te Vrugt, M., Needham, P., & Schmitz, G. J. (2022) Is thermodynamics fundamental? arXiv:2204.04352v1 [physics.hist-ph] 9 Apr 2022 doi:10.48550/arXiv.2204.04352
[Abstract]It is a common view in philosophy of physics that thermodynamics is a non-fundamental theory. This is motivated in particular by the fact that thermodynamics is considered to be a paradigmatic example for a theory that can be reduced to another one, namely statistical mechanics. For instance, the statement "temperature is mean molecular kinetic energy" has become a textbook example for a successful reduction, despite the fact that this statement is not correct for a large variety of systems. In this article, we defend the view that thermodynamics is a fundamental theory, a position that we justify based on four case studies from recent physical research. We explain how entropic gravity (1) and black hole thermodynamics (2) can serve as case studies for the multiple realizability problem which blocks the reduction of thermodynamics. Moreover, we discuss the problem of the reducibility of phase transitions and argue that bifurcation theory (3) allows the modelling of "phase transitions" on a thermodynamic level even in finite systems. It is also shown that the derivation of irreversible transport equations in the Mori-Zwanzig formalism (4) does not, despite recent claims to the contrary, constitute a reduction of thermodynamics to quantum mechanics. Finally, we briefly discuss some arguments against the fundamentality of thermodynamics that are not based on reduction.
[Citing Place (1956)]  
Citing Place (1956) in context (citations start with an asterisk *):
Section 4 Reduction and multiple realizability
* We now turn to a particularly interesting issue which may pose a problem for reduction, namely multiple realizability. The idea of multiple realizability as a problem for reduction first appeared as an argument in the philosophy of mind in opposition to the view (defended, e.g., by Place (1956) and Smart (1959)) that every mental kind is identical with a neural kind, so that pain is identical with a certain brain state. Against this, Putnam (1967) has argued that many different animals—humans, reptiles, birds, ... — all seem to experience pain, but have quite different neuroanatomies, which makes it unlikely that they all exhibit the same neural kind. Moreover, according to the identity thesis, the identity between pain and neural state would be a natural law, so that anything feeling pain would exemplify this neural kind. But while it seems plausible that Martians or sentient robots could feel pain, their neuroanatomy would be very different from that of humans, which makes it implausible that they would exhibit the same neural kinds. There is a distinction here between multiple realizability (there could be different realizations) and multiple realization (there are different realizations). Pain in humans and reptiles is an example of multiple realization, whereas pain in humans and Martians is an example of multiple realizability.