Write Down the Fact That Disagrees With You First

Write down the fact that disagrees with you the moment you meet it. Charles Darwin kept that rule for many years, and a 2009 meta-analysis of 91 studies explains why it matters: people choose agreeable information over disagreeable information by a moderate, consistent margin. The bias sits in what you pick up, which is exactly where a note intervenes.¹²

Darwin described the habit in the autobiography he wrote in 1876, which his son Francis published in 1887. He listed it among the reasons On the Origin of Species succeeded:¹

"I had, also, during many years followed a golden rule, namely, that whenever a published fact, a new observation or thought came across me, which was opposed to my general results, to make a memorandum of it without fail and at once; for I had found by experience that such facts and thoughts were far more apt to escape from the memory than favourable ones. Owing to this habit, very few objections were raised against my views which I had not at least noticed and attempted to answer."

Most retellings — including the Farnam Street essay where many readers first met the rule — smooth the wording slightly. The version above is Darwin's own, from the Project Gutenberg text of the 1887 edition.

It was not a mood. Darwin kept "from thirty to forty large portfolios, in cabinets with labelled shelves," into which he could "at once put a detached reference or memorandum."¹ The objection had somewhere to go the day it arrived.

The surprise sits in the second half of his first sentence. Darwin gave a reason: disagreeable facts "were far more apt to escape from the memory than favourable ones." The memory research mostly does not back that reason. The habit survives anyway, because the bias it corrects is real — it just lives one step earlier than Darwin thought.

The study, briefly: 91 studies, one consistent lean

The clearest measure of the bias Darwin was guarding against is a 2009 meta-analysis in Psychological Bulletin by William Hart, Dolores Albarracín, Alice Eagly and colleagues. It pooled 91 studies with just under 8,000 participants, each offering people a choice between information that supported their view and information that challenged it. People picked the supporting side, moderately and reliably.

The search turned up 67 eligible reports containing 91 studies and 300 statistically independent groups, with "just under 8,000 participants."² The authors' summary is one line: "Analyses indicated a moderate preference for congenial over uncongenial information (d=0.36)."³ In the paper's terms, congenial information supports what you already believe. Uncongenial information challenges it.

The number is not noise — the fixed-effects estimate was d = 0.36, with a 95% confidence interval of 0.34 to 0.39. The random-effects estimate was 0.38. The authors call both "a moderate congeniality bias."⁴

What d = 0.36 does not mean matters as much. It is a standardized effect size, not a percentage — nobody in these studies ignored "36%" of anything. It means that across different topics, populations and labs, the average participant leaned toward the agreeable side by a margin researchers keep detecting.

What they actually measured — and where the lean is steepest

Hart's team measured choice, not memory: which items a participant chose to read when both sides were on offer. The lean was steeper on political issues and on issues with personal stakes. And — the counterintuitive part — it was steeper when the disagreeable material was good rather than weak.

Three moderators carry the practical weight:

• Political issues. d = 0.46, stronger than for other issues.⁴
• Personal stakes. d = 0.48 when the issue was high in outcome relevance for the participant, against 0.33 when it was low — in the random-effects analysis only; the fixed-effects analysis found no difference.⁴
• Quality. In the paper's words, "the congeniality bias was larger when the uncongenial information was high or moderate in quality rather than low in quality."⁴

The last line is the one to sit with. People do not mainly avoid weak objections — those are easy to read and dismiss. They avoid the strong ones. In a notes file, the objection you skipped is likely the one that would have changed the note.

The bias also bends. The abstract reports it was weaker when the available information was low in quality, when people held their view with high confidence, and when closed-mindedness was low. In one condition it reversed outright: "an uncongeniality bias emerged when uncongenial information was relevant to accomplishing a current goal."³ When the disagreeable fact served a task in front of people, they went looking for it.

The part Darwin got wrong: we don't especially forget it

Darwin's reason — that disagreeable facts slip from memory faster — has weak support. A 1999 meta-analysis led by Alice Eagly found the memory advantage for agreeable information "inconsistent" and "small when these effects are aggregated." A 2000 follow-up found people remembered both sides equally well, even two weeks later.

The 1999 paper adds a detail that should make any reader of older studies cautious: "The tendency of the congeniality effect to decrease over the years spanned by this literature appeared to reflect the weaker methods used in the earlier studies."⁵ The better the method, the smaller the effect.

Eagly's group then tested the question directly, in three experiments on abortion and on gays in the military, with recall and recognition measured soon after the message or two weeks later. The result: "Participants remembered congenial and uncongenial messages equally well, despite their disapproval of the uncongenial information."⁶

They also found a reason. "Process findings suggested that participants' thoughtful counterarguing of the uncongenial messages enhanced their memory for them."⁶ Arguing with a fact is one way of keeping it. Readers of how to remember what you read will recognize the pattern: what you work on is what stays.⁷

Darwin half-knew. Near the end of the autobiography he calls his own memory "extensive, yet hazy." It was good enough to warn him "that I have observed or read something opposed to the conclusion which I am drawing," and, after a time, to recall "where to search for my authority."¹ Hazy is the honest word — the portfolios, not his memory, held the specifics.

Why the note is the right place for the fix

If the bias lives in selection, the fix belongs at the moment of selection — and for most note-keepers that moment is capture. A note is a decision about what is worth keeping. Darwin's rule fits because it overrides that decision for one class of fact, at once, before preference gets a vote.

Francis Bacon described the weighting problem long before anyone ran an experiment. In Joseph Devey's English edition of Novum Organum (1620), the human understanding, once committed to a proposition, "forces everything else to add fresh support and confirmation," and "in establishing any true axiom the negative instance is the most powerful."⁸

Raymond Nickerson's 1998 review gives the modern definition. Confirmation bias "connotes the seeking or interpreting of evidence in ways that are partial to existing beliefs, expectations, or a hypothesis in hand."⁹ Seeking and interpreting — two moments, and a note touches both.

The experimental fix for the interpreting half has held up across very different settings. Charles Lord, Mark Lepper and Elizabeth Preston — whose 1984 paper credits Bacon in its first sentence — asked people to consider the opposite. "In both experiments the induction of a consider-the-opposite strategy had greater corrective effect than more demand-laden alternative instructions to be as fair and unbiased as possible."¹⁰

Telling yourself to be fair did less than giving yourself a specific counter-task. Listing reasons against worked for experts too: Thomas Mussweiler, Fritz Strack and Tim Pfeiffer found, in a real-world setting with expert participants, that "listing arguments that speak against a provided anchor value reduces the effect."¹¹ Matthew Fisher and Frank Keil found that people overrate how well they can justify positions they care about, and that "explicitly considering counterarguments removes the effect of care" on that self-assessment.¹²

An ## Against heading is a standing consider-the-opposite instruction, written into the note before you need it. It also echoes Hart's goal-relevance finding: a slot to fill is a small, standing task. Whether it produces the same reversal has not been tested. Writing is where a claim gets checked, as writing is the best thinking tool argues, and the check is only as good as what reached the page.¹³

The same holds for anything downstream of your notes. An AI that summarizes them can only summarize what is in them, and it will sound just as confident about a one-sided file as about a balanced one.¹⁴

How to apply this tomorrow

The habit fits in four lines of plain Markdown and one review rule. Give every note that makes a claim an ## Against section, write the objection the moment you meet it, link the note that disagrees, and record "none found yet" with a date rather than leaving the section blank.

A worked template, for any note that asserts something:

---
title: Weekly planning beats daily standups for our team
contra: [[daily-standups-caught-blockers-earlier]]
---

## Claim
Our team ships more finished work with one weekly planning
session than with a daily fifteen-minute standup.

## For
- Sprint log, March to May: (one line on what it shows)

## Against
- none found yet (2026-10-08)

## What would change my mind
- Two months where weekly planning missed a blocker that a
  standup would have caught within a day.

Five actions, in order of leverage:

1. Put the section in the template, not the note. A heading that already exists asks to be filled. One you must remember to add rarely appears.
2. Write it at once. Darwin's phrase was "without fail and at once." A line and a source are enough; the arguing can come later.
3. Link the disagreeing note. A contra: line keeps the disagreement visible when you open either note, months apart.
4. Date the empty state. "none found yet (2026-10-08)" is a claim about a search you made. A blank section claims nothing.
5. Review the gaps. Once a month, search your notes for none found yet and look for one real objection to each.

One real objection beats a mandatory list. Lawrence Sanna, Norbert Schwarz and Shevaun Stocker asked people to list either two or ten thoughts about how an event might have turned out otherwise. "Listing many counterfactual thoughts was experienced as difficult and consistently increased the hindsight bias"; listing two did not.¹⁵ That is hindsight bias, a different bias from the one Hart measured — a design caution for the template, not evidence that the habit backfires.

The caveats

No study here tested note-taking. The habit is an application of the selection findings, not a measured intervention: Hart measured what people chose to read, Lord and Mussweiler measured judgments after a counter-instruction, and none of these studies compared notebooks with and without an Against section.

The effect is moderate and moderated. d = 0.36 is an average across very different settings. It shrinks for low-quality information and for confident holders of a view, and it reverses when the disagreeable information serves a current goal.

The memory effect Darwin described is small and inconsistent, not zero. Eagly's 1999 data show conditions under which it was stronger, and her 2000 experiments covered two contested social issues, not every kind of fact.

Bacon is quoted from Devey's edition; other translations of Aphorism XLVI word it differently. Darwin is quoted from the Project Gutenberg text of the 1887 edition. No page number is given, because editions paginate differently.

Frequently Asked Questions

Do you incorporate criticism in your notes?

You should, and in the note itself rather than a separate file. Give every note that makes a claim an ## Against section and write the objection the moment you meet it. Across 91 studies, people chose agreeable over disagreeable information by a moderate margin, so criticism that is not captured at once is often never captured.

How do I avoid confirmation bias in note-taking?

Build the counter-evidence slot into your template, not your intentions. In Lord, Lepper and Preston's 1984 experiments, a specific instruction to consider the opposite corrected judgment better than an instruction to be fair and unbiased. A heading that asks "what argues against this?" is that specific instruction, applied every time you write a note.

Should I write down counterarguments in my notes?

Yes, and write them down when you meet them. Fisher and Keil found that people overrate how well they can justify positions they care about, and that explicitly considering counterarguments removed the extra overconfidence that caring produced. Eagly's 2000 experiments also found that counterarguing a disagreeable message improved memory for it. One real objection per note is enough.

What is Darwin's golden rule?

Darwin's golden rule, described in his autobiography (written 1876, published 1887), was to make a memorandum "without fail and at once" of any published fact, observation or thought opposed to his general results. He credited the habit with the fact that very few objections were raised against his views that he had not already noticed and tried to answer.

Do we really forget facts that disagree with us?

Not much more than facts that agree with us. Eagly's 1999 meta-analysis found that memory advantage inconsistent and small, and her 2000 experiments found people remembered agreeable and disagreeable messages equally well, immediately and two weeks later. The consistent bias is in selection — what people choose to read in the first place.

How many counterarguments should a note have?

One real one is a good default. Sanna, Schwarz and Stocker found that listing ten alternative thoughts felt difficult and increased hindsight bias, while listing two did not. That study concerned a different bias, but it is a sound design caution: a quota invites filler, and filler makes the section look answered when it is not.

What is a contra note in a Zettelkasten?

A contra note is a note that disagrees with another note, linked to it directly, for example with a contra: line in the frontmatter. The link keeps the disagreement visible from both sides. Paired with an ## Against section and a dated "none found yet" placeholder, it turns counter-evidence into something a search can find.

Darwin was right about what to write down and wrong about why. The fact that disagrees with you will not slip away on its own; you will simply never pick it up — unless the note asks for it first.

mnmnote.com keeps notes as plain Markdown on your own device, so an ## Against heading is just a line you type.