In two recent papers Barry Loewer (2001, 2004) has suggested to interpret probabilities in statistical mechanics as Humean chances in David Lewis’ (1994) sense. I first give a precise formulation of this proposal, then raise two fundamental objections, and finally conclude that these can be overcome only at the price of interpreting these probabilities epistemically.
In two recent papers Barry Loewer (2001, 2004) has suggested to interpret probabilities in statistical mechanics as Humean chances in David Lewis’ (1994) sense. I first give a precise formulation of this proposal, then raise two fundamental objections, and finally conclude that these can be overcome only at the price of interpreting these probabilities epistemically.
Some of our best physical theories, notably quantum mechanics and statistical physics,
are probabilistic. A question thus is what the probabilities from the theories mean.
Since the theories provide a very successful account of the world, it is attractive to say
that their probabilities mirror real-world (“ontic”) probabilities independent of human
observers. The aim of this talk is to assess a view of such probabilities that has lately
attracted much attention under the label “Humean chances”.
The talk first clarifies what a philosophical view, or interpretation, of probabilities is
supposed to achieve and lists the most important views of ontic probabilities. I then
explain the Humean account as championed by David Lewis and Barry Loewer, among
others, and discuss it critically in view of applications to physics, in particular to statistical
physics. The main question is whether the account succeeds as a view about ontic
probabilities. The talk is also meant to give an in-depth introduction into important
issues in the metaphysics of probability.
Abstract Many theories and models from physics are probabilistic. This observation raises several philosophical questions: What are probabilities in physics? Do they reflect objective chances which exist independently of the human mind? Or do they only express subjective credences and thus capture our own uncertainty about the world? Finally, which metaphysical lessons, if at all, can one draw from the largely probabilistic character of physics? The chapters collected in this volume address these questions and provide a detailed philosophical appraisal of the status of probabilities in all of physics. Particular emphasis is laid upon statistical physics and quantum mechanics. Many chapters reflect a desire to understand probabilities from physics as objective chances. These chances are characterized, e.g., as time-averages, as probabilities from a best system in the terms of David Lewis, or using the Boltzmannian typicality approach. Other chapters are sympathetic to a Bayesian view of probabilities in physics. The chapters about quantum mechanics elucidate the peculiar characteristics of quantum correlations and discuss strategies to justify the Born Rule. Finally, the chapters of this volume demonstrate how closely interpretive issues about probabilities are entangled with other foundational problems of physics such as the Reversibility Paradox, the ontology of the quantum world and the question whether the world is deterministic.
This paper addresses the question of how we should regard the probability distributions introduced into statistical mechanics. It will be argued that it is problematic to take them either as purely ontic, or purely epistemic. I will propose a third alternative: they are almost objective probabilities, or epistemic chances. The definition of such probabilities involves an interweaving of epistemic and physical considerations, and thus they cannot be classified as either purely epistemic or purely ontic. This conception, it will be argued, resolves some of the puzzles associated with statistical mechanical probabilities: it explains how probabilistic posits introduced on the basis of incomplete knowledge can yield testable predictions, and it also bypasses the problem of disastrous retrodictions, that is, the fact the standard equilibrium measures yield high probability of the system being in equilibrium in the recent past, even when we know otherwise. As the problem does not arise on the conception of probabilities considered here, there is no need to invoke a Past Hypothesis as a special posit to avoid it.
This chapter deals with the question of how we should think of the probabilities introduced in statistical mechanics. It is usually said that these probabilities are introduced on the basis of our ignorance of the precise state of the system, or our incapacity to treat analytically the equations of motion in order to deliver a detailed account of the evolution of the system, or both. This suggests an epistemic reading. However, we make predictions on the basis of them, that are verified by experiment. This suggests that they be thought of as objective chances. It looks as if we have to employ both the epistemic and the objective concept, in an inconsistent way. Moreover, the standard equilibrium probabilities, applied to a system that has relaxed to equilibrium while isolated from its environment, cannot be either credences or chances. It is argued that the hybrid concept, epistemic chance, resolves these puzzles and is suited to play the role required of probability in statistical mechanics.
Abstract Some of the most compelling examples of the existence of truly objective probabilities come from physics, in particular quantum physics and statistical mechanics. So it is crucial to the overall success of HOC that it be compatible with the objective probabilities found in these theories. First, objective probabilities in classical (Boltzmannian) statistical mechanics (SM) are discussed. It is shown that HOC does capture the central probabilistic postulates of SM, and indeed that it may do so in two distinct ways. Second, objective probabilities in standard, non-relativistic quantum mechanics (QM), the context in which the notion that fundamental physics is at bottom chancy first became widely accepted, are discussed. It is shown that HOC is especially apt for capturing the probabilities of QM; other accounts may do equally well (though some clearly do not), but none can do the job better.
On a very intuitive way of thinking, if it is already determined that some event will happen, then there is no non-trivial chance of it failing to happen, and if it is already determined that some event will not happen, then there is no non-trivial chance of it happening. A different sort of worry that arises for an actual frequentist analysis of deterministic chance is that such an analysis seems unsuited to theories like Boltzmannian statistical mechanics or Bohmian mechanics insofar as those theories are supposed to describe the evolution of the Universe as a whole. It should be clear, then, that there are options available for understanding deterministic chance that are in keeping with several of the leading metaphysical analyses of chance. It is widely accepted that whatever chances are, they are to be contrasted with individual credences . According to classical statistical mechanics, the fundamental laws are Newtonian and thus deterministic.