Upcoming talks
16 May 2026
Chris Smeenk (Western / UCLA), “To Save the CMB”
Inflationary cosmology has remained the consensus account of early universe physics, surviving four decades of increasingly precise CMB observations. But sustained compatibility with observations is not the same as the kind of empirical success that warrants treating a theory as a secure foundation for further inquiry, the kind of success that, following George Smith’s account, comes from closing the loop — a standard we take other parts of the Standard Model of Cosmology to have achieved. This chapter argues that inflation has not yet achieved such success, and diagnoses why.
Closing the loop requires a stable theoretical framework that supports iterative refinement, which is both specific enough that discrepancies between calculation and observation are informative, and stable enough that they prompt the identification of new physical effects rather than parameter adjustments. Securing such a framework requires a prior step: identifying, across the class of models compatible with observations, the common physical structure actually doing the evidential work. But this structure has to itself be physically significant, appropriately constrained by independent lines of evidence, to serve as a stable starting point for an iterative approach. These two requirements — the search for common structure across competing models, and the demand that this structure be anchored in independently constrained physics — were the two strategies Kepler artciluated in response to skeptical challenges to astronomical theorizing.
Identifying inflation’s physical source, rather than treating the “inflaton” as a field engineered to fit observations, has been an open issue since the theory was first proposed. The effective field theory of inflation is a response to this longstanding issue: it provides a clear characterization of the common core of inflationary models, isolating a zeroth-order action shared by all single-field slow-roll models. But the core it identifies is phenomenological rather than physically significant: it captures the kinematics of a quasi–de Sitter background without specifying what physical degrees of freedom drove it. And it is fragile: the eta-problem and trans-Planckian sensitivity, neither generic to EFTs, mean that leading-order predictions depend on UV physics the framework cannot itself justify. On our analysis, the thinness and fragility of the common core show that inflation currently fails to provide the framework that closing the loop requires.
A draft book chapter will be distributed prior to the talk.
Past talks
4 April 2026
James Read (Oxford), “Affine connections for Galilean and Carrollian structures and non-relativistic Weyl theorems“
I present a classification of general Galilean/Carrollian structures, permitting affine connections with both torsion and non-metricity. I then: (i) show how non-/ultra-relativistic geometric trinities of gravity fit into this framework, (ii) leverage notions of non-/ultra-relativistic conformal structure to prove respective versions of Weyl’s famous (1921) theorem (that Weyl metrics are fixed by their projective and conformal structures), and (iii) make some conceptual remarks regarding limits of the Weyl tensor.
The talk will be particle based on James’ recent CQG paper.
7 February 2026
Ellen Shi (UC Irvine), “Does the Einstein Algebra Formalism Favor Relationalism? A New Structural Comparison”
In light of Chen (2024)’s recent objection to Rosenstock et al. (2015), this paper reconsiders the question “does the Einstein algebra formalism favor relationalism?”. Following the structural comparison approach adopted by Rosenstock et al., I propose a new formal criterion to investigate the question in place of the categorical criterion of theoretical equivalence, inspired by John Earman’s program of Leibniz algebras. Based on the new criterion, the paper shows that the Einstein algebra formalism does not favor relationalism. It re-affirms Rosenstock et al.’s conclusion with a new technical result that is not subject to Chen’s objection.
Please read Ellen’s preprint prior to the meeting.
10 January 2026
Mahmoud Jalloh (Caltech), “Constants of Nature, Law Construal, and Theory Individuation“
The debate regarding whether physics requires absolute or merely comparative quantitative structure has surfaced a question regarding the nomological status of the constants: Are their magnitudes necessary or contingent? I argue here that this is merely a special case of a more general question: To what fineness of grain ought we construe the laws? Answering this question requires balancing the epistemic significance of the laws with the robustness of our standards for individuating theories. It is found that a relatively coarse-grained construal of the laws, making the constants nomologically contingent, best balances these considerations in a manner. To show this, a methodological primitivism is adopted, such that the laws are considered as equations, with differing degrees of structure. The interpretative equilibrium found construes the laws to have algebraic and polarity structure, but not magnitude structure. Not only is this of general significance for accounts of the laws of nature, but it also provides the contingentist comparativist an answer to accusations of fallacious theory equivocation.
Please read Mahmoud’s preprint prior to the meeting.
25 October 2025
Eddy Chen (UCSD), “Typical Quantum States of the Universe are Observationally Indistinguishable”
(Joint work with Roderich Tumulka) We establish three new impossibility results regarding our knowledge of the quantum state of the universe — a central object in quantum theory. We show that, if the universal quantum state is a typical unit vector from a high-dimensional subspace H_0 of Hilbert space H (such as the one defined by a low-entropy macro-state as prescribed by the Past Hypothesis), then no observation can determine or just significantly narrow down which vector it is. In other words, the overwhelming majority of possible state vectors are observationally indistinguishable from each other (and from the density matrix of H_0). Moreover, we show that for any observation that isn’t too unlikely and most pairs of unit vectors from H_0, the observation will not significantly favor one vector over the other. We further show that the uniform distribution over the unit sphere in H_0, after Bayesian updating in the light of any observation that isn’t too unlikely, is still extremely close to uniform. Our arguments rely on a typicality theorem from quantum statistical mechanics. We also discuss how theoretical considerations beyond empirical evidence might inform our understanding of this fact and our knowledge of the universal quantum state.
Participants can read a pre-print of this work here: https://arxiv.org/abs/2410.16860
10 January 2026, Mahmoud Jalloh (Caltech)
7 February 2026, Ellen Shi (UC Irvine / UC Berkeley)
7 March 2026, Porter Williams (Pittsburgh)
4 April 2026, James Read (Oxford)
16 May 2026, Chris Smeenk (Western / UCLA)
