Chapter 2: Case Studies and Three Realities

You have probably watched this happen in real time: a scientific claim gets debunked, the researchers move on — and years later the original claim is still circulating as if nothing ever changed. Stand like Wonder Woman for two minutes before a job interview and you’ll perform better. The lead author of that study publicly walked it back a decade ago. It is still being repeated.

This chapter follows four true stories of that kind, each one traced slowly from the lab bench to the version a coworker repeats at lunch. Walking them one step at a time lets us see exactly where along the way a real finding gets lost, a shaky one sails through, or a correction quietly dies in a drawer. The payoff is noticing that the four are not four flukes but four versions of a single pattern — and naming that pattern is what the rest of the book builds on. Near the end, a fifth story — about money — walks the same ground with a different kind of truth, and sets up the distinction the chapter’s second half exists to make.

Chapter 2, Part I. Chapter 1 walked through the pipeline abstractly — the transport pipeline (The Out There → Raw Data → Insight → Theory → News → Meme) stage by stage, then the sibling selection pipeline with one gate per stage. (A gate, in plain terms, is just the test each stage applies to decide what gets through — what’s publishable, what’s newsworthy, what’s shareable.) The abstract picture is correct but loose. Ch 2 brings it to ground in two ways. First, by walking specific findings through the pipeline so we can see what each stage actually does and what each failure mode actually looks like. Second, by being explicit about which kind of reality is in transit, because the pipeline behaves differently depending on whether the content is objective (what’s actually out there), subjective (what someone privately experiences), or intersubjective (what enough people agree on that it becomes real). Harari names that third category cleanly in Nexus, and the rest of the book quietly assumes a reality type without saying which — this chapter pulls the assumption out into the open.

Two reasons for the chapter, in other words. The case studies make the structural argument concrete. The three-realities frame makes its scope explicit. The rest of the book gets more honest about both as a result.

Four pipeline outcomes

Picture the pipeline as a relay race an idea runs from the lab out to the public. An idea can drop out of the race at any handoff — and where it drops out decides what it leaves behind: a myth everyone repeats, a quietly-buried paper, a correction nobody heard, or a claim that jumped into the race halfway and was never based on anything to begin with.

Findings — and not-findings — can exit the pipeline at different stages. Each exit produces a different artifact in the world, and the four exit points the chapter walks through map cleanly to the four stages where the selection gates do most of their work. Said schematically:

• Made it all the way to meme. The finding traversed every stage. The meme-form persists in culture. Often the meme persists even after the corrective signal reaches the upstream stages, which is what makes this case the most diagnostic of the modern environment.
• Died at consensus. The finding was published — passed measurement, analysis, and the publishability gate — but the broader scientific consensus rejected it after independent attempts to confirm it failed. The meme never spread because the consensus gate did its job.
• Died at curation. The finding (often a corrective finding) passed consensus but never made it through journalism, popularization, or the cultural curation layer. The original (often wrong) finding kept spreading; the correction sat in the journals where almost no one outside the field encountered it.
• No theory at all. The “finding” never had a measurement-and-consensus origin. It was manufactured content in the strict sense — injected at the curation or meme stage and decoded by receivers as if it had originated upstream. The downstream gates cannot easily tell measured from manufactured, so the manufactured variant occupies the same stages real findings occupy.

The four case studies that follow each illustrate one of those exits. I’ll walk each one stage by stage, then read across them in the synthesis section.

Case 1: Power posing — made it to meme, the meme survives the corrective

The 2010 paper by Carney, Cuddy, and Yap reported that holding “high-power” poses (hands on hips, chest out, feet apart) for two minutes raised testosterone, lowered cortisol, and improved performance on risk-taking tasks. The paper was published in Psychological Science and presented as an actionable, free intervention with measurable physiological effects.

Each stage of the pipeline handled it as the abstract Ch 1 picture predicts. Measurement recorded saliva hormone levels and risk-task scores from 42 participants. Analysis ran the standard significance tests and reported p < 0.05 effects on both hormones and behavior. Consensus — at the publication-gate level — passed it: the paper appeared in a top journal. Curation embraced it spectacularly. Cuddy’s 2012 TED talk on the work has been viewed over 70 million times. A book deal followed (Presence, 2015). Major business publications adopted the framing. Corporate training programs incorporated power posing as a standard pre-interview intervention. The pipeline ran the full distance.

Meme stage: “stand like Wonder Woman before your interview.” The compressed form lost nearly all of the original’s qualifications — the small sample, the specific cortisol-not-just-cortisol effect, the modest behavioral effect size — and landed as a clean prescription. The compressed form did the work the pipeline rewards: actionable, identity-friendly, frictionless to apply. A 42-person hormone study became a TED-talk-and-bookshelf cultural staple in five years, and at every stage the gates passed the variant they were tuned to pass.

Then the corrective. In 2015 Ranehill et al. ran a much larger replication (n = 200) and found no hormone effects; the behavioral effects didn’t replicate either. In 2016 Dana Carney — the original lead author — publicly disavowed the work and stated she no longer believed the effects were real. The corrective passed measurement, analysis, and the consensus gate; within the psychology community the original finding is now treated as a canonical case of the replication crisis. And the meme kept spreading anyway. The TED talk is still being watched. Presence is still being read. Power posing is still being recommended in business contexts. The corrective signal made it through the early pipeline stages but largely failed at curation — Carney’s disavowal got modest coverage in scientific media and almost none in the general media that had broadcast the original. The meme outlived its evidence base by a wide margin, and continues to.

What this case shows, said plainly: the pipeline can run its full length and produce a culturally persistent meme even when the underlying finding is wrong, and the corrective signal has worse curation fitness than the original — so even when the corrective is right and the original was wrong, the corrective doesn’t catch up. The asymmetry the chapter is going to keep finding is this: at every stage the gates are tuned to pass actionable, identity-resonant, novel content, and corrective findings are systematically less of those things than the originals they correct. This is exactly the selection mechanism Chapter 10 will diagnose in more detail; for now the case study just establishes that the failure mode is real and continuous, not an occasional glitch.

Case 2: Arsenic-life — died at consensus

In December 2010 a paper by Felisa Wolfe-Simon and colleagues at NASA’s Astrobiology Institute appeared in Science claiming that a bacterium isolated from California’s Mono Lake (designation GFAJ-1) could substitute arsenic for phosphorus in its DNA backbone. If true, this would have meant terrestrial life could be built on chemistry other than the standard CHNOPS set — an enormous result for both biology and astrobiology. NASA held a high-profile press conference timed to the publication; the institutional curation layer caught the news and began packaging it for the meme stage (“alien life on Earth,” etc.) within days.

Then the consensus gate fired, harder than the publishability gate had. Within hours of the paper appearing, independent biochemists pointed out methodological problems. Within weeks, multiple labs announced attempts to replicate the central finding. Within eighteen months Science itself published two formal technical comments and a refutation showing that GFAJ-1’s DNA contained no arsenate — the original measurement had been contaminated by free arsenic in the sample preparation. The scientific consensus on arsenic-life settled to negative within a year and a half of publication.

The meme barely propagated past the initial press cycle. “Alien life found on Earth” had a forty-eight-hour spike in coverage, then went quiet as the consensus signal came through. By the time most laypeople heard about it, they heard about it as “that NASA arsenic claim that turned out to be wrong.” The pipeline’s later stages never built the cultural artifact they had been about to build, because the consensus gate caught the variant before curation could compound it.

What this case shows: the consensus stage can and does work. The book’s diagnostic chapters mostly engage cases where the consensus gate failed (Power Posing) or never operated (Astrology); arsenic-life is the worked example where the gate caught the error before the meme could compound. It is worth being explicit about this counterexample because the book’s posture would otherwise read as universal pessimism about the pipeline. The pipeline’s gates can function as designed — and when they do, the failure mode the chapter has just illustrated with Power Posing does not occur. The diagnostic question for the rest of the book is not “do the gates ever work” — they sometimes do — but what conditions make consensus catch errors versus pass them, and the conditions are visibly worse in social-science high-public-interest domains (where Power Posing lives) than in hard-physical-science extraordinary-claims domains (where arsenic-life lived). The political economy Chapter 10 diagnoses is what makes the difference.

Case 3: Stanford Prison Experiment correctives — died at curation

Philip Zimbardo’s 1971 Stanford Prison Experiment is one of the most-cited social-science studies of the twentieth century. Volunteer participants were randomly assigned to be “guards” or “prisoners” in a basement simulating a prison; within days, the guards turned cruel and the prisoners broke down. Zimbardo halted the study early. The finding — that ordinary people will behave sadistically when given a guard role in an institutional context — entered textbooks, documentaries, films (The Experiment, The Stanford Prison Experiment, 2015), and the pop-psychology canon as the canonical demonstration of situational forces overwhelming individual character.

The pipeline ran cleanly. Measurement, analysis, journal publication (in a Naval Research Reviews collection and later more prominent venues), consensus reception (initially favorable), curation (very heavy — Zimbardo became a public intellectual on the basis of the work), meme (“ordinary people become monsters in prison-guard roles”). For four decades the finding occupied the meme stage as load-bearing cultural content.

In 2018 the picture changed. Ben Blum’s Medium reporting documented archival evidence — original tapes and notes — showing that Zimbardo had explicitly coached the guards on how to behave and that one prisoner’s famous breakdown had been performed. Thibault Le Texier’s 2018 book Histoire d’un mensonge (translated 2019 as Investigating the Stanford Prison Experiment) worked through the archival record in detail and concluded that the study was, at minimum, methodologically broken, and arguably constituted academic fraud. Subsequent academic responses — Texier’s 2019 article in American Psychologist, commentary across the social-psychology community — largely accepted the methodological critique. The corrective passed measurement, analysis, and the consensus gate. By the late 2010s, within the social-psychology community, the SPE was understood as a methodologically broken study that should no longer be cited as evidence for the situational hypothesis.

And then the corrective signal died at the curation gate. Textbooks have not been updated at any meaningful scale; introductory psychology courses still teach the SPE as canonical. Pop-psychology books still cite it. The films are not being re-cut to add caveats. The TED-talk circuit has not absorbed the correction. Most non-specialists who have heard of the SPE still believe it shows what it was originally claimed to show, because the curation layer that gave them the original story has not propagated the correction. Even pieces about the replication crisis often skip over the SPE specifically, because the cleaner replication-crisis case studies (Power Posing among them) don’t require taking on a story this culturally embedded.

The mechanism this case illustrates is the out-competition failure mode at the curation layer specifically. The corrective signal has all the wrong curation properties: it is technical (archival methodology critique), unflattering to a popular cultural figure (Zimbardo), structurally disappointing (the dramatic narrative loses its mooring), and offers no clean replacement story. The original meme has all the right curation properties: dramatic, identity-relevant (“could you have been a guard?”), morally clean (a warning about institutional power), and self-contained. An accurate corrective with bad curation properties loses to a wrong original with good curation properties, even when the consensus gate has fully ratified the correction. This is exactly the failure mode Chapter 8 frames as “preservation held, training hollowed” — the journals contain the correction; the receivers who would need to be retrained on it haven’t been.

What this case shows: the consensus stage doing its job is not sufficient when curation has stopped operating as a corrective channel. Even findings that pass consensus by a wide margin can fail at curation if their curation fitness is lower than what they are correcting, and the modern curation layer’s selection criteria (engagement, drama, narrative cleanliness) systematically disfavor corrective findings over the originals they correct. This puts a load-bearing weight on what Ch 12 calls “survivable polarization” infrastructure: curation that can carry corrective signal needs to be designed to resist the engagement-fitness gradient, not ride it.

Case 4: Astrology and the no-theory cases — manufactured content occupying the pipeline

The first three cases all started with a measurement. Power posing measured hormones and risk-task scores; arsenic-life measured spectra; the SPE measured (or appeared to measure) participant behavior. The pipeline can run, can fail, can correct — but in all three the front end was real data of some kind.

Some content occupying the pipeline has no measurement at the front end at all. The most cleanly worked example is the one already in this book’s catalog: astrology. The astronomical positions astrology references are real measurements — accurate, ancient, well-documented. But the interpretive framework — that the positions at one’s birth shape personality and predict events — has no measurement origin. It was never derived from data; it was never tested in any of the senses the consensus stage would recognize; it has never passed any consensus gate at any historical period in any culture where one operated. And yet it occupies all the same downstream stages real findings occupy: it has authoritative-feeling print sources, it has popular-press curation (the daily horoscope column), it has meme-stage cultural integration (“oh, you’re so Capricorn”). The downstream gates cannot tell measured from manufactured because they were not designed to; they select on form, not on provenance.

A related modern case is the “you only use 10% of your brain” claim, which has no traceable measurement origin (the most-cited “source” — William James — said no such thing) but circulates as if it had one. It has been a textbook entry, a movie premise (Lucy, 2014), a self-help framing for decades, and several waves of corrective neuroscience explanations of why the claim is structurally false have not displaced it from cultural circulation. As with astrology, the manufactured variant occupies the same stages a measured finding would, and the consensus-and-curation correctives don’t reach the meme receivers because the corrective signal’s curation fitness is lower than the original’s.

The general shape: manufactured content can populate the later pipeline stages without ever passing through the earlier ones, because the gates between curation and meme don’t check provenance — they check form. Chapter 5b works the mechanism out in detail; this chapter just establishes that the failure is a normal and recurrent feature of the pipeline, not an exotic edge case. Astrology has been in the meme stage for two thousand years; the 10%-brain meme has been there for more than a century. Both are stable equilibria, not transient failures.

Reading the four cases

Stand back from the four cases and a pattern emerges that is worth naming directly.

The pipeline can fail at every stage (Power Posing lost at consensus, even though the journal published it; SPE-corrective lost at curation; arsenic-life succeeded at consensus when most of the rest of the chapter has been about failures there; astrology never entered the pipeline at all). The pipeline can also work at every stage — arsenic-life was caught; the journals have the SPE corrective; Power Posing’s status within the field is now appropriately calibrated. The diagnosis is not “the pipeline is broken” tout court. It is “the pipeline’s gates pass and fail along a specific gradient, set by the political economy of the gate’s host institution and by the curation-fitness of the variant being evaluated, and the gradient runs systematically against corrective signal.”

Plainly: the pipeline doesn’t break at random. It fails in a direction. A correction is almost always less dramatic, less flattering, and less immediately useful than the myth it is trying to replace — so wherever the gates reward drama, flattery, and usefulness, the correction loses the race, every time, no matter how right it is.

That phrasing matters because it tells you what the prescription has to be designed against. Building integration infrastructure that “fixes the pipeline” or “tells the truth louder” misreads the diagnosis. The pipeline tells truth fine when the gates are tuned to reward truth-correlated form; the pipeline distributes manufactured content fine when the gates are tuned to reward form regardless of provenance; the pipeline persistently fails corrective signal even when the corrective is correct because correctives have systematically worse gate-fitness. The four cases are not four failure modes; they are four samples drawn from one continuous failure surface, and the surface’s shape is what the book’s later chapters are diagnosing in formal terms.

The three realities — Harari’s frame

Walking the four cases also surfaces a question the book has been carrying without engaging: which kind of reality is moving through the pipeline? The cases above are all from the natural and social sciences — claims about hormones, bacteria, human behavior, planetary positions. Those are claims about objective reality in the philosophical sense: states of affairs that exist independent of what anyone believes about them. The pipeline can be evaluated against an external check (the next replication, the methodological audit, the empirical record), and the chapter’s diagnostic vocabulary — true, false, corrective, manufactured — applies cleanly.

Most of the book’s existing argument has implicitly been about objective claims of this kind. But the pipeline carries other kinds of content too, and the diagnostic vocabulary works differently for them. Harari, in Nexus, names the three categories cleanly:

• Objective reality. Things that exist independent of any observer or community of observers — atoms, planets, dogs, bacteria, the speed of light. These have a fact of the matter that the pipeline either transmits faithfully or distorts. The four case studies above are all in this category. The book’s diagnostic vocabulary works cleanly here.

• Subjective reality. Things that exist in one mind — pains, dreams, private preferences, qualia. These don’t really traverse the pipeline at all in the sense the book has been using; they are not propagated through institutional gates because they are not communicable in the way the pipeline transmits content. (Reports of subjective states are objective claims about the subjective; the subjective itself is not what moves.) The pipeline analysis mostly doesn’t engage this category, and the chapter mentions it for completeness rather than to do work with it.

• Intersubjective reality. Things that exist because enough people agree they exist — currencies, nations, corporations, brands, laws, religions, names, gods. These are real in a specific structural sense: they have causal consequences (you can be arrested for breaking a law; a dollar will buy you bread), but the consequences exist because the network agrees they do. The pipeline is not transmitting these realities; it is partly constituting them. Money is real because the network believes it is real; the belief is the constitutive mechanism. Brands and nations and religions work the same way.

The intersubjective category is where the book’s diagnostic vocabulary — true, false, corrective, manufactured — starts to misfit, and being explicit about this matters.

A fifth story: money

The four case studies were chosen because each traces an objective claim through the pipeline to a different exit. Before drawing out what the third reality changes in the abstract, it is worth walking one intersubjective truth the same slow way — not as a fifth exit (it will not fit the exit taxonomy, and the misfit is the point) but to get a feel for how differently the same pipeline behaves when the cargo changes kind.

Take the plainest intersubjective truth in circulation: the twenty in your pocket buys groceries. That is true — any register will confirm it. Now compare it to Case 2’s kind of truth. When the arsenic-life claim collapsed, the bacterium was unbothered: GFAJ-1’s DNA contained what it contained, and every replication could appeal to that. If every mind on Earth somehow forgot the whole episode, the phosphorus would still be there to find again. Run the same erasure against the twenty and nothing survives. Strip the agreement out of every head and the bill is ink on cotton — no residue of value sits out there waiting to be re-measured, because the value was never out there. The truth was the agreement. The fact that both sentences are genuinely true while only one has a fact outside the network to lean on is the whole content of Harari’s third category.

That difference stops being philosophical the moment the pipeline touches money in anger. Three episodes, briefly.

A failing currency cannot be corrected — only re-agreed. When a monetary agreement starts to dissolve, there is no upstream fact a corrective could carry back through the gates. In Weimar Germany in 1923, and in Zimbabwe eighty-five years later, people went on calling the notes money — the word held — while spending wages within hours of receiving them, which is the behavior the monetary agreement actually consists of. The words outlived the truth. And the recoveries, when they came, were not corrections of an error but re-constitutions: a brand-new mark, or the wholesale adoption of someone else’s dollar — a population walking out of one agreement and into another. Compare arsenic-life: the consensus gate could check the claim against the bacterium. There is no bacterium under a currency.

Coverage of the truth feeds the truth. A bank run is a claim — the bank is failing — that makes itself true by propagating: holders hear it, act on it, and their acting is the failure. Nothing was distorted in transit; the transit was the event. In 2007 a run on Britain’s Northern Rock still moved at the speed of physical queues on the pavement. In March 2023, Silicon Valley Bank lost tens of billions of dollars of deposits in a single day, the run coordinated in group chats and timelines — the agreement moving at meme speed while the institutions built to steady such agreements still moved at committee speed. For this cargo, the pipeline’s speed is not a transmission statistic. It is part of the truth’s own physics.

And constitution can now be watched from zero. Bitcoin is the rare case where the birth of an intersubjective truth sits entirely on the public record. In 2009, “bitcoin is worth something” was simply false. No measurement changed; no discovery was made; through nothing but accumulating agreement — each new holder one more unit of the thing being agreed — the claim became true, with “internet money”, Chapter 5’s worked compression, doing the recruiting. And when such a network’s agreement later splits, there is no measurement to appeal to: Bitcoin’s block-size dispute, like Ethereum’s dispute over a catastrophic theft, ended in a fork — two ledgers, two communities, both of them right, each truth fully generated by the network still holding it. The fork did for these constitutions roughly what the consensus gate did for arsenic-life, with one difference that should land uncomfortably: it didn’t pick the winner. It set how many truths there are now.

None of this fits the four-exit taxonomy, because the taxonomy assumed a referee — a fact outside the pipeline that correctives could appeal to and gates could fail against. The machinery for content without a referee — what the three truth-regimes translate into, what capture means when there is no truth to tune a gate against, why the fork is the corrective’s structural sibling — is worked out in the intersubjective-truth note; the chapters that follow flag the intersubjective case wherever it changes their conclusions.

What the three-realities frame changes for the book

For objective claims, the book’s existing diagnosis applies cleanly. The pipeline’s gates either pass truth-correlated variants or they don’t; manufactured content either gets caught at the gates or it doesn’t; corrective findings either reach the meme stage or they fail at curation. The whole structural argument can be read as a theory of how truth about objective reality fares in a particular pipeline configuration.

For intersubjective claims, the same vocabulary needs translation. The “truth” of a currency is not something the pipeline transmits and could distort; the network’s agreement on the currency is the currency’s reality. So when Ch 5c says “compression strips the decoding key” and produces preserved / inverted / orthogonal regimes, the analysis applies to objective-claim transit but not directly to intersubjective-reality production. For intersubjective realities, the relevant pipeline failure mode is not truth-loss in transit but capture of the constitutive process itself.

In plainer terms: when the pipeline mangles an objective fact, it garbles a truth that would still be true whether or not anyone believed it — the fact is sitting out there, unharmed, waiting to be measured again. But for things that are real only because we agree on them — money, laws, nations — there is no underlying truth sitting off to the side. Change how the agreement gets made and you don’t distort the reality, you change it. That is the scarier failure, because there is no outside fact left to check the result against.

A captured currency-issuance process produces a different currency; a captured legal system produces different laws; a captured brand produces a different cultural object. The capture mechanism is the same one the capture taxonomy works out — which surface is captured, by whom, and how hard the recovery is — but the consequence is different — for intersubjective realities the captured pipeline produces a different reality, not a distorted transmission of an existing one.

This has three concrete implications for the book:

• The book’s truth-compression argument (Ch 5c) is about objective claims unless explicitly extended. Ch 5c’s three regimes (preservation, inversion, orthogonality) apply to claims that have a key-against-which-to-be-decoded. For intersubjective realities, the analysis is reformulated in the intersubjective-truth note: the analogous regimes are convergent, divergent, and hollow constitution — the decodes compose into one agreement, scatter into camps affirming different agreements under the same words, or keep the words while the constituting behavior evaporates. They are failure modes of agreement-making, not of transmission.

• The book’s training-side critique (Ch 8) cuts harder for intersubjective realities. Training installs the decoding keys; for intersubjective realities the installed keys are the reality the trained receiver participates in. A captured curriculum for intersubjective content doesn’t just produce trained receivers who decode existing content wrong — it produces participants in a different intersubjective reality. The capture-asymmetry the chapter named gets sharper here.

• The book’s political-economy critique (Ch 10) works at both levels but differently. For objective content, the captured equilibrium produces truth-loss in transit. For intersubjective content, the captured equilibrium produces a different shared reality — the captured platform doesn’t just distort the truth, it shapes which “we” gets to exist in the first place. The book’s worry about platform ownership is sharper at this layer, and the prescription (Ch 12) has to engage both modes.

The fuller working-through lives in the intersubjective-truth note — the regime translation, what capture means when there is no external truth to tune a gate against, why the fork is the corrective’s structural sibling, and the volatility that follows from agreement states moving at meme speed. This chapter’s job is to surface the distinction and ground it, so the rest of the book stops sliding between reality types without flagging the slide.

Where I land

The chapter, said whole: the pipeline’s abstract structure (Ch 1) becomes concrete through case studies of specific findings exiting at specific stages — Power Posing made it all the way to meme and survived its own corrective; arsenic-life died at consensus and never propagated; the Stanford Prison Experiment correctives passed consensus but died at curation; astrology and other manufactured cases occupy the late stages without ever entering the early ones. The cases are samples drawn from one continuous failure surface whose shape is set by the political economy of the gates and the curation-fitness of the variants. The book’s diagnostic vocabulary works cleanly for objective claims (where the four cases live), works differently and harder for intersubjective realities (currencies, nations, brands, laws — where the pipeline doesn’t transmit so much as constitute), and mostly doesn’t engage subjective realities at all. Being explicit about which reality is in transit is the chapter’s second contribution, and it sharpens what the later chapters are arguing about.

The deeper claim, more provisional: the pipeline failure surface has a specific gradient — corrective signal has systematically worse gate-fitness than the original it would correct, at every stage where engagement, drama, identity-resonance, or actionability are the gate’s criteria. That single asymmetry is responsible for most of what looks like “the modern environment failing.” Building infrastructure that fixes this means building gates whose criteria reward corrective signal over original signal where the corrective is right, which is a harder design problem than “fix the algorithm” and the one Ch 12 is implicitly designing against.

What this changes for the book

• The four case studies become a recurring touchstone the later chapters can reference without re-introducing. Power Posing as the prototype made-it-to-meme-survived-correction case; arsenic-life as the consensus-worked case; SPE-corrective as the died-at-curation case; astrology as the no-theory case. The later chapters’ arguments get crisper when the abstract claim (“engagement-tuned gates fail corrective signal”) can be illustrated by a specific case the reader already knows.

• Chapter 5c’s scope is objective claims, explicitly. The three-regime model (preservation, inversion, orthogonality) applies to claims with a key-against-which-to-be-decoded; intersubjective content gets its own treatment in the intersubjective-truth note. Of the two defensible moves the book once weighed — bound the scope or extend the model — it has now done both: 5c’s regimes are bounded to objective content, and the constitutive extension (convergent / divergent / hollow) lives in the note.

• Chapter 10’s critique sharpens at the intersubjective layer. The captured platform doesn’t just distort truth about objective content; it shapes which intersubjective reality the network participates in. The political-economy critique cuts twice — at transmission failure for objective content and at constitutive capture for intersubjective content — and the second is, in some ways, the more serious case because it lacks an external check.

• Chapter 8’s training function extends to the intersubjective substrate. Receiver-training as the consumer-key substrate looks even more load-bearing when the trained-key is partly constitutive of intersubjective reality rather than just decoding-equipment for objective claims. The capture asymmetry gets a multiplier when training-substrate capture also shapes what shared reality exists.

• The pipeline-failure-modes catalog is now four named modes. Made-to-meme-and-survived-correction, died-at-consensus, died-at-curation, no-theory-injection. Subsequent chapters can refer to these by name; the failure-mode taxonomy is now part of the book’s vocabulary.

Where I’m still uncertain

• The three-realities frame now has its working-through — but the line is still live. The intersubjective-truth note does what this chapter long deferred: the constitutive regime model, capture as steering the generator, the fork as the corrective’s sibling. What remains genuinely open is the carving itself. The note carves per-claim by generator — truth from agreement vs. truth from reality — which handles the easy mixed bundles (a founding story; astrology’s false personality claims riding on a real membership practice) but has not been defended at the depth a critic would demand for the hard ones: brands’ quasi-objective promises, or whether scientific consensus tracks a generator or quietly is one. The note carries that open edge; this chapter inherits it.

• The four case studies are concentrated in psychology and the natural sciences. Power Posing and the SPE corrective are social psychology; arsenic-life is biochemistry; astrology is folk-cosmology. None of the cases come from economics, political science, history, or the humanities. The money walk-through now grounds the intersubjective case, but it is a walk, not a case study at the four cases’ stage-by-stage depth — and astrology’s double life (false objective claims bundled with a real membership practice, per the note’s per-claim carving) deserves more than the aside it gets. A wider selection would make the chapter’s claims more general; the present one is shaped by what is easiest to write about cleanly and is admittedly narrow.

• The “made-it-to-meme-and-survived-correction” case may not generalize as cleanly as Power Posing suggests. Power Posing is a particularly clean case because Carney herself disavowed and the replication failures were dramatic. Many cases where a finding goes wrong have messier correction stories — the original authors don’t disavow, the replications are partial, the field’s consensus is split rather than settled. The pipeline analysis I have given assumes a clean corrective signal; the more typical case is a noisy one, and the chapter has not engaged that more typical case.

• “Died at consensus” might be rarer than the chapter’s structure implies. Arsenic-life is a famous case partly because consensus rejection of a high-profile published claim is unusual. Most claims that “die at consensus” die quietly — the paper appears, no one engages it much, it accumulates few citations, it never reaches the curation layer because curation looks for high-profile findings. The chapter’s framing treats consensus rejection as an active gate-firing event; the more common reality is consensus attention failure, where the gate doesn’t reject anything because it doesn’t engage anything. That distinction matters and the chapter has not surfaced it.

• The astrology case may be too easy. Astrology has been in cultural circulation for two thousand years; almost no one defends it as a measurement-based account of personality anymore. The harder case for the no-theory exit is current manufactured content whose meme-stage occupancy is recent and whose corrective signal has not yet had time to fail — early-2020s wellness content, AI-generated text passed off as expert commentary, political-influence operations. The chapter cites these in passing but does not work through any of them in case-study detail, partly because the corrective signal for them is genuinely incomplete in real time. A future pass could pick up one or two.

• The corrective-signal-has-worse-gate-fitness claim is asserted, not measured. I have committed to it because the case studies illustrate it and the structural argument is consistent with it. The systematic empirical claim would be: corrective findings on average have lower engagement / curation / meme-stage propagation than the originals they correct. There is some empirical literature on this (work on the spread of misinformation versus corrections, work on retraction visibility versus original-paper citation rates), and the book should engage that literature properly rather than leaning on the four case studies alone.

---

← Chapter 1: The Information Landscape · Chapter 3: The Human Time Budget →