An apple hangs from a tree. The stem breaks, and the apple falls. Before Newton, one could have treated this as a brute fact: unsupported apples fall. Newton’s achievement was not merely to rename the event “gravity.” He introduced a quantitative cause that connected the falling apple to the Moon’s orbit, the tides and planetary motion. The same law predicted how the motion would change if the mass, distance or initial conditions changed.
That success suggests an uncomfortable question about time. Physics describes with extraordinary precision how a system’s state differs from one moment to the next. But what makes the change actually occur? Does time only label and measure change, while forces, fields and geometry specify its form? Or is there a deeper physical process by which one state gives way to another and possibilities become facts?
The question is easy to dismiss as metaphysical. That may be premature. “Objects simply fall” would also once have sounded like a reasonable stopping point. The scientific challenge is to formulate the question so that a proposed answer produces an observable contrast.
What ordinary dynamics explains
In classical mechanics, the state of a system can be represented by positions and momenta, collectively denoted by X. A dynamical law determines its rate of change:
dX/dt = F(X,t)
Given an initial state and the relevant forces, the equation determines a trajectory X(t). Newtonian gravity, electromagnetism and fluid mechanics all use this architecture. The law tells us which histories are physically allowed and which trajectory follows from specified initial conditions.
But the symbol t is not normally a substance pushing the system forward. It orders events and quantifies their separation. Nor is the function F literally a machine sitting outside the world and manufacturing the next state. Mathematically, it defines a relation between neighbouring parts of a solution.
This leaves two questions that are often conflated:
- Why does the apple follow this trajectory rather than another? Gravity answers this.
- Why is the trajectory concretely traversed—why is the apple first here and then there? Standard mechanics generally takes this occurrence for granted.
The first question is clearly physical. Whether the second can also become physical is the issue.
Relativity: time becomes geometry
General relativity deepens the problem. Gravity is no longer an ordinary force acting in a fixed arena. Matter and energy shape spacetime, while spacetime constrains the motion of matter. A freely falling apple follows a geodesic in a four-dimensional geometry.
This explains more than Newtonian gravity: why all freely falling bodies respond universally, why clocks tick at different rates in different gravitational potentials and why light bends. Yet the theory is commonly expressed as a complete spacetime history. The apple’s worldline is part of that geometry. The equations determine its shape, but do not require an external mechanism that moves the apple along it.
Introducing such a mechanism creates an immediate regress. If ordinary time t advances relative to a second parameter τ, what makes τ advance? A third time might be introduced, followed by a fourth, without adding observable content. In many cases, the extra parameter is merely a different labelling of the same trajectory:
dX/dτ = (dX/dt)(dt/dτ)
This is an important warning, but not a proof that a deeper mechanism cannot exist. It says that a new theory must add more than a hidden universal clock. It must predict something that cannot be removed by redefining the time coordinate.
Quantum theory: evolution and actualisation separate
Quantum mechanics makes the distinction sharper. The Schrödinger equation evolves a state smoothly and deterministically:
iℏ ∂|ψ⟩/∂t = H|ψ⟩
The Hamiltonian H is called the generator of time evolution. This means that it mathematically determines how amplitudes and phases change. It does not by itself explain why a measurement produces one definite result rather than a superposition of alternatives.
Different interpretations respond in different ways. Everettian quantum mechanics retains every decohered branch. Bohmian mechanics adds a definite configuration guided by the wavefunction. Operational interpretations treat outcomes as facts relative to an interaction or observer. None of these approaches has yet won decisive experimental support over the others.
Objective-collapse theories such as GRW and Continuous Spontaneous Localisation take a more radical step. They modify quantum dynamics with stochastic and nonlinear terms so that localisation is a real physical event rather than an update of knowledge. Such models can predict small losses of interference, random diffusion, heating or spontaneous radiation. They therefore turn “a possibility becomes a fact” into a potentially falsifiable physical hypothesis. They are not merely interpretations. A comprehensive review describes both their theoretical construction and their experimental tests in detail (Bassi et al., Reviews of Modern Physics).
Yet collapse models still normally assume an external time parameter. They may explain the production of a definite event without explaining why time itself passes.
Can time emerge from relations?
In general relativity there is no universal external clock for the Universe as a whole. In some approaches to quantum gravity, the fundamental equation is effectively timeless. This motivates relational accounts: one subsystem serves as a clock for another. Instead of saying “the system has state S at external time t,” we say “when clock C has reading t, system S has this state.”
The Page–Wootters construction shows how a globally stationary quantum state can contain internal correlations that an embedded observer describes as evolution. This is sometimes called “evolution without evolution.” It may explain temporal order and dynamics without an external clock. But it does not necessarily explain an objective moving present or why one correlated event is actualised after another.
Thermodynamics adds the arrow. Memories, records and irreversible processes accumulate in the direction of increasing entropy. That helps explain why the past leaves traces while the future does not. It may explain the asymmetry of our experience of time, but an entropy gradient is not automatically a mechanism that makes events occur.
A world built from events
Causal-set theory comes closer to treating becoming as physical. It proposes that spacetime is fundamentally a discrete, partially ordered set of events. In sequential-growth models, a causal set is enlarged one event at a time according to stochastic transition rules. Reality is not placed in a pre-existing continuum; spacetime geometry is expected to emerge from the accumulated causal order.
The difficult part is preserving relativity. The numerical order in which independent events are added cannot be a preferred universal time. Two spacelike-separated events may be inserted in either order, and the final physical structure must be equivalent. The classical sequential-growth programme derives families of stochastic dynamics under causality and discrete-covariance requirements (Rideout and Sorkin).
This is where higher category theory may become useful. States can be represented as objects, physical processes as morphisms, equivalences between different constructions as 2-morphisms, and higher consistency conditions as further morphisms. An infinite hierarchy of “times that move other times” can then be replaced by a hierarchy of transformations and coherence relations. Causal categories already encode processes together with signalling constraints (Coecke and Lal), while higher process theories study transformations of transformations.
This does not itself make anything happen. Category theory organises possible and composable processes; it does not select the actual one. But it may provide the right language for a local, partially ordered becoming without a global clock or arbitrary construction order.
Where physics touches metaphysics
Every proposed explanation eventually faces a choice. It can explain becoming through another process, in which case we can ask what makes that process occur. It can declare becoming fundamental. Or it can deny that there is an objective passage at all and treat the Universe as a complete four-dimensional history.
This resembles the “hard problem” of consciousness. Neuroscience can describe discrimination, memory, attention, integration and verbal report. A critic can still ask why those structures are accompanied by lived experience. Likewise, physics can describe temporal order, causal dependence, entropy and clock readings while leaving open why there is an experienced or objective occurrence.
But declaring the remainder “metaphysical” risks converting a limitation of our present conceptual tools into a limitation of possible understanding. The more productive response is to ask what a deeper theory would have to change in observation.
How to escape the metaphysical trap
Suppose a new field B controlled a local “rate of becoming”:
dX/dτ = B(x)F(X)
If B multiplies every physical process equally—every atomic transition, chemical reaction, decay, computer and brain—then it is unobservable from inside the region. It can be absorbed into the definition dt = B dτ. This would be coordinate freedom, not new physics.
A genuine discovery requires non-universal or noisy effects: different clocks responding differently, unexplained correlated decoherence, a mass-dependent production of definite events, fluctuations of causal order or a deviation from standard quantum statistics. The key observable is not an absolute speed of time but a residual ratio between different processes after all known relativistic and environmental effects are removed.
What observations currently say
So far, the evidence is mainly constraining rather than positive.
Comparisons between different atomic clocks show exceptional stability and place tight limits on changes in fundamental constants. Laboratory clocks constrain relative drift of the fine-structure constant to roughly the level of 10−18 per year, depending on the model and dataset (Murphy et al., Science). Global clock networks have also searched for correlated transient signals from hypothetical dark-matter fields and found no convincing detection, instead improving limits on possible couplings (global optical-clock network study).
Claims that radioactive decay rates vary annually or with solar activity have not produced a reproducible, isotope-independent signal. Controlled tests generally find no significant deviation once environmental and instrumental effects are considered. The persistent discrepancy between “beam” and “bottle” measurements of the neutron lifetime is a credible experimental puzzle, but it compares different apparatuses and is more plausibly due to systematics or particle physics than to a variable rate of time.
Mechanical resonators offer a more intriguing boundary. Ultra-cold cantilever experiments have observed excess force noise and used it to constrain CSL-type collapse mechanisms. Some measured noise was initially compatible with an interesting collapse-model parameter range, but no independent, universal signal has emerged. The field therefore reports upper bounds rather than a detection.
Matter-wave interference provides the complementary result: increasingly massive molecules still display coherent superposition. Interference beyond 25,000 atomic mass units strongly constrains any mechanism that destroys superpositions simply because a system is “large” (Fein et al., Nature Physics).
Perhaps the most promising frontier is the quantum clock. Relativity says that different paths accumulate different proper times; quantum theory allows a clock to occupy a superposition of paths. Theory predicts quantum corrections that distinguish a coherent superposition from a classical mixture (Smith and Ahmadi, Nature Communications). Experiments able to combine precision clocks, interferometry and mesoscopic masses could expose a failure of either standard quantum dynamics or classical spacetime.
A concrete experimental programme
A serious search should not monitor a single clock and call every drift mysterious. It should compare several co-located processes with different physical sensitivities:
- an optical electronic transition;
- a hyperfine or nuclear transition;
- a radioactive or metastable decay;
- a cryogenic mechanical resonator;
- a mesoscopic quantum superposition;
- and, where possible, spatially separated copies linked by precision time transfer.
The experiment would search for residuals that are correlated across instruments but weighted differently according to a proposed coupling. It should vary gravitational potential, orientation, velocity, mass distribution and degree of quantum coherence. A credible signal would need to survive replication with different technologies and predict new results before they were measured.
Such a programme would also have practical consequences. A controllable differential coupling to temporal evolution could improve clocks and navigation, reveal new decoherence-control methods, create unusual quantum sensors or enable operations based on non-classical causal order. A perfectly universal slowing of every process would be technologically useless because no internal observer could detect it. Applications require a gradient, selective coupling, boundary or exploitable fluctuation.
The question worth keeping open
Physics currently explains the structure of temporal evolution with astonishing success. Classical forces determine trajectories, relativity turns duration and gravity into geometry, quantum theory evolves amplitudes, thermodynamics establishes an arrow, and relational approaches reconstruct time from correlations. Objective-collapse models and causal-set growth go further by trying to make the creation of facts or events part of the dynamics.
None has yet demonstrated what makes the next state become concretely real. That may ultimately be a badly framed demand. It may also be the sign of missing physics.
The lesson from the apple is not that every intuitive question deserves a new force. It is that a question leaves metaphysics and enters physics when it is converted into a law that unifies phenomena, predicts counterfactual differences and survives experiment. Instead of asking vaguely what “pushes time,” we can ask whether the production of events has a measurable dynamics of its own.
For now, the answer is no confirmed detection. But precision clocks, mesoscopic interferometers, cryogenic resonators and tests of objective collapse are beginning to make the question experimentally accessible. That is reason enough not to close it.