Time feels like the most basic feature of reality. Seconds tick, days pass and everything from planetary motion to human memory seems to unfold along a single, irreversible direction. We are born and we die, in exactly that order. We plan our lives around time, measure it obsessively and experience it as an unbroken flow from past to future. It feels so obvious that time moves forward that questioning it can seem almost pointless.

And yet, for more than a century, physics has struggled to say what time actually is. This struggle is not philosophical nitpicking. It sits at the heart of some of the deepest problems in science.

Modern physics relies on different, but equally important, frameworks. One is Albert Einstein’s theory of general relativity, which describes the gravity and motion of large objects such as planets. Another is quantum mechanics, which rules the microcosmos of atoms and particles. And on an even larger scale, the standard model of cosmology describes the birth and evolution of the universe as a whole. All rely on time, yet they treat it in incompatible ways.

When physicists try to combine these theories into a single framework, time often behaves in unexpected and troubling ways. Sometimes it stretches. Sometimes it slows. Sometimes it disappears entirely.

The Insights section is committed to high-quality longform journalism. Our editors work with academics from many different backgrounds who are tackling a wide range of societal and scientific challenges.

Einstein’s theory of relativity was, in fact, the first major blow to our everyday intuition about time. Time, Einstein showed, is not universal. It runs at different speeds depending on gravity and motion. Two observers moving relative to one another will disagree about which events happened at the same time. Time became something elastic, woven together with space into a four-dimensional fabric called spacetime.

Quantum mechanics made things even stranger. In quantum theory, time is not something the theory explains. It is simply assumed. The equations of quantum mechanics describe how systems evolve with respect to time, but time itself remains an external parameter, a background clock that sits outside the theory.

À lire aussi :
Quantum mechanics: how the future might influence the past

This mismatch becomes acute when physicists try to describe gravity at the quantum level, which is crucial for developing the much coveted theory of everything – which links the main fundamental theories. But in many attempts to create such a theory, time vanishes as a parameter from the fundamental equations altogether. The universe appears frozen, described by equations that make no reference to change.

This puzzle is known as the problem of time, and it remains one of the most persistent obstacles to a unified theory of physics. Despite enormous progress in cosmology and particle physics, we still lack a clear explanation for why time flows at all.

Now a relatively new approach to physics, building on a mathematical framework called information theory, developed by Claude Shannon in the 1940s, has started coming up with surprising answers.

Entropy and the arrow of time

When physicists try to explain the direction of time, they often turn to a concept called entropy. The second law of thermodynamics states that disorder tends to increase. A glass can fall and shatter into a mess, but the shards never spontaneously leap back together. This asymmetry between past and future is often identified with the arrow of time.

This idea has been enormously influential. It explains why many processes are irreversible, including why we remember the past but not the future. If the universe started in a state of low entropy, and is getting messier as it evolves, that appears to explain why time moves forward. But entropy does not fully solve the problem of time.

For one thing, the fundamental quantum mechanical equations of physics do not distinguish between past and future. The arrow of time emerges only when we consider large numbers of particles and statistical behaviour. This also raises a deeper question: why did the universe start in such a low-entropy state to begin with? Statistically, there are more ways for a universe to have high entropy than low entropy, just as there are more ways for a room to be messy than tidy. So why would it start in a state that is so improbable?

The information revolution

Over the past few decades, a quiet but far-reaching revolution has taken place in physics. Information, once treated as an abstract bookkeeping tool used to track states or probabilities, has increasingly been recognised as a physical quantity in its own right, just like matter or radiation. While entropy measures how many microscopic states are possible, information measures how physical interactions limit and record those possibilities.

This shift did not happen overnight. It emerged gradually, driven by puzzles at the intersection of thermodynamics, quantum mechanics and gravity, where treating information as merely mathematical began to produce contradictions.

One of the earliest cracks appeared in black hole physics. When Stephen Hawking showed that black holes emit thermal radiation, it raised a disturbing possibility: information about whatever falls into a black hole might be permanently lost as heat. That conclusion conflicted with quantum mechanics, which demands that the entirety of information be preserved.

Resolving this tension forced physicists to confront a deeper truth. Information is not optional. If we want a full description of the universe that includes quantum mechanics, information cannot simply disappear without undermining the foundations of physics. This realisation had profound consequences. It became clear that information has thermodynamic cost, that erasing it dissipates energy, and that storing it requires physical resources.

In parallel, surprising connections emerged between gravity and thermodynamics. It was shown that Einstein’s equations can be derived from thermodynamic principles that link spacetime geometry directly to entropy and information. In this view, gravity doesn’t behave exactly like a fundamental force.

Instead, gravity appears to be what physicists call “emergent” – a phenomenon describing something that’s greater than the sum of its parts, arising from more fundamental constituents. Take temperature. We can all feel it, but on a fundamental level, a single particle can’t have temperature. It’s not a fundamental feature. Instead it only emerges as a result of many molecules moving collectively.

Similarly, gravity can be described as an emergent phenomenon, arising from statistical processes. Some physicists have even suggested that gravity itself may emerge from information, reflecting how information is distributed, encoded and processed.

These ideas invite a radical shift in perspective. Instead of treating spacetime as primary, and information as something that lives inside it, information may be the more fundamental ingredient from which spacetime itself emerges. Building on this research, my colleagues and I have explored a framework in which spacetime itself acts as a storage medium for information – and it has important consequences for how we view time.

In this approach, spacetime is not perfectly smooth, as relativity suggests, but composed of discrete elements, each with a finite capacity to record quantum information from passing particles and fields. These elements are not bits in the digital sense, but physical carriers of quantum information, capable of retaining memory of past interactions.

A useful way to picture them is to think of spacetime like a material made of tiny, memory-bearing cells. Just as a crystal lattice can store defects that appeared earlier in time, these microscopic spacetime elements can retain traces of the interactions that have passed through them. They are not particles in the usual sense described by the standard model of particle physics, but a more fundamental layer of physical structure that particle physics operates on rather than explains.

This has an important implication. If spacetime records information, then its present state reflects not only what exists now, but everything that has happened before. Regions that have experienced more interactions carry a different imprint of information than regions that have experienced fewer. The universe, in this view, does not merely evolve according to timeless laws applied to changing states. It remembers.

A recording cosmos

This memory is not metaphorical. Every physical interaction leaves an informational trace. Although the basic equations of quantum mechanics can be run forwards or backwards in time, real interactions never happen in isolation. They inevitably involve surroundings, leak information outward and leave lasting records of what has occurred. Once this information has spread into the wider environment, recovering it would require undoing not just a single event, but every physical change it caused along the way. In practice, that is impossible.

This is why information cannot be erased and broken cups do not reassemble. But the implication runs deeper. Each interaction writes something permanent into the structure of the universe, whether at the scale of atoms colliding or galaxies forming.

Geometry and information turn out to be deeply connected in this view. In our work, we have showed that how spacetime curves depends not only on mass and energy, as Einstein taught us, but also on how quantum information, particularly entanglement, is distributed. Entanglement is a quantum process that mysteriously links particles in distant regions of space – it enables them to share information despite the distance. And these informational links contribute to the effective geometry experienced by matter and radiation.

From this perspective, spacetime geometry is not just a response to what exists at a given moment, but to what has happened. Regions that have recorded many interactions tend, on average, to behave as if they curve more strongly, have stronger gravity, than regions that have recorded fewer.

This reframing subtly changes the role of spacetime. Instead of being a neutral arena in which events unfold, spacetime becomes an active participant. It stores information, constrains future dynamics and shapes how new interactions can occur. This naturally raises a deeper question. If spacetime records information, could time emerge from this recording process rather than being assumed from the start?

Time arising from information

Recently, we extended this informational perspective to time itself. Rather than treating time as a fundamental background parameter, we showed that temporal order emerges from irreversible information imprinting. In this view, time is not something added to physics by hand. It arises because information is written in physical processes and, under the known laws of thermodynamics and quantum physics, cannot be globally unwritten again. The idea is simple but far-reaching.

Every interaction, such as two particles crashing, writes information into the universe. These imprints accumulate. Because they cannot be erased, they define a natural ordering of events. Earlier states are those with fewer informational records. Later states are those with more.

Quantum equations do not prefer a direction of time, but the process of information spreading does. Once information has been spread out, there is no physical path back to a state in which it was localised. Temporal order is therefore anchored in this irreversibility, not in the equations themselves.

Time, in this view, is not something that exists independently of physical processes. It is the cumulative record of what has happened. Each interaction adds a new entry, and the arrow of time reflects the fact that this record only grows.

The future differs from the past because the universe contains more information about the past than it ever can about the future. This explains why time has a direction without relying on special, low-entropy initial conditions or purely statistical arguments. As long as interactions occur and information is irreversibly recorded, time advances.

Interestingly, this accumulated imprint of information may have observable consequences. At galactic scales, the residual information imprint behaves like an additional gravitational component, shaping how galaxies rotate without invoking new particles. Indeed, the unknown substance called dark matter was introduced to explain why galaxies and galaxy clusters rotate faster than their visible mass alone would allow.

In the informational picture, this extra gravitational pull does not come from invisible dark matter, but from the fact that spacetime itself has recorded a long history of interactions. Regions that have accumulated more informational imprints respond more strongly to motion and curvature, effectively boosting their gravity. Stars orbit faster not because more mass is present, but because the spacetime they move through carries a heavier informational memory of past interactions.

From this viewpoint, dark matter, dark energy and the arrow of time may all arise from a single underlying process: the irreversible accumulation of information.

Testing time

But could we ever test this theory? Ideas about time are often accused of being philosophical rather than scientific. Because time is so deeply woven into how we describe change, it is easy to assume that any attempt to rethink it must remain abstract. An informational approach, however, makes concrete predictions and connects directly to systems we can observe, model and in some cases experimentally probe.

Black holes provide a natural testing ground, as they seems to suggest information is erased. In the informational framework, this conflict is resolved by recognising that information is not destroyed but imprinted into spacetime before crossing the horizon. The black hole records it.

This has an important implication for time. As matter falls toward a black hole, interactions intensify and information imprinting accelerates. Time continues to advance locally because information continues to be written, even as classical notions of space and time break down near the horizon and appear to slow or freeze for distant observers.

As the black hole evaporates through Hawking radiation, the accumulated informational record does not vanish. Instead, it affects how radiation is emitted. The radiation should carry subtle signs that reflect the black hole’s history. In other words, the outgoing radiation is not perfectly random. Its structure is shaped by the information previously recorded in spacetime. Detecting such signs remains beyond current technology, but they provide a clear target for future theoretical and observational work.

The same principles can be explored in much smaller, controlled systems. In laboratory experiments with quantum computers, qubits (the quantum computer equivalent of bits) can be treated as finite-capacity information cells, just like the spacetime ones. Researchers have shown that even when the underlying quantum equations are reversible, the way information is written, spread and retrieved can generate an effective arrow of time in the lab. These experiments allow physicists to test how information storage limits affect reversibility, without needing cosmological or astrophysical systems.

Extensions of the same framework suggest that informational imprinting is not limited to gravity. It may play a role across all fundamental forces of nature, including electromagnetism and the nuclear forces. If this is correct, then time’s arrow should ultimately be traceable to how all interactions record information, not just gravitational ones. Testing this would involve looking for limits on reversibility or information recovery across different physical processes.

Taken together, these examples show that informational time is not an abstract reinterpretation. It links black holes, quantum experiments and fundamental interactions through a shared physical mechanism, one that can be explored, constrained and potentially falsified as our experimental reach continues to grow.

What time really is

Ideas about information do not replace relativity or quantum mechanics. In everyday conditions, informational time closely tracks the time measured by clocks. For most practical purposes, the familiar picture of time works extremely well. The difference appears in regimes where conventional descriptions struggle.

Near black hole horizons or during the earliest moments of the universe, the usual notion of time as a smooth, external coordinate becomes ambiguous. Informational time, by contrast, remains well defined as long as interactions occur and information is irreversibly recorded.

All this may leave you wondering what time really is. This shift reframes the longstanding debate. The question is no longer whether time must be assumed as a fundamental ingredient of the universe, but whether it reflects a deeper underlying process.

In this view, the arrow of time can emerge naturally from physical interactions that record information and cannot be undone. Time, then, is not a mysterious background parameter standing apart from physics. It is something the universe generates internally through its own dynamics. It is not ultimately a fundamental part of reality, but emerges from more basic constituents such as information.

Whether this framework turns out to be a final answer or a stepping stone remains to be seen. Like many ideas in fundamental physics, it will stand or fall based on how well it connects theory to observation. But it already suggests a striking change in perspective.

The universe does not simply exist in time. Time is something the universe continuously writes into itself.

For you: more from our Insights series:

To hear about new Insights articles, join the hundreds of thousands of people who value The Conversation’s evidence-based news. Subscribe to our newsletter.