Foundations of Thermodynamics Group

About the Group

I run the Foundations of Thermodynamics Group at NTU Singapore, which is currently funded by the Nanyang Assistant Professorship Grant. Its goal is to provide a space for research on all aspects of the conceptual foundations of thermodynamics, with particular focus on understanding the conceptual foundations of thermodynamics beyond the classical domain, e.g., in the quantum and relativistic domains, how thermodynamics relates to other non-fundamental domains such as fluid dynamics and neuroscience, and how it bears on questions in philosophy such as (but not limited to) questions about emergence, reduction, inter-theory relations, the nature of non-fundamental reality, the laws of nature, and the relation between symmetries and reality. 

The group is currently supported by Research Associate Mel Ong. 

We are in the midst of hiring two postdoctoral Research Fellows. Stay tuned for more news! 

2025 Foundations of Thermodynamics Workshop

Nanyang Technological University Singapore

16-18 July 2025

Venue: NTU School of Physical and Mathematical Sciences, SPMS-03-02 (LT3)

Organizers: Eugene Y. S. Chua (Nanyang Technological University) and 

Wayne Myrvold (University of Western Ontario)

In collaboration with Wayne Myrvold from the University of Western Ontario, the Foundations of Thermodynamics Group will be organizing the Foundations of Thermodynamics Workshop 2025. The first two days will feature a series of talks on the philosophy of thermodynamics and statistical mechanics, while the last day will feature a roundtable discussion on the status of the two dominant approaches to understanding statistical mechanics, the Gibbsian and Boltzmannian approaches. 

Attendance is free, but space is limited. There will also be a Zoom option for those who can't attend in person. There will be lunch catered for attendees. If you would like to attend, please contact Mel at melissa.ongzy@ntu.edu.sg to indicate interest. 

The Quantum Thermodynamics 2025 (QTD2025) Conference will be held the week before the workshop, on 7-11 July 2025. Afficionados of thermodynamics are encouraged to attend the conference, which will be held at the National University of Singapore. Details and registration here: https://qtd2025.quantumlah.org/ 

Speakers and Abstracts 

Eddy Keming Chen (UC San Diego)

Typical Quantum States of the Universe are Observationally Indistinguishable 

This paper, which is joint work with Roderich Tumulka, is about the epistemology of quantum theory. We establish a new result about a limitation to knowledge of its central object --- the quantum state of the universe. We show that, if the universal quantum state can be assumed to be a typical unit vector from a high-dimensional subspace of Hilbert space (such as the subspace defined by a low-entropy macro-state as prescribed by the Past Hypothesis), then no observation can determine (or even just narrow down significantly) which vector it is. Typical state vectors, in other words, are observationally indistinguishable from each other. Our argument is based on a typicality theorem from quantum statistical mechanics. We also discuss how theoretical considerations that go beyond the empirical evidence might bear on this fact and on our knowledge of the universal quantum state. 

---

Eugene Y. S. Chua (Nanyang Technological University)

Looking for Work in Quantum Thermodynamics

Recent foundational work in quantum thermodynamics (QTD) reveals a `falling apart' of various classical thermodynamic concepts, akin to the situation in relativistic thermodynamics discussed by Chua (2023). This paper focuses on one such cluster of conceptual issues for the thermodynamic work raised by a widely-discussed no-go theorem due to Perarnau-Llobet et al (2017). The theorem asserts that no unique observable for thermodynamic work can be defined in QTD, given some intuitive desiderata widely accepted in the QTD community. I diagnose the no-go theorem and argue that these desiderata encode physically distinct scenarios in quantum mechanics for reasons to do with the quantum measurement problem, explaining why a single observable cannot guarantee statistics consistent with all the desiderata at once. I raise four issues. Firstly, some historical analysis of the work concept reveals that different procedures which were consilient in defining work in the classical regime no longer agree in the quantum regime, and the classical work concept falls apart in the quantum regime. Secondly, the quantum potential, available to Bohmian mechanics and many-interacting-worlds theories, allows one to define a non-contextual quantity that explains why the desiderata in the no-go theorem are different, with some curious caveats. Thirdly, I discuss how those resistant to quantum foundations have adopted an alternative conceptual strategy -- measurement scheme contextualism -- to grapple with real-world experimental scenarios and `get work done', in light of the lack of a fundamental work observable. Finally, I end by contrasting the conceptual situation here against that of relativistic thermodynamics. 

---

Bixin Guo (Macalester College)

Making Sense of Statistical Mechanics: The Gibbsian vs the Boltzmannian Approach 

In the literature on the foundations of statistical mechanics, we find two main approaches: the so-called Boltzmannian approach and the Gibbsian approach. The Boltzmannian approach has been defended extensively in philosophy of physics literature. In contrast, less attention has been paid to the Gibbsian approach, which is often criticized as conceptually unsatisfactory, and dismissed as a set of mathematical tools, rather than as a physical theory that describes what is actually going on in thermodynamic systems. In response, I argue for the physical significance of the Gibbsian approach, and formulate a new proposal regarding its relationship with the Boltzmannian approach. I argue that statistical mechanics is a framework theory, which covers a wide range of concrete theories of different kinds of systems. In contrast, the Boltzmannian approach faces challenges in achieving such broad applicability, and is thus better viewed as a concrete theory that applies only to certain physical systems. 

---

Yichen Luo (University of Western Ontario)

Quasi-Local Black Hole Thermodynamics

Traditional arguments for black hole thermodynamics typically require black hole stationary states to be the systems of interest. Such black hole stationary states are characterized by the notion of an event horizon. However, the global and teleological features of the event horizon can be shown to be particularly problematic in the thermodynamic context; hence crucially undermine the status of black hole thermodynamics. In this paper, I discuss recent developments of quasi-local horizons, especially the framework of isolated-dynamical horizons; I argue that adopting this quasi-local framework removes the blemish in stationary black hole thermodynamics and further strengthens the physicality of black hole thermodynamics. In particular, I show that (i) the globally stationary states are a limiting case hence approximation in light of the quasi-local framework, that (ii) various thermodynamic laws defined on quasi-local horizons present their deeper physical significance, and that (iii) the quasi-local framework allows a more general discussion of black hole thermodynamics which could potentially pave the way for the underlying statistical mechanical descriptions.

---

Patricia Palacios (University of Salzburg)

Something's Gotta Give: Empirical and Non-Empirical Strategies to Cope with Information Loss 

According to theoretical calculations, black holes lose energy due to Hawking radiation, and eventually evaporate. This gives rise to what is known as the "information loss paradox": according to quantum mechanics, it should be possible, from a complete specification of the quantum state at a later time, to recover states at earlier time, but the post-evaporation state apparently contains no details about the matter that fell into the black hole. This is one of the most outstanding issues in contemporary physics, as it reveals contradictions between quantum mechanics and general relativity. In this contribution, I will analyze the use of different alternative empirical and non-empirical methods to cope with the puzzle of information loss. Although I will argue that the resolution will largely depend on our preferred theory, I will stress the role of analogue experiments, thought experiments and intertheory reduction for choosing a particular resolution to the paradox. 

---

Katie Robertson (Stirling University) - by Zoom

Roundtable Discussant

---

Jos Uffink (University of Minnesota)

The "Schism" between Boltzmannian and Gibbsian Statistical Mechanics 

Drawing upon (Uffink 2007), I will review the conceptual structure of both the Boltzmannian and Gibbsian approach to Statistical Mechanics, and highlight their most important differences. I will next discuss the ways in which more recent authors have characterized the relationship between these two approaches (e.g.: Werndl & Frigg 2018; Lazarovici 2018, Frigg & Werndl 2019, Anta 2021). From this discussion I hope to clarify the relationship between these two approaches and draw conclusions about which one is preferable depending, of course, on a choice of desiderata. 

---

Giovanni Valente (Polytechnic University of Milan)

On the Quantum Boltzmann Equation: What is the Source, if any, of Irreversibility? 

In this talk, based on joint work with Jos Uffink, I provide a philosophical analysis of the Quantum Boltzmann Equation (QBE). The question addressed is: Given that the underlying microscopic dynamics is time-reversal invariant, what, if anything, is the source of irreversibility at the macroscopic level? Snoke at al. (2012) and Snoke (2020) claim that irreversibility stems from an assumption we call "Zero Off-diagonal Terms", from which one derives a factorization condition akin to the classical Stosszahlansatz. However, we contend that, contrary to the latter, whereby factorization is assumed only for particles before collision, in the quantum case factorization holds for all pre- and post-scattering states, and thus this condition is neutral with respect to the direction of time. As such, it cannot be responsible for irreversibility. Since there is no other time asymmetric ingredient in the derivation of QBE, we argue that the equation is fully time-reversal invariant. 

---

David Wallace (University of Pittsburgh)

What Gibbsian Statistical Mechanics Actually Says: Defending Bare Probabilism 

I expound and defend the "bare probabilism" reading of Gibbsian (i.e. mainstream) statistical mechanics, responding to Frigg and Werndl's recent (BJPS 72 (2021), 105-129) plea: "can somebody please say what Gibbsian statistical mechanics says?" 

---

Pascal Rodriguez Warnier (University of Western Ontario) 

Fluctuation Theorems and (Ir)reversibility in Statistical Thermodynamics

The fluctuation theorems are a family of results in statistical thermodynamics, which enable the derivation of equalities involving the probabilities of thermal quantities and hold irrespective of the nature of the process. I propose that Jarzynski’s equality—a fluctuation theorem that equates the expectation value of the exponentiated work with the exponential of the free energy difference between two states at the same temperature—offers an alternative method for deriving equations of state without invoking reversibility.

In phenomenological and statistical thermodynamics, the derivation of relationships between state variables often involves imagining processes that come arbitrarily close to reversibility. For example, entropy is defined through reversible heat exchanges and shown to be a state function of conserved quantities like internal energy. Such a dependence can be exploited to derive further equations of state.

By taking fluctuations seriously, the alternative method discussed in this talk enables the derivation of equations of state by considering measurements on large sets of processes that are arbitrarily far from thermodynamic equilibrium.  In the limit of increasingly large numbers of processes and measurements, macroscopic reversibility is recovered statistically.

---

Charlotte Werndl (University of Salzburg) - by Zoom

The Boltzmann Equation and its Place in the Edifice of Statistical Mechanics 

It is customary to classify approaches in statistical mechanics (SM) as belonging either to Boltzmanninan SM (BSM) or Gibbsian SM (GSM). It is, however, unclear how the Boltzmann equation (BE) fits into either of these approaches. To discuss the relation between BE and BSM, we first present a version of BSM that differs from standard presentation in that it uses local field variables to individuate macro-states, and we then show that BE is a special case of BSM thus understood. To discuss the relation between BE and GSM, we focus on the BBGKY hierarchy and note the version of the BE that follows from the hierarchy is "Gibbsian" only in the minimal sense that it operates with an invariant measure on the state space of the full system.