> ## Content Index
> Fetch the complete content index at: https://www.nanosava.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Smart Enough to Fool Ourselves
- URL: https://www.nanosava.com/essays/smart-enough-to-fool-ourselves/
- Published: 2026-11-18T13:00:00.000Z
- Updated: 2026-11-18T12:59:59.000Z
- Description: When a fact starts wearing team colors, being clever stops protecting us. On the walls we keep around what we believe, and how to leave each other a way out.
- Author: Sava
- Tags: Emergent Stupidity

## Two games on one field

On a Saturday in late November 1951, Princeton played Dartmouth at Palmer Stadium. Princeton was unbeaten, and its star, Dick Kazmaier, won that year's Heisman Trophy. In the second quarter Kazmaier left the field with a broken nose and a mild concussion. Later in the game, a Dartmouth player broke his leg. Princeton won, 13 to 0.

Then the arguments began. The Princeton student newspaper charged that Dartmouth's players had been out to "get" Kazmaier. A Princeton student remembered it decades later: "It was no longer a football game, it was a vendetta." The *Dartmouth Alumni Magazine* replied that everyone who watched the slow-motion films had seen "nothing more than a typical, hard-fought college game."

Two psychologists decided to look: Albert Hastorf of Dartmouth and Hadley Cantril of Princeton, himself a Dartmouth graduate. They showed the same film of the game to students from both schools and asked them to mark every infraction. Princeton students counted, on average, 9.8 infractions by Dartmouth. Dartmouth students, watching the identical reel, counted 4.3\. Their 1954 paper, "They Saw a Game," concluded that the "game" had in fact been many different games.

Two days after the game, a student in the room recalls, Cantril told his 8 a.m. class there had been three games: the one Princeton fans saw, the one Dartmouth fans saw, and "Then there was the real game."

Nobody in this story was dim, and nobody was lying. Each side saw what it needed to be true, and was sure it was simply looking.

## The question

If being clever does not protect us when a fact wears our team's colors, what does? And what happens to a society when more and more of its facts arrive already sorted by team?

## The same table, twice

About six decades later, Dan Kahan of the Cultural Cognition Project at Yale Law School, with Ellen Peters, Erica Cantrell Dawson, and Paul Slovic, built a sharper version of that film.

They showed 1,111 American adults a small table from a fictional study of a new skin cream. Of the patients who used the cream, 223 got better and 75 got worse. Of those who did not, 107 got better and 21 got worse. Did the cream help?

Most people say yes, because 223 is the biggest number. Compare the proportions instead: about a quarter of the cream users got worse, against about one in six of the others. The cream made rashes worse. Overall, 59 percent of subjects got the problem wrong, and people higher in numeracy, the skill and habit of using numbers, did much better.

Then the researchers kept the numbers and changed the story. The rows became cities that had recently banned carrying concealed handguns in public and cities that had not; the columns, crime down or crime up. Same table, same arithmetic.

This time the answers split along political lines, and in this study numeracy widened the gap. The most numerate people tended to get the right answer when it supported their side and miss it when it did not, in both directions: liberal Democrats were sharpest when the data said the ban reduced crime, conservative Republicans when it said the ban increased it. On the authors' estimates, a high-numeracy partisan whose outlook the correct answer affirmed was 45 percentage points more likely to get it right than one whose outlook it threatened. Among low-numeracy partisans, the gap was 25 points.

Now the honest part. That headline result, more skill and more polarization, has not held up well. In 2023, Michael Stagnaro, Ben Tappin, and David Rand ran the task with 2,355 Americans from a probability sample, across five contested issues. People were more accurate when the answer suited their politics, but higher numeracy helped across the board; it did not amplify the bias. A 2024 Western European study on immigration found the same: motivated reasoning, but no extra polarization among the numerate. Stagnaro and colleagues count one independent replication of the amplification effect, numerous failures, and some mixed results.

Where that leaves us: the pull of the team is robust; the claim that intelligence makes it worse is contested; and nothing I found says intelligence shrinks the pull.

## The greatest ignorance

[Part I](https://www.nanosava.com/essays/the-stupidity-no-one-chose/) set out three mechanisms that systems run. This part follows the second, insulation: the distance a system puts between a decision and the inconvenient fact. Bonhoeffer described both of its forms: facts that contradict one's prejudgment "simply need not be believed," and irrefutable ones are "pushed aside as inconsequential."

Plato had a name for the inner version. The greatest *amathia* of the *Laws*, Book III, is not a gap in knowledge but a split between what we judge good and what we love: knowing the good and not wanting it. It can belong, he says, to "expert calculators." Kahan's table puts data behind that line. Everyone in his survey would have agreed that evidence matters. In the moment, some loved their side's answer more than the table, and the love won.

Two more details in Plato matter. The first is why this ignorance is the greatest: it "belongs to the main mass of the soul," in R. G. Bury's translation, the part that feels pleasure and pain, which Plato compares to "the mass of the populace in the State." He finds the same disorder at both scales, counting it among "the most discordant forms of ignorance, whether in the State or the individual." A city could have *amathia* too, which is the nearest thing in Plato to this series' claim about emergence.

The second is in the *Sophist*, where he names the other face of it: "Thinking that one knows a thing when one does not know it. Through this, I believe, all the mistakes of the mind are caused in all of us." Harold Fowler's translation renders *amathia* there as "stupidity." In all of us.

## The three mechanisms, here

**Insulation: what the system provides.** In Kahan's data it looks less like a wall than like stopping early. The numerate subjects, the authors write, were likely to "terminate their engagement with the evidence" when the quick reading gratified their politics. Nobody tore the second row off the table. They simply never got to it. *Amathia* is the inner side of that wall: a mind that could read the second row and does not want to. The outer side is what the system provides: incentives under which, as the next section shows, the fact can arrive and still not matter.

**Delegation: what the system offers.** Most of our beliefs are borrowed, and that is not a vice. Kahan and his colleagues note that we use far more science than we could ever master by being expert at something else: working out who knows what, usually by asking people who share our outlook. That works until positions on a question of fact "come to be seen as symbols of membership in and loyalty to competing groups." Then the default authority shifts, unannounced, from the people who know to the people who belong, and borrowing a belief means borrowing a badge.

**Alignment: what the system rewards.** Identity-protective cognition, the paper says, can be seen as "a psychic self-defense mechanism that steers individuals away from beliefs that could alienate them from others on whose support they depend." The reward is belonging. The cost of disagreeing with your friends lands on you, tonight.

## Why nobody chose it

Here is what makes it emergent. The authors argue that, for each of us, this reasoning is rational. What any one citizen believes about a contested policy fact, they write, is "too inconsequential to influence the risk" that person faces. But holding the belief your group rejects "could expose him or her to an array of highly unpleasant consequences."

So each of us does the sensible thing. We protect our standing, which is near and real, at the cost of accuracy, which is far and shared. Add that up across millions and you get what the authors call a "tragedy of the science communications commons": each person's interest in beliefs that bind them to their people works against everyone's interest in policies that actually work. They describe otherwise intelligent people caught in "the unenviable situation of having to choose between crediting the best available evidence or simply being who they are."

No one designed this, and it is not confined to politics. It shows up wherever being right has become a sign of belonging: a medical specialty and its signature procedure, a leadership team and the strategy it announced. A football game can survive two versions of itself. A public health emergency, a bridge, or a bank cannot. Reality sends one bill, and it does not check which side you sat on.

## More and more

The cost today is mostly invisible: slower decisions, studies that persuade only those who already agreed, experts spending their credibility proving they have no team.

The risk is in the ladder. On the first rung, a few questions carry team colors, and the rest can still be argued on evidence. On the second, the set grows: a food, a medicine, a technology, a statistic becomes a badge, one less fact we can examine together. On the third, the instruments of correction get sorted too. The statistician, the scientist, the reporter, and the court are each assigned to a side, so even those whose job is to tell us we are wrong can be waved away by label.

That third rung is insulation at the scale of a society, and it is the one that matters. A society can be wrong about a great deal and still recover, as long as it shares a way of noticing. When every fact arrives pre-sorted by team, it loses that shared ability, and keeps being wrong in two directions at once.

I do not know how fast we are climbing; these studies are snapshots, not trends. But the systems we now live inside make sorting cheaper every year, and nothing in this research suggests that getting smarter will lift us off the ladder on its own. Kahan and his colleagues agree: better science education, "while immensely important," cannot be expected to end these conflicts. The fix lies in the commons, in keeping facts from acquiring a team at all.

## Noticing it

The Three Checks and the Scale Question, for a belief you hold dearly, ideally one your friends share.

1. **The delegation check.** Where did I get this? Could I state the evidence for it myself, or do I hold it because people like me hold it?
2. **The insulation check.** What would I prefer not to find out about this, and what keeps it from reaching me? When did I last read the strongest case against it, in words its holders would sign?
3. **The alignment check.** Would I still believe this if my side did not? What would it cost me to say out loud that I had changed my mind, and to whom?
4. **The Scale Question.** If everyone treated evidence the way I am treating this piece of it, could we ever find out together that we were wrong?

One practice comes straight from the experiment: **the swap test.** When a finding pleases you, relabel it in your head so that it supports the other side, and read it again. If you would now go hunting for the flaw, go hunting for it now. The table does not know whose side it is on.

## Leaving a way out

The checks are for me. For others, what helps most is not a better argument but an exit.

At stake is what the sociologist Erving Goffman called face: "the positive social value a person effectively claims for himself." We defend our own, he observed, and protect each other's. Once a belief is a badge, conceding a fact costs face, a price few pay in public, to an opponent, on a deadline.

Does a secure self lower the price? In 2000, Geoffrey Cohen, Joshua Aronson, and Claude Steele gave students with strong views on the death penalty a persuasive report against those views. Those first affirmed by writing about a value they cared about judged the report more favorably (72 students), though their attitudes moved no more than the others'. After praise for a skill (80 students), 62 percent moved substantially toward it, against 40 percent of the rest.

The samples were small; later political tests found less. In three experiments, Brendan Nyhan and Jason Reifler saw self-affirmation "not consistently increase willingness to accept corrective information"; in 2022, several research teams found its "treatments have little effect" on political attitudes and beliefs. The lab tools are weak; the problem is not. Alignment puts a price on changing one's mind, paid in face, and we can lower it for each other.

What an exit looks like:

- **Ask; do not rule.** A conclusion people reach themselves costs less face. "What would we expect to see if this were true?" puts you both in front of the same table; "You are wrong" puts one of you on trial.
- **Separate the fact from the flag.** They can keep their values, their people, and their dignity and still update this one fact.
- **Go first.** Name something you were wrong about.
- **Let it happen later.** Few people change their minds mid-argument; give them time.
- **Do not collect the win.** "I told you so" makes them pay in face after all, and teaches others to stay put.

## The real game

Cantril told his students there were three games that Saturday: Princeton's, Dartmouth's, and the real one. The real game is always there. It is just the one nobody in the stands is watching. Seeing it does not take a better brain. It takes, for a moment, loving the truth a little more than the team, and making that a little cheaper for the people around us.

**This week's ask: a sighting.** Where did you see a fact get sorted by team before anyone looked at it, in your own field? No names, no shaming. Send it as a comment below (free subscribers can comment) or to hello@nanosava.com. I read every sighting, and the best ones, anonymous unless you say otherwise, will shape later parts of [the series](https://www.nanosava.com/series/emergent-stupidity/).

## Sources

1. Albert H. Hastorf and Hadley Cantril, "They Saw a Game: A Case Study," *Journal of Abnormal and Social Psychology* 49(1), 1954, pp. 129 to 134\. [https://doi.org/10.1037/h0057880](https://doi.org/10.1037/h0057880?ref=nanosava.com) (paywalled; the 9.8 and 4.3 figures, from the paper's Table 2, with 49 Princeton and 48 Dartmouth students, and the "many different games" conclusion were checked against a scan of the article.)
2. Brett Pelham and Curtis Hardin, "People See Things as They Expect Them to Be, But that Doesn't Have to Spell Political Disaster," *Character & Context*, Society for Personality and Social Psychology, 2019\. [https://spsp.org/news-center/character-context-blog/people-see-things-they-expect-them-be-doesnt-have-spell](https://spsp.org/news-center/character-context-blog/people-see-things-they-expect-them-be-doesnt-have-spell?ref=nanosava.com)
3. C.E.W., "Concerning the Princeton Game," *Dartmouth Alumni Magazine*, January 1952\. [https://archive.dartmouthalumnimagazine.com/article/1952/1/1/concerning-the-princeton-game](https://archive.dartmouthalumnimagazine.com/article/1952/1/1/concerning-the-princeton-game?ref=nanosava.com)
4. Ken Ackerman, "Unforgettable game and lecture," letter, *Princeton Alumni Weekly*. [https://paw.princeton.edu/inbox/unforgettable-game-and-lecture](https://paw.princeton.edu/inbox/unforgettable-game-and-lecture?ref=nanosava.com)
5. "Albert H. Hastorf *49," memorial,* Princeton Alumni Weekly\*. [https://paw.princeton.edu/memorial/albert-h-hastorf-49](https://paw.princeton.edu/memorial/albert-h-hastorf-49?ref=nanosava.com)
6. Dan M. Kahan, Ellen Peters, Erica Cantrell Dawson, and Paul Slovic, "Motivated numeracy and enlightened self-government," *Behavioural Public Policy* 1(1), 2017, pp. 54 to 86\. [https://www.cambridge.org/core/journals/behavioural-public-policy/article/abs/motivated-numeracy-and-enlightened-selfgovernment/EC9F2410D5562EF10B7A5E2539063806](https://www.cambridge.org/core/journals/behavioural-public-policy/article/abs/motivated-numeracy-and-enlightened-selfgovernment/EC9F2410D5562EF10B7A5E2539063806?ref=nanosava.com) (abstract). First circulated as a Cultural Cognition Project working paper, 2013\. Figures and quotations here are from the preprint at [https://hdl.handle.net/1794/22105](https://hdl.handle.net/1794/22105?ref=nanosava.com); the published wording may differ.
7. Michael N. Stagnaro, Ben M. Tappin, and David G. Rand, "No association between numerical ability and politically motivated reasoning in a large US probability sample," *PNAS* 120(32), 2023, e2301491120\. [https://pmc.ncbi.nlm.nih.gov/articles/PMC10410746/](https://pmc.ncbi.nlm.nih.gov/articles/PMC10410746/?ref=nanosava.com)
8. Paul Connor, Emily Sullivan, Mark Alfano, and Nava Tintarev, "Motivated numeracy and active reasoning in a Western European sample," *Behavioural Public Policy* 8(1), 2024, pp. 24 to 46\. [https://ideas.repec.org/a/cup/bpubpo/v8y2024i1p24-46\_2.html](https://ideas.repec.org/a/cup/bpubpo/v8y2024i1p24-46%5F2.html?ref=nanosava.com)
9. Cultural Cognition Project, "About us," Yale Law School. [https://www.culturalcognition.net/about-us/](https://www.culturalcognition.net/about-us/?ref=nanosava.com)
10. Plato, *Laws*, Book III, 689a to e, trans. R. G. Bury, *Plato in Twelve Volumes*, vols. 10 and 11, Harvard University Press, 1967 and 1968 (Perseus Digital Library). [http://www.perseus.tufts.edu/hopper/text?doc=Perseus:text:1999.01.0166:book=3:page=689](http://www.perseus.tufts.edu/hopper/text?doc=Perseus:text:1999.01.0166:book=3:page=689&ref=nanosava.com)
11. Plato, *Sophist*, 229c, trans. Harold N. Fowler, *Plato in Twelve Volumes*, vol. 12, Harvard University Press, 1921 (Perseus Digital Library). [http://www.perseus.tufts.edu/hopper/text?doc=Perseus:text:1999.01.0172:text=Soph.:section=229c](http://www.perseus.tufts.edu/hopper/text?doc=Perseus:text:1999.01.0172:text=Soph.:section=229c&ref=nanosava.com)
12. Dietrich Bonhoeffer, "After Ten Years," prologue to *Letters and Papers from Prison*, Dietrich Bonhoeffer Works (English), vol. 8, ed. John W. de Gruchy, trans. Barbara and Martin Rumscheidt, Fortress Press, 2010, section "On Stupidity," p. 43\. Quotations checked against the publisher's sample of the prologue: [https://ms.fortresspress.com/downloads/0800697030Prologue.pdf](https://ms.fortresspress.com/downloads/0800697030Prologue.pdf?ref=nanosava.com)
13. Erving Goffman, "On Face-Work: An Analysis of Ritual Elements in Social Interaction," *Psychiatry* 18(3), 1955, pp. 213 to 231\. [https://doi.org/10.1080/00332747.1955.11023008](https://doi.org/10.1080/00332747.1955.11023008?ref=nanosava.com) Quoted from the reprint in Goffman, *Interaction Ritual: Essays on Face-to-Face Behavior*, Anchor Books, 1967 (Pantheon reprint, 1982): the definition of face on p. 5; the "defensive" and "protective" orientations, paraphrased here, on p. 14.
14. Geoffrey L. Cohen, Joshua Aronson, and Claude M. Steele, "When Beliefs Yield to Evidence: Reducing Biased Evaluation by Affirming the Self," *Personality and Social Psychology Bulletin* 26(9), 2000, pp. 1151 to 1164\. [https://doi.org/10.1177/01461672002611011](https://doi.org/10.1177/01461672002611011?ref=nanosava.com) Read in the copy once hosted by the Stanford Graduate School of Education: <https://web.archive.org/web/20150620131731/https://ed.stanford.edu/sites/default/files/when%5Fbeliefs%5Fyield.pdf> (Study 1: 72 participants analyzed; Study 2: 80, of whom 61.9 percent of the affirmed and 39.5 percent of the unaffirmed shifted two or more points on a seven-point scale.)
15. Brendan Nyhan and Jason Reifler, "The roles of information deficits and identity threat in the prevalence of misperceptions," *Journal of Elections, Public Opinion and Parties* 29(2), 2019, pp. 222 to 244\. [https://doi.org/10.1080/17457289.2018.1465061](https://doi.org/10.1080/17457289.2018.1465061?ref=nanosava.com) (quoted from the abstract; authors' manuscript, December 21, 2017: [https://sites.dartmouth.edu/nyhan/files/2021/03/opening-political-mind.pdf](https://sites.dartmouth.edu/nyhan/files/2021/03/opening-political-mind.pdf?ref=nanosava.com))
16. Benjamin A. Lyons, Christina E. Farhart, Michael P. Hall, John Kotcher, Matthew Levendusky, Joanne M. Miller, Brendan Nyhan, Kaitlin T. Raimi, Jason Reifler, Kyle L. Saunders, Rasmus Skytte, and Xiaoquan Zhao, "Self-Affirmation and Identity-Driven Political Behavior," *Journal of Experimental Political Science* 9(2), 2022, pp. 225 to 240 (online February 2021). [https://doi.org/10.1017/XPS.2020.46](https://doi.org/10.1017/XPS.2020.46?ref=nanosava.com) (quoted from the abstract, via [https://ideas.repec.org/a/cup/jexpos/v9y2022i2p225-240\_6.html](https://ideas.repec.org/a/cup/jexpos/v9y2022i2p225-240%5F6.html?ref=nanosava.com))