closer to sanitization. The January 5, 2026 Substack title is deliberately misleading in a playful way. Bloom knows that professors are overwhelmingly politically liberal. His inversion is that academics are “conservative” about changing the institutions from which they benefit.

Nine days later The Chronicle published essentially the same argument as “Why Professors Fear the Future.” It explicitly identifies the essay as a version of the Small Potatoes piece.

The title “Why are so many professors conservative?” invites a reader to misunderstand Bloom for a moment. “Why Professors Fear the Future” tells the reader what the argument is.

That is a reduction in Substack-style bait. There are small lexical edits too. On Substack, Bloom recalls presenting Yale’s MOOC proposal to faculty and says, “I got my ass handed to me.” In The Chronicle it becomes simply, “It did not go well.” On Substack, rejecting “dumbass ideas” isn’t the same as fearing innovation. In The Chronicle they become “dumb ideas.” The hashtag “#NotAllProfessors” becomes “Not all professors, of course.”

Those are exactly the sorts of changes predicted by the e-personality hypothesis. Internet-native diction gets normalized when the argument enters a professional publication. But the effect is selective. “ChatGPT can quickly handle many of the bullshit tasks that fill a professor’s day” survives. “A hell of an extrapolation” survives. “Being a Democrat is weak sauce” survives. His criticism of “Republican radicals” who want to abolish tenure survives. And most important, the big claim survives. Professors really are conservatives in the sense Bloom means. Their radicalism applies to ideas and politics, while their conservatism appears when their own institutional arrangements are threatened.

This gives us the clearest picture yet of the filter. The editor doesn’t remove the provocation. The editor removes some evidence that the writer is enjoying being provocative.

IV. “A lot of developmental psychology isn’t worth doing” → Theory and Society → Two Psychologists Four Beers

This is the most important academic case. The Substack essay appeared July 29, 2024 under “A lot of developmental psychology isn’t worth doing,” with the joking subtitle “Not your studies, of course. Yours are great!” The Substack version is swimming in Bloom persona. He warns that 99 percent of his subscribers won’t care, tells the “civilians” to leave, then announces, “Ok, the civilians are gone. Now it’s just us.”

He describes developmental research as a recipe. He imagines challenging speakers during question periods by asking who cares whether some capacity appears at age three or five and what theory would change depending on the answer. He confesses that this same approach earned publications for him and his students. Finally he admits that he hesitated to publish the post because he wondered, “Why piss people off?”

Then something remarkable happens. Theory and Society publishes “Much of developmental psychology is not worth doing” on October 28, 2024. It is classified as a Brief Communication and runs only five pages.

The journal title is not safer. It is stronger. “A lot of” becomes “Much of.” The comic subtitle disappears, but the indictment remains right in the title.

The abstract is conventionally academic. It speaks of age-comparison studies, theoretical justification, research priorities and developmental mechanisms. Quotations from the published paper reproduced by another researcher show that some of the characteristic Substack rhetoric survives inside the article as well. Thomas Chazelle describes the paper as having a “going for the throat” style and reproduces Bloom’s discussion of the recurrent conference-talk recipe.

That matters. We aren’t just seeing the same abstract proposition re-expressed in disciplinary language. At least some of the conversational attack survives the journal boundary.

Two weeks after the paper appeared, Two Psychologists Four Beers released an episode called “Psychology Worth Doing.” The description says Bloom discusses his recent paper arguing that much developmental psychology is not worth doing. Then, in the same episode, they discuss whether progressive-friendly social-science findings deserve extra skepticism. In other words, two recent Bloom provocations, one that reached Theory and Society and one that reached The Chronicle, are recombined in conversation.

This begins to look like an ecosystem.

V. “Be Right Back” → EconTalk

This one gives us a tight chronology. Bloom published “Be Right Back” on March 25, 2024. Its subtitle called AI simulations of people both “a blessing and an abomination.” Two days later Russ Roberts recorded an EconTalk episode with Bloom. At 1:09 Roberts explicitly says the topic is Bloom’s recent Substack essay and repeats the “blessing and an abomination” formulation. Bloom responds that he likes those two words and their Biblical resonance, then develops the same scenario about AI simulations of living and dead people.

Here there is essentially no de-provocation. The podcast host takes the Substack hook and elevates it into the governing hook of the conversation. What changes is that Bloom now has to answer questions. His thought experiment acquires objections, distinctions and spontaneous qualifications. The medium generates caveats through dialogue rather than editing.

Podcasts look more like ordinary human interaction than written social media. We hear Bloom hesitate, laugh and respond to another person. But the subject has already been selected by the online Paul Bloom.

The Substack persona is supplying the agenda to the apparently less virtual medium.

VI. “The humiliation theory of American politics” → Robert Wright

Bloom published “The humiliation theory of American politics” on January 13, 2025, building an argument around Trump, Obama and the political consequences of humiliation. His January 16 conversation with Robert Wright included a segment called “The humiliation theory of American politics.” When Bloom posted the podcast the following day, he explicitly said they had discussed humiliation, “the topic of my Substack earlier this week.”

Three days from essay to podcast. Small Potatoes is generating talking points for his public appearances.

VII. The AI-companionship sequence is the most important case, but it destroys a simple causal story

This one does not begin on Substack. On December 29, 2023, Trends in Cognitive Sciences published online “In praise of empathic AI,” by Michael Inzlicht, C. Daryl Cameron, Jason D’Cruz and Bloom. They argued that AI could simulate empathy without some human biases and might improve human welfare, while acknowledging serious dangers.

Then Bloom starts worrying at the subject on Small Potatoes. “Showing and sharing” in January 2024 asks what human pleasures of being known, understood and shown things imply for artificial companions. “Be Right Back” in March develops AI substitutions for friends, partners and the dead. In June he publishes “My friend thinks it’s a good idea for us to spend most of our time with AI companions.” Bloom explicitly says it grew out of “Be Right Back.” He recounts trying to tell a friend how dystopian such a world would be, only to discover that he had trouble explaining what was actually wrong with it. That failure becomes the essay. Part II follows a week later and moves from the advantages to the dangers.

Then something happens. Inzlicht and Cameron publish a guest post on Bloom’s own Substack disagreeing with “Paul’s gloomier takes,” despite the fact that Bloom is their coauthor on “In praise of empathic AI.” Bloom continues the exchange with “Get Real” in November. At the end he explains why he is doing this. Writing helps him get his head straight. He wants to continue his dialogue with Inzlicht and Cameron, and he wants a larger audience to read the argument and give him feedback.

That is evidence for Small Potatoes functioning as an intellectual laboratory.

Then comes The New Yorker.

“A.I. Is About to Solve Loneliness. That’s a Problem” appeared in July 2025. The title is pure Bloom inversion. Solving loneliness sounds obviously good. Bloom will argue that success itself creates the danger.

The article starts from “In Praise of Empathic A.I.” and the debate it caused. But it also draws together concerns Bloom had been worrying over on Substack. The need to be understood. The difference between real and simulated companionship. AI’s endless availability. The possibility that a companion designed to satisfy us might make actual humans less attractive. The magazine article then supplies a much stronger organizing proposition. Loneliness may be corrective feedback.

Bloom compares it to hunger and other unpleasant biological signals. Rejection and loneliness hurt partly because that pain pushes us to alter our behavior and make ourselves capable of sustaining human relationships. An AI that perfectly soothes loneliness might interrupt the feedback loop.

That is a synthesis. And Bloom tells us that directly. On the day the article appeared, he wrote that his original proposed structure wasn’t good. He spent weeks discovering that other ideas worked better, repeatedly rethinking and rewriting. He specifically credits Henry Finder for “wise and meticulous editorial work.”

This is the strongest evidence in the whole study for a professional gatekeeper changing Bloom’s thought.

Notice, however, what Finder does not purge. The first person remains. Bloom tells a domestic story about sitting with his wife and two-year-old while neither would speak to him. The inversion remains. The humor remains. The unsettling scenarios remain. Elsewhere in the article even “bullshit” survives.

Bloom has written elsewhere that Finder and other magazine editors sometimes improve both his ideas and his individual sentences. His description is that reading a New Yorker piece under his name can feel like looking in a mirror and seeing a better-looking version of himself.

That is the perfect counterpoint to Aboujaoude. Aboujaoude imagines a more assertive virtual self progressively affecting the offline person. Bloom describes an editor taking the virtual or unedited self and producing a better-looking version of it.

Then the idea moves again. In February 2026, Emily Zohar, Bloom and Inzlicht published “Against frictionless AI” in Communications Psychology. Its argument is that AI’s ability to eliminate effort and unpleasant friction can be harmful precisely because struggle, frustration and corrective feedback can generate learning, meaning and growth. Loneliness is one of the forms of friction in the argument.

But here we need a major control against overclaiming. The authorship statement says Zohar wrote the original draft. Bloom, Zohar and Inzlicht jointly conceptualized the paper and jointly revised it. So we cannot call its prose Bloom’s Substack voice entering scholarly publication. What we can call it is conceptual migration through a collaborative research network.

Then at TED2026, Bloom made the “counterintuitive case” that AI companionship may cure loneliness at a cost. Eleven days later his Robert Wright podcast listed “Paul’s TED talk on AI and loneliness” as a segment.

So this lineage is not: Substack → real intellectual work.

It is: academic collaboration → Substack exploration → reader and colleague criticism → more Substack exploration → New Yorker synthesis → academic collaboration → TED → podcast.

That is a productive feedback loop.

In December 2023 Bloom published “We don’t want moral AI” on Small Potatoes. The subtitle is “What I argued in my New Yorker article (and more).” He explicitly describes it as a shorter, snappier version of the magazine piece with additional speculation suitable for his “small and friendly audience.”

When Bloom moves from professional Bloom to Substack Bloom, he says the new version becomes “shorter and snappier” and permits additional speculation. That is an experimental manipulation of register. It also cautions us against imagining a one-way contamination process. Bloom moves material in both directions. The same individual can take a polished magazine argument and de-formalize it for Substack, then take a Substack argument and professionalize it for The Chronicle.

Across the three direct Small Potatoes → Chronicle pairs, the pattern is consistent. All three retain the core provocative thesis. All three retain substantial first-person material. The politically sensitive examples remain when they are central to the argument. Sweeping field-level or institutional claims remain. Humor remains. Qualifications remain and sometimes become stronger. Even profanity survives surprisingly often. “Bullshit,” “fuck you,” “pissing people off,” “hell of an extrapolation” and “weak sauce” all make it through professional publication somewhere in these three essays.

What reliably disappears? The subscription solicitations disappear. Talk about the process of choosing a title disappears. Hashtag syntax disappears. “Here’s another ending because my mood changed” disappears. References to future Substack posts and “teaser trailers” disappear. Some casual vulgarity is downgraded. Reader-address and comment-section choreography are reduced. Factual claims are occasionally tightened. And the argument becomes more linear. This lets us distinguish three kinds of gatekeeping.

The Chronicle mostly performs subtraction. It strips off the platform furniture while leaving the Paul Bloom argument remarkably intact. The New Yorker performs architecture. It can cause Bloom to discover that his original organizing idea doesn’t work and rebuild the article around a better one. Bloom himself says that happened with the loneliness piece. Podcasts perform stress testing. They take a Substack proposition and force Bloom to defend, modify and elaborate it against another mind in real time.

Scholarly publication performs genre conversion. The developmental-psychology case retains a surprisingly aggressive thesis and some conversational rhetoric, while converting the overall intervention into something citable as a journal contribution. The AI case shows that the resulting scholarship may also be collaborative enough that we should stop talking about a single Bloom “voice.”

Aboujaoude’s original mechanism in Virtually You is still relevant. He argues that the online personality can become bolder and less restrained, and that traits cultivated online may eventually become incorporated into offline behavior. But Bloom requires a revision. The evidence does not show a reckless digital Paul Bloom escaping containment and gradually conquering Professor Bloom. It shows something else. Bloom has created an experimental intellectual self that professional institutions increasingly use as an upstream source.

Small Potatoes supplies ideas quickly. It lets him make the thesis maximally visible. It allows jokes, overstatement, personal confession and rhetorical risk. Comments and paid subscribers supply criticism. Podcasts supply argument. Then editors or scholarly collaborators can select from that material. In other words, the online persona has become R&D.

That is why “seepage” is slightly misleading. Seepage suggests an accidental leak.

Much of this is deliberate transmission.

And one of the biggest surprises in the study is that the professional world apparently wants much of the e-personality. The Chronicle does not take Bloom’s inflammatory essay about academic cowardice and make it bland. It publishes the sex-and-race taboos, the “fuck you,” the “pissing people off,” the jokes and the accusation that professors are timid. It simply removes the part where Bloom tells his Substack readers that he couldn’t decide what to call the essay.

The market value of Bloom’s professional persona may now depend partly on qualities developed or exercised intensively in his online persona.

That brings us back to the trajectory question. The question is no longer whether Small Potatoes is making Bloom’s academic prose sound like Substack. I don’t think the evidence is strong enough for that. The more defensible claim is that Small Potatoes is changing where Bloom’s professional ideas come from, how rapidly he develops them, which of them receive reinforcement, and how they travel. That is a much larger effect than whether he says “bullshit” more often.

There is one particularly elegant causal loop here: Bloom publishes a culture-war essay. The audience rewards it. Bloom notices the reward and publicly worries that he is being rewarded for culture-war writing. The Chronicle notices the same success and immediately republishes the essay. The professional reward therefore confirms the audience reward. That is audience selection pressure operating through two institutions at once.

The result is not capture because Bloom keeps writing conclusions that frustrate the coalitions his titles might attract.

Small Potatoes has become part of the machinery by which Paul Bloom produces intellectual work. The online version of Bloom generates, exaggerates and tests propositions. Podcasts pressure-test them. Professional editors remove the platform performance and tighten the argument. Scholarly collaborators convert some of the ideas into research discourse. The traffic also runs backward, with polished professional work being loosened and extended on Substack.

That vindicates the most interesting part of Aboujaoude. The boundary is dissolving. He describes the danger as online and offline personality progressively becoming difficult to separate. With Bloom, that dissolution is happening at the level of intellectual production.

By 2026, asking whether an idea belongs to “Substack Bloom,” “academic Bloom,” “podcast Bloom” or “magazine Bloom” is the wrong question. It belongs to a Paul Bloom production system in which Small Potatoes has become a central node.

Posted in Paul Bloom, Psychology | Comments Off on Paul Bloom’s Small Potatoes Archive