Resetting Time to Its Proper Place in Physics and Beyond

Time Matters. But Matter matters more.

Author's note: This essay grew out of an extended conversation between me and OpenAI's ChatGPT (GPT-5.6 Sol). The underlying hypothesis and line of inquiry were developed iteratively through that dialogue, with ChatGPT contributing scientific context, critical examination, research, structure and editorial development. The final argument and publication are the author's responsibility.

Resetting Time to Its Proper Place in Physics

Time matters. But matter matters more.

Physics tells us that time slows down near massive objects, that velocity causes time dilation, that different observers can accumulate different amounts of proper time, and that matter follows worldlines through spacetime. These statements have precise meanings within relativity, and the theory predicts the corresponding observations with extraordinary accuracy. Yet the language can also encourage a surprisingly misleading mental picture of what the underlying experiments actually show.

Grammatically, time behaves like an ordinary noun. Time passes, time slows down, gravity affects time, we move through time, time stops, and we travel through spacetime. These sentences make it remarkably easy to imagine time as a physical participant: gravity acts upon time, time consequently runs more slowly, and slower time then causes atoms, clocks and people to change more slowly. Likewise, spacetime can become a four-dimensional environment through which matter seems to travel, making yesterday and tomorrow sound a little like places. Amsterdam exists at one set of spatial coordinates, so perhaps Amsterdam in 1650 exists somewhere else along a temporal coordinate.

None of that is required to use relativity successfully. Strip the temporal terminology away for a moment and look at what we physically do. We take two physical systems, compare them, let them follow different trajectories through different physical environments, and compare them again where possible. Their states may now differ. Atomic clocks may have accumulated different numbers of transitions, other suitable clocks show corresponding differences, and under sufficiently extreme relativistic conditions even biological ageing can diverge.

General relativity predicts these relationships with extraordinary accuracy, but what we physically observe is not “time itself” acting on anything. We compare the history of change in one physical system with the history of change in another. That distinction is the subject of this essay. Before saying what time did, it is often worth asking a more concrete set of questions: which physical systems are being compared, which state changes were measured, what were they measured against, and what physical histories produced the difference?

We measure change with change

Every clock humanity has ever constructed is a changing physical system. A pendulum changes position, a quartz crystal oscillates, and an atomic clock uses a reproducible atomic process. Even observing the clock requires further physical change: photons reach a detector, electrical states change, information is stored and neurons respond.

The official definition of the second makes this wonderfully explicit. The SI second is defined by fixing the frequency associated with a particular transition of caesium-133 at exactly 9,192,631,770 hertz. Before atomic standards, our temporal units ultimately depended on astronomical changes such as the rotation of the Earth. Our clocks improved because we found physical processes whose changes provided more stable references. The reference process changed, but the underlying principle did not.

Suppose I want to determine how long some experiment takes. I place a clock beside it. The experiment is one changing physical system and the clock is another. During a particular sequence of state changes in the experiment, the reference system completes a particular number of reproducible changes. Calling the result “five seconds” is extraordinarily convenient, but underneath the unit we are still comparing one physical history with another.

This suggests a slightly different starting point from the familiar statement that “time measures change.” At the empirical level, physical systems undergo sequences of state change, and we compare those histories with other physical histories. Time is the abstraction physics uses to express their ordering and relationships.

The matter-box experiment

Imagine two, as far as physically possible, identical sealed boxes, A and B. Each contains identical atoms, identical atomic clocks, identical chemical processes and, if we want to make the thought experiment more vivid, an identical human observer. We begin with the boxes together and establish corresponding initial states. We synchronize the processes we intend to compare and call this initial configuration (S_0).

Then we separate them. Box A follows one route through the universe while Box B follows another. They may travel at different velocities, accelerate differently, pass at different distances from massive bodies, or experience different gravitational environments. One could remain near Earth's surface while another travels far from Earth; in a more extreme version, one could approach a black hole.

Where physically possible, we eventually bring them together again and compare them. Their states are no longer necessarily equivalent. Their clocks may disagree, corresponding physical processes may have accumulated different numbers of transitions, and if the relativistic difference is sufficiently large the observers themselves may have aged differently.

We should be careful here with the phrase “amount of change.” There is no obvious universal scalar called change. A system can move enormous distances while changing little chemically, or undergo trillions of electromagnetic oscillations without much macroscopic displacement. What we can do is select corresponding reproducible processes in the two boxes and compare their histories. The remarkable discovery underlying relativity is that these comparisons exhibit a universal structure, and relativity captures that structure mathematically.

Putting proper time back into the experiment

Within relativity, proper time has a precise definition. A physical system follows what the theory calls a worldline through spacetime, and the spacetime metric assigns an invariant quantity along that trajectory. The potential confusion begins when we reverse the explanatory direction.

It is tempting to say that Box A accumulated less proper time than Box B and therefore the physical processes in A advanced less far. Within the formalism, that is perfectly ordinary language. Operationally, however, how did we establish the proper-time difference? We compared physical processes. There is no second device hidden behind the atomic clock that independently measures a substance called proper time and tells us whether the atoms obeyed it correctly. The clock itself is a physical system undergoing reproducible state changes.

A state-first description therefore runs in the opposite direction. Boxes A and B follow different physical histories. When we compare corresponding processes, they have accumulated different numbers of state transitions. Those differences display a universal relationship. General relativity represents that relationship geometrically, and the invariant scale along each history is what we call proper time.

In this sense, proper time is not an additional mechanism making matter behave differently. It is the geometric measure relativity assigns along a physical history, physically realized by an ideal clock travelling with that history. The distinction does not change the predictions of relativity, but it does change the conceptual hierarchy through which we understand them.

Gravitational time dilation is useful shorthand

The same issue becomes especially visible in the phrase gravitational time dilation. We commonly say that gravity makes time run slower. It is compact, memorable and mathematically meaningful, but unpack the statement experimentally and ask what physical systems were compared, which processes changed, and what they were measured against.

Take two sufficiently accurate clocks, compare them initially, let them follow histories at different gravitational potentials, and then compare their physical processes according to an appropriate procedure. Their readings diverge in precisely the systematic manner predicted by general relativity. Modern optical clocks can detect extraordinarily small differences associated with elevation, which makes this one of the most striking examples of relativity becoming directly measurable in ordinary terrestrial conditions.

The phenomenon is real, but “gravity slowed time” is already a compressed description of that phenomenon. What the experiment directly gives us is a relationship between changing physical systems that followed different histories relative to the surrounding distribution of mass-energy.

There is an essential qualification. We should not replace “gravity slows time” with the equally misleading idea that gravity merely interferes with the mechanism of a caesium clock. If that were all that happened, replacing caesium with an unaffected process should make the effect disappear. That is not what relativity predicts. Suitable clocks based on different physical processes exhibit the same relativistic relationship. This universality is precisely why proper time and spacetime geometry are such powerful abstractions.

So the complete operational description is cumbersome, which is why terms such as gravitational time dilation exist. There is nothing wrong with shorthand; the problem begins when we forget that it is shorthand.

Location matters, coordinates do not cause anything

Our matter boxes also make another distinction important. Coordinates are labels. If I describe exactly the same physical situation using another coordinate system, I have not changed the experiment. The numbers themselves exert no influence on the boxes.

But physical location absolutely matters. A box on Earth's surface occupies a different physical relationship to Earth's mass-energy distribution from a box far away. A box travelling close to a neutron star or black hole follows a dramatically different physical history from one travelling through comparatively weak gravitational conditions. Likewise, the route matters. Two boxes can begin together and eventually reunite while following radically different trajectories in between.

So saying that coordinates do not matter physically can itself become misleading. Arbitrary coordinate labels do not matter, but the physical relationships represented by location and trajectory certainly do. This is another place where language can hide an important distinction between the representation and what is represented.

A black hole is an extreme test

A black hole makes the distinction vivid. Imagine Box A remaining far from a black hole while Box B travels close to it and later returns, assuming we choose a trajectory for which returning is possible. When reunited, corresponding physical processes in the two boxes need not have accumulated identical histories.

The usual shorthand says that B experienced gravitational time dilation. Our alternative description simply delays introducing that terminology. The boxes began in corresponding states, followed different physical histories through dramatically different gravitational conditions, and returned with systematically different states. General relativity predicts the relationship.

Now allow B to cross the event horizon. The event horizon is not merely an arbitrary coordinate line. Particular coordinate systems can behave badly there, but changing coordinates does not remove the causal distinction associated with being inside the horizon. Yet the falling box does not locally encounter a sign saying “TIME STOPS HERE.” For a sufficiently large black hole, its atoms continue changing normally as it crosses the horizon, its clock continues operating, chemical processes continue and its observer continues thinking.

A distant observer obtains a very different picture through the signals received from the falling box. This makes statements such as “time freezes at the event horizon” excellent examples of the problem discussed here. Without specifying whose clock, which signals and which comparison procedure we mean, an efficient piece of relativistic shorthand becomes an apparently causal statement about something called time. The more useful questions are concrete: what is happening to the falling physical system, what is happening to the distant system, what information can pass between them, and which changing processes are being compared?

Spacetime is a framework for relationships among histories

This brings us to perhaps the most powerful term of all: spacetime. Relativity demonstrates that spatial and temporal relationships cannot generally be separated in the Newtonian manner. Minkowski's geometrical reformulation of special relativity and Einstein's general relativity provide an extraordinarily successful framework for describing events, trajectories, gravity and causality.

Nothing in this essay requires rejecting that framework. But spacetime is also a word, and words carry intuitions. It is very easy to imagine a four-dimensional substance containing three familiar directions called space and another comparable direction called time. Matter then seems to move through this container, and past and future begin to sound like locations within it.

From there the science-fiction intuition almost writes itself. Amsterdam exists at one set of spatial coordinates, while Amsterdam in 1650 exists at another temporal coordinate. Perhaps a sufficiently advanced machine merely needs to find the correct route. The mathematics does not automatically require that intuitive ontology.

For the purposes of this essay, I find a more cautious description useful: spacetime is the geometric framework through which relativity represents relationships among physical events and histories. Our matter box follows a physical history, its state changes, other boxes follow other histories, signals propagate between them, and matter-energy relates to the gravitational geometry that constrains possible trajectories and causal relationships. Relativity describes all of this mathematically with extraordinary accuracy.

The map is extraordinarily good. That does not mean every intuitive property suggested by the language of the map must be projected back onto the territory.

The clock is another matter box

There is a useful consequence of looking at the problem this way: a clock stops being conceptually special. It is another physical system following its own history.

Suppose Box A contains the experiment I care about and Box C contains my reference clock. To measure the duration of some process in A, I correlate changes in A with reproducible changes in C. Now send C along another physical trajectory and the reference process itself acquires a different relationship to A.

This is exactly the kind of situation relativity teaches us how to handle. There is no privileged universal reference clock sitting outside the universe. Every actual clock is itself a physical system with a physical history. Proper time can therefore usefully be thought of as the local invariant scale along a particular timelike history, while an ideal clock physically realizes that scale through its own state changes.

That observation makes the success of relativity more impressive rather than less. Despite every clock being another participant in physical reality, suitable clocks exhibit relationships that can be captured by a common geometry. Viewed from this direction, relativity describes relationships among histories of physical processes, and temporal quantities provide an extraordinarily efficient language for expressing those relationships.

The observer is another changing system

Observation itself cannot escape this structure. For one physical system to observe another, something must interact. A photon reaches a retina, an electrical state changes in a detector, a magnetic state stores a bit, neurons alter their firing patterns, and information becomes physically represented somewhere.

The observer therefore cannot stand outside the universe and inspect time independently. The object changes, the reference clock changes, the measuring apparatus changes and the observer changes. Our empirical description is constructed from correlations among those changes.

This does not require us to settle any deeper metaphysical question about time. The narrower point is enough: every measurement through which we assign temporal quantities is realized through physical processes and comparisons among physical histories. Relativity describes the structure of those comparisons extraordinarily well.

So what is “now”?

The same perspective changes the way I think about the present. Ordinary intuition imagines now as a universal boundary sweeping forward through reality. Relativity already undermines the universal part of that intuition because spatially separated events that one observer regards as simultaneous need not be simultaneous for another.

Our brains complicate the subjective version. Conscious experience is not an instantaneous snapshot. Signals propagate at finite speeds, sensory information is processed, and the brain integrates information over intervals before producing our experience of a coherent present.

From a state-oriented perspective, “now” can be treated more modestly as the current physical configuration from which a system can interact. The past survives in that configuration through physical traces such as memories, photographs, fossils, scars, documents, geological structures and radiation. The future appears differently, through predictions and representations of possible subsequent states.

My memory of yesterday exists physically now as a current configuration of my brain, while my prediction about tomorrow also exists physically now. Both are present states containing information about other states.

The arrow becomes an asymmetry of physical histories

We often say that time flows from past to future, but again it is useful to ask what we actually observe. Heat disperses, eggs break and almost never spontaneously reassemble, organisms age, records accumulate, causes leave traces in subsequent states, and our brains contain memories correlated with earlier conditions rather than memories of tomorrow.

There is an enormous asymmetry in physical histories. Thermodynamics describes a fundamental part of that asymmetry through entropy. Many microscopic laws possess time-reversal symmetries or closely related forms of reversibility, yet macroscopic state evolution displays an overwhelming directionality.

Rather than beginning by saying that an entity called time flows forward and everything else follows it, we can begin with the observed asymmetry of physical state evolution. That does not solve the arrow-of-time problem, but it places the explanatory burden somewhere more concrete: why do physical histories exhibit such a strong statistical asymmetry?

Travelling into the future becomes differential history

The matter-box model gives us a cleaner way to discuss time travel. Suppose I want to encounter an Earth that has undergone fifty years' worth of familiar physical change while undergoing as little change myself as possible. The conceptual requirement is straightforward: Earth and I must follow very different physical histories.

Perfect suspended animation provides an imaginary example. If my biological state could somehow be preserved almost perfectly while Earth continued changing, I could be revived into a world transformed by fifty years while my own physical state had changed comparatively little. We cannot currently do this to humans reversibly, but the thought experiment makes the relationship obvious.

Relativity provides a genuine physical route to differential ageing. Send our matter box along an appropriate high-velocity trajectory and later reunite it with Earth. Corresponding clocks and other physical processes can then show dramatically different accumulated histories.

We call this relativistic time dilation. In the vocabulary developed here, future-directed time travel is differential state evolution along different physical histories. Nothing needs to reverse.

Travelling into the past asks for something else

Now ask for the opposite: I want to travel back to 1976. The familiar time-travel picture encourages me to imagine 1976 as a destination that still exists somewhere along a temporal axis, just as Paris exists elsewhere in space. A sufficiently advanced machine merely needs to find the right route.

The state-oriented question is different: what physically made 1976 be 1976? It was a particular configuration of matter, radiation, fields, organisms, information and relationships. People had particular bodies and memories, buildings occupied particular states, photons propagated in particular directions, heat was distributed differently, records that exist today had not yet been created, and many people alive today did not yet exist.

If by returning to 1976 I mean literally restoring that earlier physical situation, those changes have to be undone. Trees must become younger, chemical reactions must reverse, radiation must return to previous configurations, heat that dispersed must reconcentrate, memories accumulated since then must disappear and physical records of subsequent events must be removed.

Then we encounter the traveller problem. Suppose the surrounding universe is somehow restored perfectly to its earlier configuration while I remain my present-day self. We have already failed to restore the original state, because the universe now contains something the original configuration did not contain: me, carrying physical information about events that, according to the reconstructed surroundings, have not happened.

We have created a new state resembling 1976. That is not the original 1976.

Reversing movement is not enough

My original intuition about backwards time travel was simpler: perhaps I could remain unchanged while everything else in the universe retraced its movements. That captures part of the idea, but movement is too narrow.

Physical reality includes radiation, fields, thermodynamic distributions, quantum states and correlations, and chaotic systems in which tiny differences can amplify enormously. Restoring an earlier physical configuration would therefore require much more than reversing the visible velocities of macroscopic objects. The relevant physical state and its relationships would have to retrace their history with extraordinary precision.

So the original intuition becomes more precise: literal return to the past, if by that we mean restoring the actual past, is a problem of state restoration rather than merely motion reversal. That is a fundamentally different requirement from arranging for two systems to accumulate different histories before meeting again.

A causal loop is not a restored past

General relativity introduces an important complication. Some mathematical solutions contain closed timelike curves: trajectories through spacetime that loop into their own causal past. These are often described as examples of general relativity permitting time travel, although whether physically realizable versions can exist is a much deeper question involving global spacetime structure, unusual physical conditions and possible quantum constraints.

Even granting such a curve for the thought experiment, our distinction still matters. Put the matter box on the curve. Inside the box, its physical processes need not reverse; its atoms continue changing, its clock continues operating and its observer can continue forming memories.

The box could therefore encounter an earlier part of the surrounding causal history as a later physical state of itself. That is not state restoration. The surrounding environment may correspond to an earlier portion of its history, but it now interacts with a system containing information accumulated subsequently. That is precisely why causal loops produce such strange paradoxes.

Ordinary language therefore collapses several profoundly different ideas into the phrase travelling through time. Future-directed relativistic travel is differential state evolution, literal restoration of an earlier world is state restoration, creating a new present that resembles an earlier state is state reconstruction, and a closed timelike curve would instead represent causal looping. These are not four technologies for doing the same thing; they are four different physical propositions.

The closest we already come to travelling backwards

Once the past is treated as a previous physical and informational state, something interesting happens: humans have been reconstructing previous states for thousands of years. Memory does it internally, history does it symbolically, archaeology infers earlier states from surviving physical traces, paintings preserve representations, photography captures information carried by light, film reconstructs moving images and sound, and museums combine surviving matter with reconstructed context.

Games add interaction and virtual reality adds immersion. These technologies can be placed on a continuum of state reconstruction. A sentence about ancient Rome reconstructs relatively little, a historical novel reconstructs considerably more conceptually, film adds audiovisual information, a historically sophisticated game adds spatial interaction and causal possibilities, and virtual reality can add embodiment.

At the far end of that continuum sits one of science fiction's more interesting inventions: the Star Trek holodeck. Perhaps the holodeck, rather than the DeLorean, is the more physically interesting model of a time machine.

Amsterdam, 1650

Imagine asking a sufficiently advanced simulation system to reconstruct Amsterdam in 1650. It reconstructs streets, canals, buildings, weather, clothing, language, sounds and smells. You can walk into a tavern, pick up an object, speak with an inhabitant and continue through the city.

Now add everything we know historically. Known inhabitants receive reconstructed biographies, archaeological evidence constrains buildings and objects, shipping records constrain trade, historical documents constrain politics, religion and social relationships, and environmental evidence constrains weather and ecology. Where information is missing, sufficiently sophisticated models infer plausible states while distinguishing inference from recovered fact.

At some point, the experiential distinction between learning about Amsterdam in 1650 and visiting Amsterdam in 1650 could become remarkably small. Yet nothing travelled backwards. Everything is happening in the present physical configuration, where present physical systems have been arranged to reproduce information and relationships characteristic of an earlier state.

That may be the closest physically plausible approximation of backwards time travel available to us: not transporting ourselves into the past, but reconstructing enough of an earlier physical state in the present that we can interact with it again.

Reconstruction eventually runs into information

There is an unavoidable limit. To reconstruct an exact historical state, we would need the information constituting that state, and much of it is no longer accessible to us. Photons have escaped into space, documents were destroyed, people died without recording almost everything they experienced, microscopic correlations became distributed throughout environments, and the exact physical state of almost every historical object is unknown.

AI can improve inference enormously, but inference is not recovery. A future simulation might construct an astonishingly accurate Roman marketplace and produce a merchant who behaves exactly as our best historical, archaeological, linguistic and psychological models predict. Wherever the relevant information is inaccessible, however, the simulation must infer rather than recover.

The encounter is a new physical event constrained by surviving information about an earlier world. State reconstruction can therefore approach extraordinary historical fidelity without becoming state restoration.

Our brains already reconstruct earlier states

There is a biological parallel. Human memory is reconstructive. When I remember childhood, my brain does not physically return to its childhood configuration; it changes now into a physical state representing aspects of an earlier state.

Imagining the future works similarly. My brain changes now into a configuration representing a possible subsequent state. Biological cognition therefore already performs primitive versions of historical reconstruction and future simulation, while books, paintings, films, games and virtual reality externalize that capability. AI can increasingly participate in reconstructing missing structure.

A holodeck would be an extreme technological continuation of something brains already do. Notice that throughout this discussion we never needed to imagine the past as a physical destination in order to talk meaningfully about recovering information from it, representing it or recreating aspects of it.

Putting the abstraction layers back in order

We can now return to where we started. There is the physical world: systems, matter, fields, radiation, interactions, locations, trajectories and changing states. There is measurement: choosing reproducible physical processes and comparing their histories with other physical processes. Then there is the mathematical and conceptual framework through which we describe those comparisons: seconds, coordinate time, proper time, worldlines, metrics and spacetime.

The third layer is not a mistake. It is one of the greatest achievements of science. Without abstraction, physics would scarcely be possible. The whole point of a successful abstraction is that we no longer have to repeat the cumbersome underlying description every time. “Gravitational time dilation” is vastly easier to say than a paragraph describing two physical systems, their trajectories, gravitational conditions, signal exchanges and subsequent clock comparisons.

But successful abstractions carry a particular danger: the better the abstraction works, the easier it becomes to mistake the map for the territory. That is particularly easy with time because our language already treats it as an actor. Time flows, passes, slows, catches up with us; we save it, lose it and travel through it. Physics gives some of these expressions precise technical meanings, while ordinary language can quietly give those technical meanings an ontology they never demonstrated.

Resetting time to its proper place

This is why I think a useful discipline when discussing time is to reverse the usual explanatory order. Before saying that time slowed down, identify the physical histories being compared. Before saying that gravity affected time, identify the physical systems, their trajectories relative to the surrounding mass-energy distribution, the state changes used as clocks and the procedure through which they were compared.

Before saying that one observer experienced less time, ask which physical processes accompanying the observers accumulated different changes. Before imagining spacetime as a container holding accessible past and future destinations, ask what relationships among physical events and histories the geometry actually represents. And before saying that something travelled into the past, ask whether we mean differential evolution, state restoration, state reconstruction or causal looping.

None of these questions weakens relativity. Quite the opposite: once we remove the intuitive baggage carried by words such as time and spacetime, what remains is arguably more remarkable. Physical systems following different histories exhibit universal relationships among their internal processes, and relativity describes those relationships with astonishing precision. A clock is not an observer standing outside this process; it is another changing physical system participating in it, and an observer is not outside it either.

We compare physical change with physical change and one history with another. Time gives us an extraordinarily powerful language for expressing those relationships, and perhaps that is already enough.

Once we keep that hierarchy in mind, even something as fantastic as time travel separates into much clearer physical questions. Travelling into the future means arranging radically different histories of state evolution. Restoring the past would mean reconstructing an earlier physical state with impossible or near-impossible fidelity. A causal loop, if nature permits one at all, would be something different again.

The closest route backwards may therefore turn out not to be a hidden road through a fourth dimension. It may be learning to reconstruct increasingly rich versions of earlier physical states from the traces they left behind. We already do this with memory, history, archaeology, books, photographs, film and simulation, while AI and virtual reality may eventually take us considerably further.

Not backwards through time, but forwards into increasingly convincing reconstructions of what came before. Throughout all of it, the same question remains useful: what physically changed, and what did we compare it with?