Can You Turn Back Time?

Time Is Broken showed how time bends near mass and motion. Does Time Even Exist? asked the deeper question of whether it fundamentally exists at all. This post asks a narrower one: can you move through it in the wrong direction? The answer, according to the equations, is “maybe” – and the universe isn’t letting us know.

Forward time travel is easy

Before tackling the hard direction, it’s worth noting that forward time travel is a solved problem. It’s been happening since the universe had mass and relative motion – we’ve just been proving it since 1971.

The twin paradox is real. Move fast enough relative to someone else, and less time passes for you. The Hafele-Keating experiment confirmed it with caesium clocks on commercial airliners. GPS satellites confirm it every second of every day. Scott Kelly came back from the ISS 5 milliseconds younger than his twin.

If you want to travel a thousand years into the future, the recipe is straightforward: accelerate to a significant fraction of the speed of light, cruise for a while (by your clock), decelerate, and come home. The energy requirements are absurd – accelerating a modest spacecraft to 99% of light speed would require more energy than the entire world currently produces in a year – but the physics is not in dispute. You would arrive in the future. Everyone you knew would be dead. Going back would need a different mechanism entirely – which is where this gets interesting.

Gravitational time dilation offers another route. Park yourself near (but not inside) a black hole, wait a while by your clock, then fly away. Less time passes for you than for the universe outside. The film Interstellar got this broadly right: the characters who visited the planet near the black hole aged hours while decades passed outside. The specific numbers in the film were dramatised, but the principle is textbook general relativity.

Forward time travel isn’t speculative. It’s engineering.

Backward time travel: the equations say yes

Going backward is where things get interesting – and contested.

The equations of general relativity describe the geometry of spacetime. They’re not suggestions. They’re constraints: given a distribution of mass and energy, the equations tell you exactly how spacetime curves. And some solutions to those equations contain closed timelike curves (CTCs) – paths through spacetime that loop back on themselves. Travel along a CTC, and you return to your own past. A handful of such solutions are known; each one is mathematically valid, and each one is physically strange in its own way.

Gödel’s rotating universe. The solution where CTCs were first recognised for what they were. In 1949, Kurt Gödel found a solution to Einstein’s equations describing a universe that rotates as a whole, in which sufficiently long journeys through spacetime loop back to their starting point in time. You could, in principle, attend your own birth. CTCs had quietly been present in an earlier solution – Willem van Stockum’s 1937 infinite rotating dust cylinder – but nobody noticed until Frank Tipler pointed it out in 1974. Gödel’s was where the phenomenon became impossible to ignore.

Gödel presented this solution as a birthday gift to Einstein. It’s unclear whether Einstein was delighted or horrified. Gödel’s universe doesn’t match ours – ours expands (his doesn’t) and the cosmic microwave background shows no sign of global rotation to extremely tight bounds – but that’s not the point. The point is that general relativity, taken at face value, permits time travel. The equations don’t forbid it. Gödel proved that any argument of the form “time travel is impossible because it violates general relativity” is wrong. The theory allows it. Whether the universe chooses to use that allowance is a different question.

The Kerr metric. Another CTC solution – and one we can point a telescope at. In 1963, Roy Kerr found the solution for a rotating black hole. The Event Horizon Telescope has since imaged M87* and Sgr A* directly; LIGO routinely catches pairs of spinning black holes merging. The geometry is real; whether the CTC region of the geometry is real is another question. Kerr’s solution contains closed timelike curves deep in the interior, behind the inner event horizon. In the mathematical solution, you could pass through the ring singularity and emerge in a region where time loops are possible.

Whether this is physically meaningful is debated. The interior of the Kerr solution may be unstable – perturbations might destroy the closed timelike curves before anything could traverse them. But the mathematical structure is there, and it’s a solution to the same equations that predict GPS corrections and gravitational waves.

Wormholes. In 1988, Kip Thorne (who would later win a Nobel Prize for LIGO) showed that if traversable wormholes exist – shortcuts through spacetime connecting distant regions – they could be converted into time machines. The recipe: take one end of a wormhole, accelerate it to near-light speed, then bring it back. Time dilation means less time has passed at the accelerated end. Enter the “slow” end and you emerge from the “fast” end at an earlier time. You’ve gone backward.

Thorne wasn’t trying to design a time machine. He was responding to a question from Carl Sagan, who was writing Contact and wanted the physics to be plausible. But the analysis was rigorous, published in Physical Review Letters, and it launched a serious research programme into the physics of time travel that continues today.

The catch is that we don’t know if traversable wormholes can exist. They require “exotic matter” with negative energy density to keep them open. Quantum field theory allows negative energy densities in certain configurations (the Casimir effect is a real example), but whether you can get enough of it, concentrated enough, to hold open a wormhole is unknown.

The Tipler cylinder. The same Frank Tipler who’d dredged van Stockum’s CTCs out of obscurity didn’t stop at reanalysing other people’s work. In the same 1974 paper, he showed that an infinitely long, extremely dense, rapidly rotating cylinder would drag spacetime around it hard enough to create closed timelike curves of its own. Finite cylinders don’t work – Stephen Hawking proved that the closed timelike curves require the cylinder to be infinite. This makes it impractical (to put it mildly) but it’s another example of the equations permitting what intuition forbids.

The grandfather paradox and self-consistency

If backward time travel is possible, what stops you from killing your own grandfather before your parent is born? This is the oldest and most intuitive objection to time travel.

The Novikov self-consistency principle offers one resolution. Proposed by Igor Novikov in the 1980s, it states that any events on a closed timelike curve must be self-consistent. You can travel to the past, but you can’t change it – because you didn’t. Whatever you do in the past has already happened. It’s already part of the history that led to you travelling backward in the first place.

Think of it like a jigsaw puzzle. You can’t place a piece that doesn’t fit. If you travel back and try to kill your grandfather, something prevents it – you slip, you miss, you change your mind. Not because of magic, but because the version of history where you succeed is logically inconsistent and therefore doesn’t exist. Only self-consistent histories are allowed.

This isn’t as strange as it sounds. We already accept that physical laws constrain what’s possible. You can’t build a perpetual motion machine, not because someone stops you, but because the laws of thermodynamics don’t permit it. The Novikov principle says that self-consistency is a similar constraint: the laws of physics, applied to closed timelike curves, only admit solutions where the timeline is internally coherent.

The Deutsch model takes a quantum approach. David Deutsch, in 1991, applied quantum mechanics to the grandfather paradox and showed that closed timelike curves are consistent if you allow the universe to be in a mixed quantum state. Roughly: the traveller who emerges from the time loop is not identical to the one who entered it. They’re a quantum mixture – partly themselves, partly a version from a slightly different history. This avoids paradoxes at the cost of letting quantum mechanics redefine what “the traveller” even means. Which, given everything else about quantum mechanics, is perhaps not a high price.

The quantum eraser: does the future affect the past?

In 1999, Yoon-Ho Kim and colleagues performed an experiment that seems to suggest the future can influence the past. It’s called the delayed-choice quantum eraser, and it’s one of the most unsettling experiments in physics.

Here’s the setup, simplified. You send photons through a double slit. Normally, they produce an interference pattern on a detector – the signature of quantum mechanics, showing the photons behaving as waves. But if you add a detector that tells you which slit each photon went through, the interference pattern disappears. The photons behave as particles. This much is standard quantum mechanics.

Now the twist. Kim’s experiment split each photon into two entangled partners. One partner (the “signal”) went to a screen. The other (the “idler”) went on a longer path to a second detector, where the “which-path” information was either preserved or erased – after the signal photon had already hit the screen.

When the experimenters later compared the data, they found that the signal photons whose idler partners had their which-path information erased showed an interference pattern. The ones whose idler partners retained the information did not. The choice about the idler – made after the signal photon hit the screen – appeared to retroactively determine whether the signal photon behaved as a wave or a particle.

This is not, despite appearances, evidence of backward causation. The interference pattern only becomes visible when you sort the signal photons using information from the idlers. If you look at all the signal photons together, without sorting, there’s no interference pattern. The “retrocausal” effect is an artefact of post-selection, not a signal travelling backward in time. Still, the lesson is real: quantum correlations don’t respect our intuitions about the order of cause and effect. The universe doesn’t care which measurement happened first. The entanglement ties the results together regardless of timing.

Hawking’s party

In 2009, Stephen Hawking threw a party for time travellers. He prepared champagne, put up a banner reading “Welcome, Time Travellers,” set coordinates, and waited. Nobody came.

He published the invitation afterward – so that future time travellers would know when and where to show up. The fact that nobody arrived was, Hawking suggested with a grin, “experimental evidence that time travel is not possible.”

It was a joke, mostly. The absence of guests doesn’t prove much – perhaps time travellers can’t travel to before the machine was built, or perhaps they chose not to come, or perhaps the invite was lost in the noise of history. But it illustrates Hawking’s own position: he believed the universe has a chronology protection mechanism that prevents closed timelike curves from forming.

His chronology protection conjecture, published in 1992, argues that whenever conditions approach those needed for a time loop, quantum effects – specifically, a divergence in the stress-energy tensor of the vacuum – intervene and destroy the loop before it can form. The back-reaction of quantum fields near a forming CTC generates enough energy to collapse the would-be time machine.

“It seems there is a chronology protection agency which prevents the appearance of closed timelike curves and so makes the universe safe for historians,” Hawking wrote. The conjecture is unproven. It might be wrong. But the fact that it was needed at all – that someone of Hawking’s stature felt the need to propose a law preventing time travel – tells you how seriously the equations permit it.

Retrocausality: a serious proposal

Most of this post has treated backward-in-time effects as paradoxical or impossible. But a growing number of physicists are taking retrocausality – genuine backward-in-time influence – seriously as a foundation for quantum mechanics.

The motivation is Bell’s theorem. In 1964, John Bell proved that quantum mechanics cannot be explained by any theory where particles have pre-existing properties and influences travel no faster than light. Experiments have repeatedly confirmed quantum mechanics. So at least one of those assumptions must be wrong.

Most physicists give up the pre-existing properties (this is the standard “Copenhagen” or “many-worlds” approach). But a minority – including Huw Price at Cambridge and Ken Wharton at San José State – argue that we should instead give up the assumption that causes always precede effects. If influences can travel backward in time, Bell’s theorem is satisfied without giving up realism. Particles do have definite properties; it’s just that future measurements can influence past states.

This isn’t crackpot physics. Price and Wharton’s work is published in peer-reviewed journals and taken seriously by the foundations-of-physics community. It’s a minority position, but it’s a legitimate interpretation, and it has the advantage of preserving something that most quantum interpretations sacrifice: the idea that things have definite properties even when nobody’s looking.

The price is steep. Retrocausality means that the state of a particle right now depends partly on what will happen to it in the future. Not in a way that lets you send messages backward (that would violate other constraints), but in a way that makes the universe’s bookkeeping work out. The future doesn’t cause the past in the way you’d normally use the word. It constrains it, the way a jigsaw puzzle constrains which pieces can go where.

What we actually know

1. Forward time travel is real. We’ve measured it. GPS depends on it. It’s engineering, not speculation.

2. General relativity permits closed timelike curves. Multiple exact solutions to Einstein’s equations contain them. This is mathematics, not handwaving.

3. We don’t know if the universe actually allows them. Hawking’s chronology protection conjecture says no, but it’s unproven. Quantum gravity might resolve this, but we don’t have a theory of quantum gravity.

4. The grandfather paradox has solutions. The Novikov principle (self-consistency) and the Deutsch model (quantum mixed states) both resolve it without contradiction.

5. Quantum mechanics is weird about time. Entanglement doesn’t respect temporal ordering. The delayed-choice quantum eraser demonstrates this without actually sending information backward.

6. Retrocausality is a legitimate interpretation. A minority of physicists take it seriously as a foundation for quantum mechanics.

7. Nobody came to Hawking’s party. Make of that what you will.

The physics of time travel isn’t a closed question. It’s an open one, sitting at the intersection of general relativity, quantum mechanics, and quantum gravity – precisely the intersection where our best theories break down. Until we have a theory that works at that intersection, the equations say “maybe” and the universe isn’t talking.

There’s another clock to examine, though – the one inside you. It has no caesium atom and no GPS correction. It runs on light, adenosine, and a cluster of twenty thousand neurons behind your eyes. And it sets the terms for how you experience every other clock in this series.

The Clock Inside You is next – the biology of jet lag, shift work, and why your body refuses to run on UTC.