Title page of our paper, “The Politics of Black Classification: Sociopolitical Cues and Racial Perception,” with Lauren Davenport (Stanford) and Hunter Rendleman (UC Berkeley), dated April 14, 2026. Abstract: What makes someone Black in American society today? From Donald Trump questioning Kamala Harris’s racial identity to Joe Biden’s claim that hesitant Black voters “ain’t Black,” American politics frequently brings questions of racial authenticity and belonging to the surface. Yet political science often approaches race as a fixed attribute rather than a social construction. Here, we seek to understand how Americans define blackness in social and political life. Using a conjoint experiment with a racially diverse sample that includes Black, white, and mixed race Black-white respondents, we evaluate how ascribed and acquired traits influence perceptions of blackness. The results show that inherited characteristics—particularly parentage and skin tone, which are the strongest determinants of racial classification—play a central role, while sociopolitical cues such as partisanship, neighborhood context, and spousal race also influence racial classification. Using a continuous measure, we also show that respondents make graded assessments of blackness rather than purely binary classifications, with some individuals perceived as more Black than others. Black respondents are more likely than white respondents to classify a broader set of profiles as Black, consistent with a more inclusive understanding of racial membership, yet they also place greater emphasis on shared political identity. These findings clarify how racial categories are socially constructed and why that construction carries real political and social consequences.

Our paper, “The Politics of Black Classification: Sociopolitical Cues and Racial Perception” (w/ Lauren Davenport & @hrendleman.bsky.social), has been conditionally accepted at Perspectives on Politics!

Sharing abstract below. Long time coming, but we are really proud of this paper.

More soon!

SimBench now at #ICLR2026!
Often in social simulations, the goal is not to predict what one specific person will do. It is to estimate how a group will respond, whether in pre-testing a real polling question, or in stress-testing a policy or intervention before running it in the real world.

Such amazing news, congrats!!!

🧵 Very proud our special collection on vocal communication in corvids in Animal Cognition (@springernature.com) is now fully published! 🐦

Huge thanks to all collaborators and contributors who made this possible! (1/15)
link.springer.com/collections/...

Empowerment Gain and Causal Model Construction: Children and adults are sensitive to controllability and variability in their causal interventions Learning about the causal structure of the world is a fundamental problem for human cognition. Causal models and especially causal learning have proved to be difficult for large pretrained models usin...

New preprint of a paper with Eunice Yiu to appear in Philosophical Transactions A, Special issue: World models, 2026. The theoretical link between empowerment in RL and Bayesian causal models with cool new data. arxiv.org/abs/2512.08230

I am excited to share my first publication in Behavior Research Methods! 🚀

We’ve released ChildLens: 109 hours of open-access, annotated egocentric video/audio from children (ages 3–5).

Paper: rdcu.be/fdfAI

New paper out in Proceedings of the Royal Society B: we apply linguistic tools to sperm whale vowels.

The result: sperm whale vowels do not just look like human vowels. They also behave like them.

We found several parallels. Like in Latin, whales have short and long vowels.

Medial temporal lobe encodes cognitive maps of real-world social networks | PNAS Humans routinely solve social problems by navigating densely interconnected networks—gossiping strategically, brokering across cliques, and coordin...

Now out in PNAS with @jaeyoungson.bsky.social, Alice Xia, @apaxon.bsky.social & @orielf.bsky.social. Medial temporal lobe encodes predictive representations of people's real-world social networks which afford them key advantages in social navigation. www.pnas.org/doi/10.1073/... 🧵

How do people decide whether it’s wrong to harm a pig? A chimp? A baby?

In the West, these judgments are based more on *experience* (being able to feel) than on *agency* (being able to think and act)

New study finds that same pattern across other cultures

journals.sagepub.com/doi/10.1177/...

New preprint (by Vandendriessche et al.)

In everyday life, choices often lead to multiple simultaneous outcomes — some positive, some negative. Yet most reinforcement learning research has focused on situations where each choice produces only a single outcome 1/5

osf.io/preprints/ps...

Rapid concerted switching of the neural code in the inferotemporal cortex - Nature Face cells in the macaque inferotemporal cortex are initially able to detect faces and then rapidly switch to a face-specific neural code to discriminate between different face identities.

This looks like a significant discovery from Doris Tao's lab:

Rapid concerted switching of the neural code in the inferotemporal cortex
@nature.com

"..our findings indicate that there is a previously unknown mechanism for neural representation:.."

www.nature.com/articles/s41...

In Barcelona for #chi2026! Presenting our work on eliciting LLMs' assumptions about users, and how this mismatches with user expectations, in the Tues poster session! (Spoiler: users assume that LLMs give objective info much more than they actually do, which leads to sycophancy)

The Cognitive Tools Lab at Stanford (cogtoolslab.github.io) is recruiting two new research staff members to join in AY 26-27.
Full-Time Lab Manager: forms.gle/UVwfx5wbY9Km....
IRiSS Predoc Researcher: iriss.stanford.edu/predoc/2026-....
Please share widely in your networks, thank you!!

CHI 2026 is right around the corner. Here is some cool stuff from my lab.

1. Counterfactual Explanations for Agentic Workflows (github.com/eilab-gt/eil...) . A new API for making AI pipelines that automatically makes your pipeline explainable. Counterfactuals let you explore what *could* happen.
1/

Our paper finding that infants infer helpers’ relationships, and not their dispositions, is now out in PNAS! Sharing in case anyone needs something to read on the way home from #CDS2026 ;)

doi.org/10.1073/pnas...

Parallel by design? Meaning and grammar in single-stream and dual-stream neural network architectures A major current approach in theoretical linguistics proposes that there is only a partial, defeasible correspondence between syntax and semantics: meaning and grammar are autonomous, parallel compo...

New paper with Olivier Michalon. We ask whether transformers that separate syntax and semantics into parallel streams classify sentences better than single-stream models. They do not. We think this matters for how the autonomy of meaning and grammar should be explained. 1/6
doi.org/10.1080/0952...

🚨 NSF is already quietly eliminating the SBE Directorate, despite Congress’ mandate that NSF support the behavioral & social sciences.

Steps to counter this are in motion.

If you
- have an SBE proposal under review
- serve on an SBE grant panel

You can help! Fill out this form: shorturl.at/xuKw2

www.science.org/doi/10.1126/...

"mental imagery reactivates the same sensory codes used during visual stimuli, suggesting the existence of a generative model capable of synthesizing detailed sensory contents from an abstract, semantic representation."

Quantifier-Specific Usage Patterns Shape Learning—A Corpus Analysis on Universal Quantifiers in English and Dutch Child-Directed Language - Mieke Sarah Slim, Elizabeth Tobyn, Caroline F. Rowland, 2026 Quantifiers specify semantic relations between sentence constituents. Due to their meanings, they form a learning challenge for children. This challenge is made...

Quantifier-specific usage patterns shape learning—A corpus analysis on universal quantifiers in English and Dutch child-directed language. Final version by @miekeslim.bsky.social , Elizabeth Tobyn & @carorowland.bsky.social
doi.org/10.1177/01427237251385837

📄 New preprint: The blessing and curse of Value-Shaping imitation
(by @isabellehoxha.bsky.social)
osf.io/preprints/ps...
Imitation is central to human learning: but not all imitation processes are equally adaptive. We study their computational properties using reinforcement learning models.

What enables human language? A biocultural framework Explaining the origins of language is a key challenge in understanding ourselves as a species. We present an empirical framework that draws on synergies across fields to facilitate robust studies of language evolution. The approach is multifaceted, ...

“The rich complex nature of human social ecologies creates new challenges, which select for cognitive capacities that support distinctive forms of communication & facilitate creation of linguistic conventions in turn.” @thomscottphillips.bsky.social & co in a valuable eLetter response to our paper.🧪

Massive budget cuts for US science proposed again by Trump administration Budget proposal would also curb federal payments for scientific publishing.

Concerned about the possible dismantling of the Social, Behavioral, and Economic Directorate of NSF?? www.nature.com/articles/d41...

Find out what's happening and what we can do to protect behavioral science with @fabbs.org Friday April 10th at 1:30 ET. Register at us02web.zoom.us/meeting/regi...

🔥 We're very pleased to release our latest study 🧠: "Temporal structure of the language hierarchy within small cortical patches"
Paper → arxiv.org/abs/2604.03021
🧵 Summary thread below: 1/7

happy #CDS2026 to those who celebrate!

The narrow search effect and how broadening search promotes belief updating | PNAS Information search platforms, from Google to AI-assisted search engines, have transformed information access but may fail to promote a shared factu...

The paper, led by Eugina Leung, is here:
www.pnas.org/doi/10.1073/...

We document that the human tendency for confirmation bias in question framing (the tendency of people to frame search in terms of their prior beliefs) and algorithms that optimize for relevance combine to impede belief updating.

Screen shot of part of an award's event history with sources: 2025-05-26: Award termination date in TAGGS PDF (TAGGS PDF) 2025-05-27: 'TERMINATED' action added (HHS TAGGS) 2025-05-27: End date cut off from 2026-05-31 to 2025-08-31 (USAspending awards) 2025-05-31: Other amendment (USAspending File C) 2025-06-03: Grant marked as terminated (RePORTER website) 2025-06-13: Added to TAGGS PDF of terminations (TAGGS PDF) 2025-06-18: Court-ordered reinstatement (MA v. RFK ) (Court case) 2025-06-18: Court-ordered reinstatement (APHA v. NIH ) (Court case) 2025-06-30: End date extended from 2025-05-31 to 2026-05-31 (USAspending File C) 2025-07-03: Removed from TAGGS PDF of terminations (TAGGS PDF)

We've released a major update to our #NIH tracking at Grant Witness. We now publish a complete, multi-source history of events to track the complex administrative and legal history of each grant.
grant-witness.us/posts/2026-0...

📣 So happy to share with you the results of my main PhD study, where two human predators worked together to hunt a human prey in Minecraft, and had their brains scanned simultaneously using fNIRS 🎮🧠

We find that the brains of the predators are synchronised in the prefrontal cortex during hunting! 💡

Some bears will appear in ACL Findings this year! Congrats to Sriram and Siyuan for their hard work -- I am excited about this paper for two main reasons: 1) I got to express a little bit about what I think about language and thought; 2) this was one of the smoothest projects i've worked on!

American and Israeli missiles hit Sharif University in Tehran tonight. Often called #Iran’s MIT, the institution has been home to some of the country’s best and brightest minds in science, technology, engineering, and mathematics since 1966. 1/2

The world building here is dark imo. In this world a cop doing that test at that moment is pretty blatant racial discrimination. So the "uh oh" isn't about the test, but about what the Knight has just learned re the dispositions of the cop who pulled him over. The body cam will be off.