Home > Positioning > Persons > Albert

David Albert (1954–)

Albert is a philosopher of physics whose work addresses two problems: why time has a direction, and what happens when a quantum measurement is made. His most influential contribution is the formulation of the Past Hypothesis — the postulate that the universe began in a state of extraordinarily low entropy — as an explicit foundational element in statistical mechanics. Boltzmann’s framework explains why entropy increases given a low-entropy starting point; Penrose quantified the improbability of that starting point; Albert named it and argued for its role as a fundamental postulate on which the reliability of all records, inferences, and sciences of the past depends. His work on quantum mechanics — particularly the measurement problem and the case for dynamical-collapse theories — is pursued in parallel, and the two programmes connect: the arrow of time and the interpretation of quantum mechanics are, in Albert’s view, not separate problems.

Life

Born 1954. Undergraduate and doctoral training in theoretical physics at Rockefeller University, New York. The move from physics to philosophy of physics was driven by the foundational questions that physics itself raised but did not address — what the formalism of quantum mechanics says about the world, and what grounds the asymmetry of time. Professor of philosophy at Columbia University since the early 1990s; Frederick E. Woodbridge Professor of Philosophy. Columbia under Albert has been a centre for the philosophy of physics, particularly the foundations of statistical mechanics and quantum mechanics; his teaching and supervision have shaped a generation of work in the field.

The most sustained collaboration is with Barry Loewer, a philosopher at Rutgers. The mentaculus — the programme for grounding all special-science regularities in the Past Hypothesis, the Statistical Postulate, and the dynamics — is their joint project, developed across papers and conversations over three decades. Much of the recent argument about how statistical mechanics underwrites the reliability of the special sciences is jointly authored or jointly conceived.

Albert writes in a style unusual in the philosophy of physics — direct, conversational, often combative, impatient with evasion. Quantum Mechanics and Experience reads more like a sustained argument in a seminar room than a textbook. The directness is deliberate: Albert treats clarity about what a theory actually says as a precondition for taking it seriously.

The Past Hypothesis and the foundations of statistical mechanics

Time and Chance (2000) is Albert’s central work on the arrow of time. The argument begins from a familiar observation: Boltzmann’s statistical mechanics explains why entropy increases — there are overwhelmingly more high-entropy configurations than low-entropy ones, so a system migrates toward what is statistically dominant. But the reasoning is time-symmetric: it applies equally well in both temporal directions. Run the statistics backward and they predict that the system was at higher entropy in the past — which is false. The early universe was at lower entropy than the present. The statistical mechanics, taken alone, makes the wrong retrodiction.

Albert’s response: the correct application of statistical mechanics requires a supplementary postulate — the Past Hypothesis. The universe began in a macrostate of extraordinarily low entropy. This postulate is not derivable from the dynamics; it must be added as a boundary condition. Combined with the dynamics (whether classical or quantum) and a probability distribution over the microstates compatible with the initial macrostate (the Statistical Postulate — roughly, that the actual microstate was typical of those compatible with the macro-condition), the Past Hypothesis grounds the asymmetry of time. Entropy increases toward the future because the Past Hypothesis constrains the past; the statistics are free to run in the future direction but not in the past direction, because the past is pinned.

The scope of the claim is broad. Reichenbach had shown, in The Direction of Time (1956), how records form: a subsystem interacts with a larger system, acquires a low-entropy imprint, and then evolves in quasi-isolation — a branch system. Albert takes the next step: what makes all such records reliable is the Past Hypothesis. A record — a footprint, a photograph, a memory — is informative about the past only because the recording system was in an initialised ready state before the interaction. That initialisation, in turn, was possible only because of the low-entropy past. Without the Past Hypothesis, there is no ground for trusting any inference about the past — including the inferences of geology, evolutionary biology, and cosmology. The arrow of time is not one topic among others in physics; it is the condition for the reliability of every empirical science.

The mentaculus. Albert and Barry Loewer coined this term (from Tristram Shandy) for the complete probability map of the world: the Past Hypothesis, the Statistical Postulate, and the dynamical laws together. The mentaculus is supposed to ground not only thermodynamics but all the special-science regularities — all the patterns that chemistry, biology, and the social sciences rely on. Whether this is too ambitious — whether a single probability distribution over initial microstates can really ground the regularities of economics or psychology — is debated. The claim is that it can in principle, because those regularities ultimately depend on the thermodynamic asymmetry, which the mentaculus grounds.

Quantum mechanics and the measurement problem

Quantum Mechanics and Experience (1992) — both a textbook and a philosophical argument — presents the measurement problem as the central issue in the interpretation of quantum mechanics. The problem: quantum mechanics describes systems as evolving in superpositions (the Schrödinger equation is linear and deterministic), but measurements always find definite outcomes. Standard quantum mechanics handles this with the collapse postulate — upon measurement, the superposition collapses to one outcome. But what counts as a measurement? The theory does not say. If the measuring apparatus is itself a quantum system (and it is), then the apparatus-plus-system should evolve into a superposition, not collapse. The measurement problem is the question of how definite outcomes arise from a theory whose dynamics do not produce them.

Albert’s preferred resolution is the GRW theory (Ghirardi, Rimini, Weber, 1986): a dynamical-collapse theory in which the Schrödinger equation is modified so that superpositions spontaneously and randomly collapse at a rate that is negligible for individual particles but effectively instantaneous for macroscopic objects. The collapse is not triggered by measurement or observation; it is a feature of the dynamics. GRW is a precise, observer-independent theory — it says exactly what happens and when — and Albert has argued at length that this precision is what the foundations of quantum mechanics require. The theory is empirically adequate but has not been experimentally distinguished from standard quantum mechanics; the collapse rate is a free parameter constrained only by upper bounds.

Albert’s critique of the Everett (many-worlds) interpretation is sustained and specific. He argues that Everettian quantum mechanics faces a probability problem: if every outcome of every measurement actually occurs (in different branches), it is unclear what it means to say that one outcome is more probable than another. Everettians have proposed solutions (decision-theoretic arguments, self-locating uncertainty); Albert has contested them in each case. The debate is unresolved.

In After Physics (2015), Albert argues that the two programmes are not independent — they are aspects of the same problem. The Past Hypothesis is a claim about the initial quantum state of the universe. What that state is, what it means for it to be “low-entropy,” and how it evolves all depend on which interpretation of quantum mechanics is correct. A GRW universe and an Everettian universe have different accounts of what the Past Hypothesis is a hypothesis about. The foundations of statistical mechanics cannot be settled without settling the foundations of quantum mechanics, and vice versa. Albert’s insistence on this entanglement — that you cannot do the philosophy of one without doing the philosophy of the other — is the central argument of the book.

Where Albert stops

Albert is explicit about the status of the Past Hypothesis: it is a postulate, not a derivation. It says that the universe began in a low-entropy macrostate. It does not explain why. Albert argues that the postulate is necessary — that statistical mechanics does not work without it — but he does not claim to have explained the boundary condition. Whether the Past Hypothesis can be derived from something deeper is an open question. Penrose’s conformal cyclic cosmology, Carroll’s baby-universe scenarios, and various inflationary models each attempt to make the low-entropy beginning follow from more fundamental principles. None has succeeded to general satisfaction. Albert’s position is that until such a derivation exists, the Past Hypothesis must be accepted as a fundamental postulate on the same footing as the dynamical laws — a position some physicists find unsatisfying, since it elevates a contingent fact about the initial state to the status of a law.

The mentaculus — the claim that the Past Hypothesis, the Statistical Postulate, and the dynamics together ground all special-science regularities — is the most ambitious element of the programme. The Statistical Postulate assumes that the actual initial microstate is typical of those compatible with the initial macrostate. Whether this assumption of typicality is justified — and what “typicality” means in a deterministic universe where the initial microstate is whatever it is — is a contested point in the foundations of statistical mechanics. If the mentaculus works, it provides a unified foundation for the sciences; if the typicality assumption fails, the foundation has a gap at its base.

Albert’s preference for GRW over Everett is philosophically motivated and empirically underdetermined. Both interpretations are consistent with current experimental evidence. The choice between them may ultimately be settled by experiment — GRW predicts deviations from standard quantum mechanics at scales that future experiments might probe — but it has not been settled yet. Albert has been clear that the choice matters for the foundations: different interpretations give different accounts of what the world is like, and the foundations of statistical mechanics may depend on which account is correct. Whether this dependence is real or whether the arrow-of-time results are interpretation-neutral is itself debated.

Key works

See also: Boltzmann · Reichenbach · Penrose