Soren Schlassa

I'm a PhD candidate in the NYU Department of Philosophy, and undergrad emeritus at Carleton College. 

I work on theoretical questions about computation and representation, both biological and artificial. Mostly, I study what gives representations their structure. 

One project is about what it is to perceive objects, i.e. treat some parts of the world as "going together in a unified whole." I think the answer is computational. Another project is about the difference between analog and symbolic representation. A third is on the various structures computational objects can have. You can compute with vectors, magnitudes, strings, trees, lists... What is the difference?

I also teach about philosophical issues in AI, both normative and theoretical. 

Language Models and Philosophy

Barnard College, Fall '26

Philosophy of Artificial Intelligence

Barnard College, Fall '25

Logic

NYU, Fall '25

Minds and Machines

NYU, Summer '25

Philosophical Applications of Cognitive Science

NYU, Summer '22

Global Ethics (TA)

Instructor: Kwame Anthony Appiah, Spring '23

Philosophy of Language (TA)

Instructor: Matt Mandelkern, Fall '22

Philosophy of Mind (TA)

Instructor: Verónica Gómez Sánchez, Spring '22

Philosophical Applications of Cog Sci (TA)

Instructor: Michael Strevens, Fall '21

My philosophy-teaching philosophy: I really like teaching philosophy. My first goal is to get students thinking like philosophers. To me that means framing hypotheses, clarifying them, applying them to cases, revising them, and rooting out misunderstanding. I think studying philosophy is a particularly good way to practice these things. I put a lot of effort into making my classes welcoming and collaborative.

My dissertation committee is Michael Strevens, Ned Block, and David Chalmers.

Papers I'm working on:

"How to See Things"

Develops a computational account of what it is to visually represent objects, drawing heavily on empirical work. 

[paper under review]

Criticizes recent accounts of the iconic/analog/symbolic. Proposes a fix.

"A Theory of Computational Structure"

Gives a theory of the space of structures of computational vehicles and their implementation conditions. Uses case studies from neural networks and brains.

Technically, I have an Erdös number of 4 (Erdös - M. Kac - T. Jacobson - B. Allen - me), if you count these papers with nearly a thousand authors each. Maybe a better metric would be to weight each link by the log of the number of authors on the paper. Alas, on that definition my number goes up to 15...