The Wisdom Audit, part 2 of 9. This series tests famous sayings against published research and ends each test with a verdict.
We inherit a lot of advice about which signals to trust.
Don’t judge a book by its cover.
A picture is worth a thousand words.
Actions speak louder than words.
Honesty is the best policy.
These four sayings make four competing claims about where truth lives: in appearances, in images, in behavior, or in words. They can’t all be equally right, so I ranked them. Here’s what the research actually says about the predictive power of each signal, and where the sayings hold up or fall apart.
Don’t judge a book by its cover
Origin: The exact wording shows up in American print by 1849 and was already being called “an old proverb” by 1854.
What it actually claims: snap judgments from appearance are unreliable, so you shouldn’t use them.
The data says people are better at this than the proverb thinks. In 1992, Nalini Ambady and Robert Rosenthal pooled every study they could find where people made judgments from “thin slices” of behavior, meaning less than five minutes of observation, and found those judgments predicted real outcomes at an average correlation of .39. (Correlations run from 0, no connection, to 1, perfect prediction, so .39 is a solid signal by the standards of psychology.) In their famous follow-up, students watched three ten-second silent clips of a teacher, thirty seconds total with no audio, and their ratings correlated .76 with that teacher’s end-of-semester evaluations from actual students. Alexander Todorov’s lab pushed it further: 100 milliseconds of looking at a face is enough to form trait judgments that barely change with unlimited time, and judgments of competence from candidates’ faces alone predicted 68.8% of the 2004 US Senate races.
The first half of the literature, then, says covers carry real signal. The problem is the second half: the halo effect. Dion, Berscheid, and Walster showed in 1972 that we assume attractive people have better character and better futures, and Hamermesh and Biddle found plain-looking workers earn roughly 5 to 10 percent less than average-looking ones, with a smaller premium for the attractive. That part is not signal. That part is bias with a paycheck.
Verdict: Split. First impressions are far more predictive than the saying admits, and they are simultaneously contaminated in ways you can’t turn off.
The rewrite: Judge the book by its cover, but know which parts of the cover are ink and which are glare.
A picture is worth a thousand words
Origin: This one did not come from ancient China. An advertising executive named Fred Barnard put “one look is worth a thousand words” in the trade journal Printers’ Ink in 1921, inflated it to “one picture is worth ten thousand words” in 1927, and attached a fake Chinese proverb label to make it sound wise. Newspaper editor Arthur Brisbane had said a version in 1911. This saying is ad copy that escaped the lab.
What it actually claims: images communicate and persist better than text.
The data backs the ad man. In Lionel Standing’s 1973 experiments, people viewed 10,000 pictures once and recognized about 83 percent of them afterward. Picture memory scales in a way word memory doesn’t. But the effect has a dark side: Newman and colleagues showed in 2012 that attaching a photo that proves nothing (a photo of the Mona Lisa next to the claim “the Mona Lisa has no eyebrows”) makes people more likely to judge the claim true. They call it truthiness. A picture is worth a thousand words even when the thousand words are false.
Verdict: Holds, and it holds uncomfortably well.
The rewrite: A picture is worth a thousand words, which is exactly why you should ask who picked the picture.
Actions speak louder than words
Origin: John Pym told the English Parliament in 1628 that “actions are more precious than words.” The near-modern wording appears in Thomas Manton’s sermons by 1693, and the exact phrase was in American print by 1736.
What it actually claims: behavior is a more reliable signal than speech.
Here the folk wisdom is right and the famous statistic behind it is wrong. You’ve probably heard that communication is 7 percent words, 38 percent tone, 55 percent body language. That “rule” comes from two 1967 studies by Albert Mehrabian that tested single spoken words, only about feelings, with inconsistent cues, on all-female samples. Mehrabian himself has said publicly that the ratio does not apply to communication in general. The number is a misquote that outcompeted the truth, which is its own lesson about how sayings spread.
What does hold is the principle underneath: revealed preference beats stated preference. Across the behavioral literature, what people do predicts what they’ll do next far better than what they say they’ll do. The diary studies on lying make the point from the other side. DePaulo’s participants averaged one to two lies per day, and Serota’s survey of 1,000 Americans found the distribution wildly skewed, with 60 percent reporting zero lies in the past day and about half of all lies coming from 5 percent of people. Words are cheap for a specific minority, and that minority is enough to devalue the currency.
Verdict: Split. The saying survives; its most famous supporting evidence does not.
The rewrite: Actions speak louder than words, and no, it’s not 55 percent.
Honesty is the best policy
Origin: Sir Edwin Sandys, an English statesman, wrote it in 1599, and Benjamin Franklin popularized it in America.
What it actually claims: truthfulness produces better outcomes for the truth-teller than deception.
The data mostly agrees, for a reason the proverb never gives. In 2018, Emma Levine and Taya Cohen assigned people to be completely honest in every conversation for three days. Participants predicted it would be painful. It wasn’t. Honest conversations turned out to be more enjoyable and less socially damaging than people expected, in both directions of the conversation. The barrier to honesty isn’t that honesty goes badly. The barrier is that we systematically overestimate how badly it will go.
Verdict: Mostly holds, not because lying never pays, but because we misprice honesty.
The rewrite: Honesty is the best policy you keep overestimating the cost of.
The ranking
| Signal | How well it predicts | Where it goes wrong |
|---|---|---|
| Behavior (actions) | This is the strongest signal of the four: what people do predicts what they will do next better than anything they say | The famous 7-38-55 statistic attached to this saying is a misquote |
| Images (pictures) | Pictures are remembered and believed at rates text never reaches | They persuade just as effectively when the claim is false |
| Appearance (covers) | Thin-slice judgments predict real outcomes far better than chance (r = .39) | The halo effect and the beauty premium contaminate the reading |
| Words (honesty) | Honest words are undervalued because we overestimate their cost | A small group of prolific liars devalues everyone else’s words |
One more thing is worth noticing before this series goes further: of these four sayings, the one with the fakest pedigree (a 1921 ad slogan wearing a Chinese proverb costume) is the one the data supports most cleanly. Wisdom doesn’t care who said it. That cuts both ways, and it’s where this series is headed.
Next in the series: If It Ain’t Broke, Check Anyway.
Sources: Ambady & Rosenthal 1992 (Psychological Bulletin 111); Ambady & Rosenthal 1993 (JPSP 64); Willis & Todorov 2006 (Psychological Science 17); Todorov et al. 2005 (Science 308); Dion, Berscheid & Walster 1972 (JPSP 24); Hamermesh & Biddle 1994 (AER); Standing 1973 (QJEP 25); Newman et al. 2012 (Psychonomic Bulletin & Review); Mehrabian & Wiener 1967, Mehrabian & Ferris 1967; DePaulo et al. 1996 (JPSP 70); Serota, Levine & Boster 2010 (Human Communication Research 36); Levine & Cohen 2018 (JEP: General 147).
Laurie Stach 