Theory
Formalize how institutions compress social reality, allocate judgment, and create representation wedges.
Behavior · computation · governance
We study how institutions represent the world, how those representations shape judgment and welfare, and how technology can help institutions reason more faithfully—and responsibly.
The institution retains enough detail to act while making its simplifications visible and contestable.
The argument
Our point of departure
A risk score, legal category, case file, performance metric, or algorithm is a compressed model of reality. The model may be necessary. It is never neutral.
Rules, categories, procedures, and technologies are representations of reality. They make collective action possible by deciding what will count.
Where we build
Two major initiatives
The program
How the work fits together
Formalize how institutions compress social reality, allocate judgment, and create representation wedges.
Study how information, discretion, timing, language, and organizational rules change real decisions.
Build datasets and observatories that make previously invisible institutional systems measurable.
Design and test tools that help institutions reason without surrendering responsibility.
Selected work in motion
The Lab
Faculty Director
PhD Student
PhD Student
PhD Student
PhD Student
Assistant Project Scientist
Assistant Specialist
Bellwether Postdoctoral Scholar