The weirdest people in the world?

Behavioral and Brain Sciences 33(2–3):61–83 (target article), with open peer commentary and authors' response (pp. 83–135), doi 10.1017/S0140525X0999152X. Author-hosted PDF of the full 75-page BBS treatment from Henrich's former UBC page. Read by researcher R4e (innate against learned) for research batch R4, 5 October 2026. Provenance: papers/base/henrich2010_weird.provenance.json.

What was read

• Read in full: the target article, every line (pdftotext lines 1–2405: abstract, sections 1–8, acknowledgements, notes 1–16); response sections R4 ("Are WEIRD populations actually unusual?") and R5 ("Does variability conceal psychological universals?"), lines 6703–6968; the combined reference list, lines 7146–8975.
• Not read (the brief allowed skipping commentaries): the 28 commentaries (lines 2406–5754); the rest of the authors' response: its introduction and R1–R3 (lines 5755–6702) and R6–R8 (lines 6969–7145). Figures exist only as captions in the text.
• Caution: the text extraction comes from a two-column PDF; a few passages are out of order on the page, but no content is missing.

The argument

• Who psychology studies. In the top journals of six subfields (2003–2007), 68% of subjects were from the US and 96% from Western industrialised countries, which hold 12% of the world's population. In JPSP, 67% of US samples (80% of others) were psychology undergraduates. A random American undergraduate is more than 4,000 times more likely to be a research subject than a random person from outside the West. 70% of psychology citations come from the US, against 37% in chemistry.
• The claim: WEIRD subjects (Western, Educated, Industrialised, Rich, Democratic) are not just a narrow sample but frequent outliers, "among the least representative populations one could find for generalizing about humans".
• What they do not claim: that humans share little. They "expect humans from all societies to share, and probably share substantially, basic aspects of cognition, motivation, and behavior". Their point is that, with this database, we cannot tell reliably developing aspects of psychology from developmentally, culturally or environmentally contingent ones.
• Scope: they exclude domains expected to vary between individuals (personality traits, values, emotional expressiveness), psychopathology, and societal-level universals (cooking, kinship, grammar). They take no position on whether differences are cultural, developmental or genetic.
• Method: four "telescoping" contrasts, each against the available comparative data.

Contrast 1: industrialised against small-scale societies

Similarities across the divide:

• some illusions (Perspective Drawing);
• colour perception, with terms affecting it "at the margins";
• recognition of basic facial emotions: 58% of variance universal against 9% culturally specific (Elfenbein and Ambady 2002); pride displays;
• passing the explicit false-belief task (in all populations studied, but at ages from 4 to at least 9, industrialised populations earliest);
• analog number sense (Amazonian groups, infants, primates);
• confusing people within the same relationship category, though the categories differ;
• psychological essentialism, though its degree may vary.

Contrast 2: Western against non-Western societies

• Antisocial punishment: in about half of the undergraduate samples from a diverse set of industrialised countries (Herrmann 2008; the article gives no count), many subjects paid to punish people who contributed more than they did, wiping out the cooperation-raising effect of punishment. Western samples behaved like the original Zurich students. The behaviour correlates with formal institutions, civic norms and GDP.
• Self-concept: Westerners describe themselves by internal traits and attitudes; non-Western groups by roles and relationships.
  • Self-enhancement is much stronger in the West and null or reversed in East Asia. Response R5 gives the numbers: self-serving attributional bias d = 1.05 for Americans against −0.30 for Japanese; across 31 methods, d = 0.87 for Westerners against −0.01 for East Asians, with only the self-esteem IAT showing no difference.
  • Westerners value choice more, and conformity is weaker in the West (meta-analysis of 17 societies, Bond and Smith 1996).
• Analytic against holistic thought. Both systems exist in all normal adults, but Westerners attend more to objects than to context, explain behaviour by dispositions (the "fundamental attribution error" is "less fundamental elsewhere"), and classify by rules rather than resemblance. They sit at the extreme, and some effects reverse elsewhere.
• Moral reasoning. Kohlberg's post-conventional reasoning was found in all Western urban samples and in no small-scale society (Snarey 1985, 27 countries). Outside the West, and among Western religious conservatives, morality also rests on community and divinity, not only justice and harm.
• Similarities:
  • mate preferences in 37 cultures: men weight physical attractiveness more in all 37, women weight ambition more in 34. But the preferred waist-to-hip ratio of about 0.7 failed to replicate in small-scale societies (note 10).
  • Five-factor personality structure in most of 51 countries. It was weaker where data quality was poor, and weaker when derived from indigenous trait words (notes 11–12).
  • Willingness to punish free-riders did not differ once demographics were controlled.

Contrasts 3 and 4: Americans, and American students

• Americans are the most individualistic Westerners; in a survey of 100 societies, only American parents gave babies a separate room. Their worldview defence after death reminders is r = 0.37, against 0.30 in other Western and 0.26 in non-Western samples.
• Students against other adults: across more than 7,000 studies (N > 650,000; Peterson 2001), students and non-students differed in direction or size for about half the phenomena. Students are more susceptible to attitude change and social influence. Missouri adults gave 47% in the Dictator Game, undergraduate freshmen 32%. Game behaviour stabilises around 24, trust around 30.
• Heritability depends on the environment (Turkheimer 2003, cited). For high-SES 7-year-olds, genes explain 70–80% of IQ variance and shared environment under 10%; for low-SES children, genes 0–10% and shared environment about 60%. The authors: "much of what we think we have learned from behavioral genetics may be misleading, as the data are disproportionately influenced by WEIRD people."
• The same allele, opposite effects: a serotonin-receptor variant goes with more attention to focal objects in Americans and less in Koreans (Kim et al., in press).
• WEIRD psychology is recent. IQ rose 18 points in half a century across industrialised nations, mostly on analytic subtests (Flynn). Indian students' attitudes to tradition resemble American grandparents' more than American students'.

General conclusions

• Variation is commonplace in size, existence (an effect present or absent) and direction. It includes domains assumed "basic": visual perception, memory, attention, fairness, categorisation, induction, spatial cognition, self-enhancement, moral reasoning and the heritability of IQ.
• Similarities can arise three ways: reliably developing adaptations (theory of mind), by-products of innate adaptations, or independent invention and diffusion of useful learned responses (counting, cooking, fire).
• No reliable a priori test of universality yet.
  • The intuition that attention, memory and perception vary little is not supported by their data.
  • Physiological or genetic measures do not settle it either: the same gene or the same task can work differently across populations.
  • Presence in other species is weakened by culture–gene coevolution.
  • Infant phenomena are "reasonably assumed to be more universal", but later development can reverse early biases (the spatial case).
• Children and primates do not replace cross-cultural comparison. Differences appear early (folk biology) and can be larger in children than in adults (Müller-Lyer).
• The mechanism they propose: evolution gave humans ontogenetic programs, including cultural learning, that tune brains to local physical and social environments. Cultural evolution has changed physical environments, cognitive environments (counting systems, colour terms, writing, grammatical structures, categories, heuristics) and social environments (norms, institutions). "Broad patterns of psychology may be – in part – a product of our genetic program's common response to culturally constructed environments."
• WEIRD samples remain fine for existence proofs. They recommend explicit statements of generalisability, reporting of samples, data sharing and diverse subject pools.

From the response (R4, R5)

• R4: the first society rich and specialised enough to amass psychological data may be unusual for the same reasons (individualism, markets, schooling). Instruments honed on WEIRD samples may also make WEIRD people look unusual; whether they stay unusual once phenomena important elsewhere are counted is open.
• R5: universality must be inferred from phenotypes, so narrow samples can mislead about what is universal underneath. Personality structure and some sex differences in mate preference are solid functional universals; fairness, moral principles and spatial perception vary widely. "Universal ontogenetic processes that give rise to different psychological processes under different conditions" are the alternative to universal psychological processes. Epigenetics and culture–gene coevolution (lactose tolerance) make population differences possible without a single universal process.

Limits

• A review assembled to make a case. The authors invite counter-examples, and the 28 commentaries (unread here) contest parts of it.
• Much evidence comes from two-population comparisons. Several cited studies were in press or under review in 2010.
• No numbers on how much of total psychological variance lies between cultures as opposed to within them, except emotion recognition (58% universal against 9% culture-specific).

What it means for Amadeus (inference)

• "The part of the brain we all share" is smaller than intuition says, and intuition is a bad guide. Several supposedly basic processes are culturally calibrated: illusion susceptibility, spatial frames, categorisation, reasoning style, moral foundations, self-enhancement, conformity, fairness norms. A base built from WEIRD intuitions about what is basic would build in WEIRD psychology.

• What does look shared:

  • learning machinery (ontogenetic programs, cultural learning);
  • analog number;
  • emotion recognition;
  • theory-of-mind ability (not its timing);
  • essentialism;
  • hierarchical folk taxonomy;
  • the five-factor structure (not trait levels);
  • punishing free-riders.

  Several of these match Spelke's core systems. They are candidates for base architecture or priors.

• Culture is a layer between base and person. Many differences are population-level, acquired by everyone growing up in a society: the frame of reference their language uses, their basic-level categories, their fairness norms. They are neither universal (base) nor individual (slot). The population prior (B8) must be the person's own reference population, not a global or WEIRD default.

• A text-trained vocabulary engine is WEIRD by construction if its corpus is mostly Western online text. Its "commonsense" defaults about self, choice, morality and reasoning are then those of an outlier population. This matters for P1: those defaults are committed values on dimensions where people differ.

• Heritability estimates themselves are WEIRD. The Turkheimer SES result shows that variance shares (R4e question 1) depend on the environment studied. Behaviour-genetic numbers from Western twin registries should not be read as human constants.