My research asks how speakers package information so that a listener can recover it; teaching is the same problem in practice. The goal is students who can evaluate evidence for themselves and argue for their conclusions.
In the classroom
9.59 Laboratory in Psycholinguistics, MIT
Teaching assistant since 2023, with Ted Gibson. The course replaces exams with papers in which students report experiments they run themselves, and I designed its statistics curriculum backward from that goal: every lab and problem set builds toward the students’ own experiments, from design to mixed-effects models.
On the first day, students rate sentences on their phones. That dataset returns all semester as the analyses grow more sophisticated.
New ideas arrive as answers to problems students can already see: Anscombe’s quartet for why we plot data, Simpson’s paradox for why we need mixed-effects models.
Every assignment feeds into a final paper written to journal standards. One became a paper in Cognition with its student author (Poliak, Chen, & Gibson, 2025).
Statistics across MIT Brain and Cognitive Sciences
Practica on interpreting linear models, making effective figures, mixed-effects models, and Bayesian statistics, and consulting on analyses for projects across labs.
Computer science, Harvard
Teaching fellow for four courses, including CS50, 2019–2021.
English for speakers of other languages
Volunteer teacher with The Welcome Project, 2023–2024.
From students
“…the absolute care that Moshe had for students, the lack of judgement and the desire for students to learn.”
“I feel like I actually understand statistics for the first time.”
From 9.59 course evaluations, 2023–2025, in which 20 of 25 students rated the overall teaching 7 out of 7.
Courses I am ready to teach
Laboratory in Psycholinguistics, for advanced undergraduates and PhD students
Introduction to data analysis and coding in R
Linear and mixed-effects models, and advanced statistical methods
Introduction to psychology or to linguistics
Seminar: Language as rational communication, connecting human comprehension and production with language models in AI
Mentoring research
Since 2023 I have mentored more than 30 undergraduate researchers from MIT, Wellesley, and Harvard. Many cross-linguistic studies started this way: students whose native language is not English adapt an English experiment to their own language, which is how our production experiments came to span 17 languages.
Goals, set together
Each semester begins with a written reflection. The student sets out their learning goals, I set out the project’s goals, and we build a roadmap together, revisiting it at midsemester and at the end.
From guidance to independence
In a first semester I guide every detail. Students who stay take on more of the decisions, much like junior graduate students, with growing ownership of the design and materials.
A bar worth reaching
A challenging bar tells students their abilities are trusted, and invites them to go past it.
Hannah Kimura joined the misheard-lyrics project by choosing songs and reading about noisy-channel processing. When she asked for more of a challenge, she fine-tuned the language model BERT on song lyrics, and by the end she was building whole studies on her own. She is a co-author on our CogSci 2024 paper and is now pursuing a master’s in computer science at Georgia Tech.
Student researchers
Students have joined through MIT’s UROP program, the MSRP-Bio summer program, and funded positions. Among those who stayed for several semesters or co-authored papers:
Ayla Sumer · 2026
Aimee Wang · 2024–2026
Curtis Chen · a 9.59 class project that became Poliak, Chen, & Gibson (2025), Cognition
Hannah Kimura · 2023 · misheard song lyrics, CogSci 2024 · now at Georgia Tech
Getting involved
Interested in research on language?
Undergraduates from any major are welcome, and no prior research experience is needed. If you speak a language other than English, there may be a study to run in it, but that is not a requirement.
To start a conversation, email moshepol@mit.edu with a few lines about what interests you. If you think I can help with something else, such as statistics or applying to graduate school, please write as well.