Abstract
Away from equilibrium, no local quantity predicts where a driven system will settle. Low rattling theory recovers prediction by giving up generality: typical driven systems accumulate where they are slow to leave, producing order with no designer, and only when the drive is predictable. I trace this upward, from thermodynamic self-organisation to systems readable as Bayesian reasoners to agency as an effective causal description relative to an observer's means of measuring and acting. The ladder ends by turning on itself, with consequences for how we practise science and study our own minds.
About the Speaker
Pavel Chvykov works across complexity science, contemplative practice, and the meta-questions of how research itself gets done. His PhD and postdoc at MIT, with Jeremy England, produced "Low Rattling" (Science, 2021) - a predictive principle for self-organisation of driven matter, and “Causal Geometry" with Erik Hoel (Entropy, 2021), on the observer-dependence of causal models. He is now connecting that work to active inference and to other models of agency and observation. This bridge between systems physics and the observer’s cognition led him to co-found the Complexity Science and Contemplative Studies Consortium, following a conviction that rigorous inquiry should transform the inquirer, not only the literature.
Practical
This event is entirely online through Microsoft Teams. Join here.
