--- title: '10 Questions for Joseph Henrich' author: Gwern, GPT-4.5, ChatGPT-4 o1-pro description: "A short log of using an LLM interview-prompt to generate Joseph Henrich questions, with notes on what made the generated questions useful or generic." created: 2025-03-15 modified: 2026-05-10 status: finished importance: 0 confidence: log css-extension: dropcaps-de-zs toc-not backlink: False ... ::: abstract [Dwarkesh Patel](https://en.wikipedia.org/wiki/Dwarkesh_Patel){.link-live} [interviewed](https://www.dwarkesh.com/p/joseph-henrich) anthropologist [Joseph Henrich](https://en.wikipedia.org/wiki/Joseph_Henrich){.link-live} in early 2025. Curious about the process of developing good interview questions (which is central to the high quality of [*Playboy*](https://en.wikipedia.org/wiki/Playboy){.link-live}, [*The Paris Review*](https://en.wikipedia.org/wiki/The_Paris_Review){.link-live}, [*The New Yorker*](https://en.wikipedia.org/wiki/The_New_Yorker){.link-live}, and Dwarkesh Patel interviews), I experimented with using [ChatGPT](https://openai.com/blog/chatgpt/) to come up with *interesting* questions by considering hypothetical counterfactual answers. I developed what I eventually called [**the interview prompt**](/interview-inkhaven#interview-prompt), which initializes a brainstorming session by writing a biographical intellectual summary of a subject (serving as cognitive semantic [priming](https://en.wikipedia.org/wiki/Priming_(psychology)){.link-live} and avoiding the flat empty 'view from nowhere' [LLM](https://en.wikipedia.org/wiki/Large_language_model){.link-live} default). I found the resulting questions interesting enough to continue tinkering with the interview prompt and reusing it later. I think that '*N* Questions for *X*' might be an interesting writing genre, in that asking a good set of questions is a subtle art, done poorly by most interviewers, often interesting and thought-provoking in their own right, underrepresented training data for LLMs, gives people permission to ask the questions they've always idly wondered but never dared ask a famous person directly, and might well inspire the subject to write up their own response and answer some questions! ::: My initial biographical prompt: ::: collapse > We are brainstorming ideas for an upcoming Dwarkesh Patel podcast interview of American anthropologist Joseph Henrich, famous for cultural selection theses. We would like you to assist in coming up with questions. > > 'Good questions for Henrich', for us, are ones that are tailored to him, and couldn't be asked of anyone else; they are not redundant with past interviews, or cover obvious questions; they can challenge him, or focus on contradictions or unpleasant implications; they do not necessarily need to focus on his work, as we are often also interested in his career or psychology or how he works (similar to questions like [Tyler Cowen's](https://en.wikipedia.org/wiki/Tyler_Cowen){.link-live} interview questions, or the [*Paris Review* writing questions](/morning-writing)) or the role serendipity and luck played in making 'Joseph Henrich'. > > To do so, please write up a summary of what you know about Henrich: his ideas, his career, his professional relationships (major influences, mentors, advisers, students, allies, enemies, critics etc), what he might have already talked about. Then start brainstorming questions. > > To brainstorm a possible question, imagine hypothetical dialogues with Henrich: ask him a question, and list summaries of 3 of the responses he might give; then think of a followup question for each one. Do this for 10 possible questions, and pick the most interesting one and describe why. > > Here are some example questions that we might ask Henrich: > > - How do you feel about the poor [replication of social psychology literature](https://en.wikipedia.org/wiki/Replication_crisis){.link-live}, like WEIRD results? Or criticisms of anecdotes like ['caribou hunting randomization'](/doc/sociology/1983-vollweiler.pdf)? > > - Henrich says the critical moment was the Western church's crackdown on cousin marriage. But we see the beginnings of the split already in the ancient Greeks and Romans. The Romans were already proto-WEIRD with a sort of inchoate affection for impersonal institutions (eg. Brutus killing his sons), so maybe the Middle Ages were not what mattered? Why don't the ancient DNA studies on inbreeding measures like [runs of homozygosity](https://en.wikipedia.org/wiki/Runs_of_homozygosity){.link-live} seem to support this abrupt crackdown? > > - According to his theory, didn't modernity require something like the opposite of Chesterton's fence, where societies were dismantling kin-based networks without fully understanding their function, and it was through this period of disruption that modernity emerged? Was that just dumb luck by Western civilization, in his paradigm? that this craven, selfish, manipulation by the Church just happened to be the lottery ticket to the [Great Divergence](https://en.wikipedia.org/wiki/Great_Divergence){.link-live} and world dominance? > > - What about Sheilagh Ogilvie's critiques of the relevance of "the European marriage pattern" for economic development? > > - If institutions are the product of cultural evolution, does that mean we should all be cultural conservatives---was Chesterton basically right about everything, including his fence? > > - If cultural selection is the end-all be-all, then hasn't it shut down now that societies are no longer many tiny competing genocidal units but a single near-monolithic global culture where genocide is very déclassé? Shouldn't there be 'cultural dysgenics' as coordination mechanisms gradually decay under lack of selection of culture units? > > - What happens when AIs ingest all human culture? > > Before we start, please ask me any clarifying questions which might help you come up with really good questions for Henrich. > > ...1. I think ideally his trajectory & experiences would be of most interest; contradictions risk devolving into nitpicking or having been already addressed at length in his publications. 2. similarly, methodology/process are more likely to be fallow territory. 3. It would depend on whether you think Henrich can give good responses; we do not want to waste questions on the equivalent of a 'no comment'. That would be a bad question. 4. They can, if, similarly, there is the possibility of a good answer, rather than shying away protectively. 5. I'm not sure. What areas do *you* think are underexplored or neglected? That would be a good topic to imagine question-dialogues on and might yield a good question. ::: The [GPT-4.5](https://openai.com/index/introducing-gpt-4-5/) overview of Joseph Henrich: ::: collapse > Joseph Henrich is an anthropologist whose work integrates cultural anthropology, psychology, economics, and evolutionary theory. He is best known for developing and popularizing concepts such as: > > - **[WEIRD](https://en.wikipedia.org/wiki/WEIRD){.link-live} psychology**: Henrich argues psychological research has overly relied on samples drawn from Western, Educated, Industrialized, Rich, and Democratic societies. This biases our understanding of human psychology, suggesting human cognition is far more culturally malleable than typically assumed. > - **[Cultural evolution](https://en.wikipedia.org/wiki/Cultural_evolution){.link-live} & cumulative culture**: Henrich emphasizes cultural selection, arguing that human cognitive and behavioral traits evolved primarily through cultural mechanisms---particularly the transmission of knowledge, norms, and practices across generations---not just biological evolution. > - **Marriage practices and the [Catholic Church](https://en.wikipedia.org/wiki/Catholic_Church){.link-live}**: Henrich attributes Europe's distinct historical trajectory (eg. institutional innovation, individualism, and impersonal trust) to the medieval Catholic Church's [prohibitions on cousin marriage](https://en.wikipedia.org/wiki/Cousin_marriage#Catholic_Church_and_Europe){.link-live}, weakening kinship ties, and fostering cooperation among unrelated strangers. > - **Mentors, Allies, Influences**: Henrich studied under cultural evolutionists [Robert Boyd](https://en.wikipedia.org/wiki/Robert_Boyd){.link-live} and [Peter Richerson](https://en.wikipedia.org/wiki/Peter_Richerson){.link-live}, who shaped his interdisciplinary evolutionary perspective. He engages constructively with scholars like [Steven Pinker](https://en.wikipedia.org/wiki/Steven_Pinker){.link-live} (psychology/evolution), [Jonathan Haidt](https://en.wikipedia.org/wiki/Jonathan_Haidt){.link-live} (moral psychology), and [Daron Acemoglu](https://en.wikipedia.org/wiki/Daron_Acemoglu){.link-live} (economic institutions). > - **Critics & debates**: Henrich faces critiques from scholars like [Sheilagh Ogilvie](https://en.wikipedia.org/wiki/Sheilagh_Ogilvie){.link-live}, who questions his historical claims on marriage practices and economic outcomes, and from anthropologists skeptical of broad cultural evolutionary narratives that oversimplify complexity or imply cultural determinism. > - **Intellectual style and methodology**: He is data-driven, broadly synthetic, and interdisciplinary. His major publications (eg. [*The Secret of Our Success*](https://www.amazon.com/Secret-Our-Success-Evolution-Domesticating/dp/0691166854), [*The WEIRDest People in the World*](https://en.wikipedia.org/wiki/The_WEIRDest_People_in_the_World){.link-live}) extensively blend ethnography, historical evidence, economic theory, psychological experiments, and evolutionary modeling. > - **Career and serendipity**: Originally trained as an aerospace engineer before shifting to anthropology, Henrich's path exemplifies disciplinary boundary-crossing and luck-driven shifts. His early experiences living in indigenous communities shaped his understanding of cultural variation profoundly. ::: Initial GPT-4.5 batch: ::: collapse > **Personal Intellectual Journey**: > > - Your career began in aerospace engineering, seemingly unrelated to cultural anthropology. How did engineering shape your anthropological methodology or your approach to studying human culture? > > Has your engineering background ever led you astray, perhaps causing oversimplifications or blind spots? > > - Looking back, what's the luckiest event or encounter that importantly shaped your professional path? > > Do you think academia systematically underestimates the role luck plays in intellectual success? > > - Early in your fieldwork among indigenous communities, were there cultural practices or phenomena that you fundamentally misunderstood or undervalued? > > Did recognizing these early mistakes importantly alter how you approached subsequent research? > > **Cross-disciplinary Insights and Frictions**: > > - You've worked across multiple disciplines (anthropology, economics, psychology). Did you ever experience intellectual isolation or difficulty when crossing these disciplinary boundaries? > > Did these struggles reshape or refine any of your core ideas or methods? > > - Have you ever dismissed major ideas prematurely, only later realizing they held substantial value? > > Has reflecting on these mistakes changed how you evaluate or adopt new theories today? > > **AI and Cultural Evolution**: > > - You emphasize imitation as crucial to human cumulative culture. Yet, AI imitation-learning research has been relatively unsuccessful compared to [reinforcement learning](https://en.wikipedia.org/wiki/Reinforcement_learning){.link-live}. Does this gap suggest your imitation-based cultural evolution theory might be missing something fundamental? > > Could AI's difficulty in replicating human imitation indicate that cultural learning is cognitively special---or conversely, that your theories overestimate imitation's importance? > > - You've famously highlighted how WEIRD societies produce distinct psychologies. Does the global dominance of Internet culture---mostly Western-driven---risk spreading WEIRD psychology universally, diminishing psychological diversity? > > Would global cultural homogenization via the Internet be culturally adaptive or dangerous in your evolutionary framework? > > - Could the rise of advanced AI systems capable of curating or creating cultural content lead to mass extinction of human-generated cultural traditions? > > How might the loss of human-generated cultural traditions impact human identity and societal stability? > > - How does online anonymity---widespread on the Internet---alter cultural evolution, particularly given your emphasis on reputation and social norms? > > Could anonymity lead to cultural dysgenics, or might it actually create valuable cultural innovations? > > **Future Directions and Neglected Areas**: > > - What specific area of cultural evolution research do you believe is currently unjustifiably neglected---by yourself or the broader research community? > > Why do you think this area remains underexplored, despite its apparent importance? ::: [ChatGPT-4 o1-pro](https://openai.com/index/introducing-chatgpt-pro/) list of questions: ::: collapse > 1. **From Engineering to Anthropology**: How did your transition from aerospace engineering to anthropology shape your thinking on cultural evolution? > > - Did you ever clash with traditional anthropologists who found your engineering mindset too quantitative? > > - What advice would you offer to future scholars interested in crossing disciplinary boundaries? > > 2. **Origins of WEIRD Psychology**: Why do you see the 'Western, Educated, Industrialized, Rich, Democratic' populations as psychologically unique, and what are the most surprising real-world consequences of that difference? > > - What's the next major myth about human universals you think might crumble under cross-cultural evidence? > > - How do you address critics who claim your WEIRD framework overstates Western exceptionalism? > > 3. **Church Bans on Kin Marriage**: You argue that the medieval Church's crackdown on cousin marriage catalyzed Western individualism. How do you square that with earlier proto-individualism in ancient Rome or Greece? > > - Why did Church-enforced marriage norms have more lasting structural impact than civic laws in classical societies? > > - Could a similarly sweeping norm shift happen today through a different institution (state, multinational organization)? > > 4. **The Puzzle of Inbreeding Evidence**: Genetic data doesn't always show a clear drop in inbreeding at the time you propose. Are we missing some subtle effect, or might the genetic record challenge the scope of your claim? > > - If the Church's ban partially reduced but never eliminated cousin marriage, was that still enough to reshape social norms profoundly? > > - How do you recommend combining historical DNA with ethnographic or archival data to settle this question? > > 5. **Balancing Macro Models and Ethnographic Nuance**: Critics say your large-scale cultural-evolution claims can gloss over local complexity. Where do you draw the line between insightful simplification and missing key details? > > - Have you ever majorly revised a model once deep ethnographic findings contradicted your expectations? > > - How might future research integrate micro-level case studies without losing sight of big-picture patterns? > > 6. **Sheilagh Ogilvie and Counterexamples**: Sheilagh Ogilvie questions whether dismantled kin structures truly drove economic development. What specific data would convince you that tight-knit family or guild networks can be just as beneficial? > > - Could the European Marriage Pattern have had different effects across social classes or regions, making it more variable than your model suggests? > > - How do you see your macro claims holding up in the face of localized studies that point to alternate explanations for Europe's economic rise? > > 7. **Cultural Group Selection vs. Traditional [Multi-Level Selection](https://en.wikipedia.org/wiki/Multi-Level_Selection){.link-live}**: How do you differentiate your notion of cultural group selection from classic biological group-selection models, which many still find controversial? > > - In real-world settings, how do we measure when cultural competition among groups overrides individual-level incentives? > > - Do you rely on genetic predispositions for cooperation, or is your framework fully cultural? > > 8. **Top-Down Influence in the Internet Age**: Today, social-media algorithms and tech corporations shape which norms go viral. Does this top-down curation alter the bottom-up group-selection logic central to your work? > > - Could the internet's echo chambers create "digital tribes" competing as if they were distinct cultures? > > - Is there a risk that algorithmic curation accelerates harmful cultural norms before counter-norms can emerge? > > 9. **AI as a Cultural Participant**: Generative AI can produce art, text, and even interact conversationally. If AI 'agents' begin shaping norms, do we consider them part of cultural evolution or external forces manipulating it? > > - Could AI systems develop their own internal "traditions" or iterative learning that parallel human cultural evolution? > > - Where would accountability lie if AI-driven norms turn out to be maladaptive on a large scale? > > 10. **Global Cultural Convergence or Divergence**: In a highly connected world, do you foresee new, distinct cultural evolutions arising from digital spaces---or are we headed for a monolithic global culture where meaningful variation and competition vanish? > > - What might cultural evolution look like when every local tradition is instantly exposed to global scrutiny or adoption? [\[cf. ["Internet Community **Design**: Slow To Fast"](/internet-community)\]]{.editorial} > > - Does a lack of inter-societal competition risk stagnation, akin to a species in a non-competitive ecological niche? ::: [GPT-5.5 Pro](https://openai.com/index/introducing-gpt-5-5/#pro) final revision: > 1. **Engineering Habits**: You began as an aerospace engineer. Which engineering habit helped you see cultural evolution, and which habit was poisonous? > > - What did anthropologists notice that modelers missed? > - Name one case where fieldwork overruled a clean model. > > 2. **Fieldwork Error**: What did you most misunderstand in early fieldwork because your default psychology was WEIRD? > > - Did that mistake later become part of the theory? > - Are there still ethnographic cases where your model feels like a bad fit? > > 3. **WEIRD Survivors**: Since the WEIRD critique became unavoidable, which claimed human universal has most deservedly died? > > - Which alleged WEIRD difference has survived better cross-cultural data better than you expected? > - What psychological result still embarrasses the field because it is probably a convenience-sample artifact? > > 4. **Rome Before Church**: If Greece and Rome already had impersonal law, civic institutions, and partial nuclear-family tendencies, why give medieval Church kinship policy causal priority? > > - What evidence distinguishes "the Church caused Western individualism" from "the Church amplified an older Mediterranean trajectory"? > - What would you expect to see in a Rome-without-Christianity counterfactual? > > 5. **Kinship Without Genes**: If medieval aDNA or runs of homozygosity fail to show a sharp inbreeding break, which part of the Church thesis survives: mating, inheritance, household formation, corporate landholding, or norm psychology? > > - What genetic result would most trouble the simple version of the thesis? > - What archival result would rescue it? > > 6. **Ogilvie's Best Case**: What is the strongest Ogilvie-style case where guilds, kin networks, or closed corporate institutions increased development rather than blocking it? > > - What would make you downgrade the anti-kinship mechanism? > - Did the European Marriage Pattern have opposite effects across classes, regions, or occupations? > > 7. **Cultural Selection Test**: What is the cleanest empirical example of cultural group selection where we can specify the units, selection pressure, counterfactual, and outcome without post-hoc storytelling? > > - Which celebrated cultural-evolution story is most at risk of being a just-so story? > - What would a failed cultural group-selection prediction look like? > > 8. **Chesterton's Fence**: If useful institutions emerge by blind cultural selection, how much deference should reformers grant traditions they cannot explain? > > - When should a society dismantle a tradition before it understands its function? > - Is your framework more Burkean than you are personally comfortable with? > > 9. **Selection After Peace**: If group competition historically selected institutions, what selects institutions after conquest, exit, fragmentation, and replacement are suppressed by states, globalization, and humanitarian norms? > > - Is "cultural dysgenics" a real category or a misleading analogy? > - What modern institution is decaying because the old selection pressure on it disappeared? > > 10. **AI as Selector**: LLMs ingest human cultural variation and emit cheap, RLHF-filtered recombinations. Are they accelerants of cultural evolution, homogenizers, or new selectors outside the human imitation ecology? > > - What would count as an AI "tradition" rather than model weights plus prompt habits? > - What part of your cultural-learning model breaks first when the transmitter is non-human? ```{=HTML}
[]
```