The Twin Paradox

Two siblings synchronise their watches, and one of them leaves. She boards a ship, runs it up to a fair fraction of the speed of light, coasts out to a distant star, swings around, and comes home. Her watch and her brother's no longer agree — hers reads less. Not as an illusion, not as a trick of signalling delay, but as a permanent physical fact stamped into the two watches when they are laid side by side on the kitchen table. The traveller is younger. The famous "paradox" is the protest that follows: motion is relative, so from the traveller's deck it was the Earth that flew away and came back, and by the very same time-dilation rule the stay-at-home ought to be the younger one. Both cannot be true. Something must break the symmetry — and that something is not a matter of opinion. This article takes the proper time owned by aq-gr-proper-time — the invariant arc length $c\,\mathrm{d}\tau=\sqrt{-g_{\mu\nu}\,\mathrm{d}x^\mu\mathrm{d}x^\nu}$ that a clock reads along its own worldline — and turns it on the most argued-about scenario in all of relativity. The resolution is not a verbal dodge about "who really accelerated"; it is a theorem. Between two timelike-separated events, the straight inertial worldline carries more proper time than any bent one, because the minus sign in the metric turns the triangle inequality upside down. The stay-at-home twin walks the straight path, the traveller walks the bent one, and the geometry does the rest. Everything else — the acceleration, the leaping now-line, the Doppler ledger — is commentary on that single geometric fact.

Key takeaways

Two twins part at one event and reunite at another; the one who travels out, turns around, and returns has aged less than the one who stayed inertial — a real, measured effect.
There is no contradiction: the situation is not symmetric. Only the traveller changes inertial frames, so the reciprocal "each sees the other slow" argument cannot be applied both ways.
The clean statement is geometric: elapsed age is the proper time $\tau=\tfrac1c\int\sqrt{-g_{\mu\nu}\mathrm{d}x^\mu\mathrm{d}x^\nu}$ along the worldline, a path-dependent functional, not a global clock reading.
Between two timelike-separated events the inertial (straight) worldline maximises proper time — the reverse of the Euclidean triangle inequality, forced by the metric's minus sign.
For a constant-speed round trip the traveller's age is shorter by exactly $1/\gamma=\sqrt{1-v^2/c^2}$ ; the "missing" time is bookkept by the relativity of simultaneity at the turnaround, not by the size of the acceleration.
Key idea: proper time is a path-dependent functional of the worldline, $c\,\tau=\int\sqrt{-g_{\mu\nu}\mathrm{d}x^\mu\mathrm{d}x^\nu}$ . Between two timelike-separated events the inertial worldline maximises it, so the twin who never accelerates always ages the most.

The core insight

Proper time is a path-dependent functional of the worldline, $c\,\tau=\int\sqrt{-g_{\mu\nu}\mathrm{d}x^\mu\mathrm{d}x^\nu}$ . Between two timelike-separated events the inertial (straight) worldline maximises proper time, so the twin who never accelerates always ages the most — the asymmetry is geometry, not a debate about frames.

Where this lives today

The paradox stopped being a thought experiment the moment atomic clocks grew portable. Its literal realization is the 1971 Hafele–Keating experiment — caesium clocks flown east and west around the world aboard airliners, returning offset from a ground reference by tens to hundreds of nanoseconds, the sign depending on flight direction because the honest comparison frame is the non-rotating geocentric one. Every Global Positioning System satellite is a continuous twin experiment: its orbiting clock ticks slow by about 7 microseconds per day from its speed and fast by about 45 microseconds per day from its higher gravitational potential, a net advance near 38 microseconds per day that, left uncorrected, would poison positioning by roughly 10 kilometres daily. Cosmic-ray muons play the travelling twin every second — born high in the atmosphere with a 2.2-microsecond rest-frame lifetime, far too short to reach the ground classically, they arrive in abundance because their proper time runs short along their fast worldline, and the CERN muon storage ring pinned this to sub-percent accuracy at $\gamma\approx29$ . Looking outward, the relativistic-rocket calculation of the third worked example is the honest arithmetic behind every serious discussion of crewed interstellar travel: a starship crew could in principle cross the galaxy within a lifetime while millennia elapse on Earth, the twin paradox scaled from a puzzle to a one-way ticket.

The physical picture

Think of a clock's proper time as the odometer wound along its worldline. Two cars driving between the same two towns can clock different mileages depending on the route; two clocks carried between the same two events read different elapsed times depending on the worldline that joins them. The one surprise, and it is the whole subject, is that in spacetime the straight route reads the most, not the least — because the timelike part of the metric enters with a minus sign, so wandering off through space subtracts from the accumulated time. The stay-at-home twin coasts straight up the time axis; the traveller detours out into space and back; and when their worldlines meet again the straight one has logged more. The first figure draws both paths and shows exactly where the traveller's "missing" years went.

Figure

Two worldlines meet at events P and Q. Alice coasts straight and ages τ = T, while Bob flies out and back at v = 0.8c and ages only τ = 0.6 T. At the turnaround Bob's now-line at home leaps from the lower dashed line to the upper one, skipping 0.64 T of Alice's history — the relativity of simultaneity, not the acceleration, is where the missing time hides

The symmetry-breaking is concrete, and you can hear it. Let each twin beam a continuous radio signal carrying an image of their own clock face. On the outbound leg each receives the other's signal redshifted — running slow; on the inbound leg, blueshifted — running fast. The catch is that Bob switches from slow-receiving to fast-receiving at the midpoint of his trip, spending half his time getting slow signals and half getting fast ones. Alice, though, keeps receiving Bob's slow outbound signal long after Bob has already turned around, because the light showing his turnaround is still crawling home across the gap. She spends most of her wait on slow signals and only a short final stretch on fast ones. Tally the Doppler shifts over each leg and the asymmetric total drops out exactly — no acceleration formula required. The second figure quantifies the punchline directly: how much the traveller's clock lags as a function of speed.

Figure

The age ratio of traveller to stay-at-home for a constant-speed round trip, τ traveller over τ home = √(1 − v²/c²). It is essentially 1 at low speed, so no difference is measurable; it falls to 0.80 at v = 0.6c and 0.60 at v = 0.8c, and plunges toward 0 as v → c, so the traveller can return arbitrarily young while decades pass at home

Now picture the turnaround as instantaneous. Bob's plane of simultaneity — the set of events he calls "now, back home" — swings around as he changes velocity. In the single instant he reverses, that "now" line at Earth sweeps forward through a large block of Alice's history. That swept-out interval, the green wedge in the first figure, is precisely the chunk of Alice's life that Bob's naive reciprocal reasoning failed to count. There is no time truly missing and no clock running wrong; the gap was always hidden in the relativity of simultaneity at the corner of Bob's worldline, and the moment you draw the worldlines it becomes visible.

The core equations

The operational quantity is the proper time increment: the reading a comoving clock accumulates over an infinitesimal step of its worldline, built from the invariant interval with the metric as the ruler. In signature $-+++$ the timelike interval is negative, so we flip its sign and divide by $c^2$ to get a positive $\mathrm{d}\tau^2$ . Everything below is a specialization, an integral, or a ratio of this one line.

\mathrm{d}\tau=\frac{1}{c}\sqrt{-\eta_{\mu\nu}\,\mathrm{d}x^\mu\mathrm{d}x^\nu}=\sqrt{\mathrm{d}t^2-\frac{1}{c^2}\left(\mathrm{d}x^2+\mathrm{d}y^2+\mathrm{d}z^2\right)} \qquad (1)

where: $\tau$ — proper time read by a clock on the worldline, an invariant scalar; $\eta_{\mu\nu}$ — Minkowski metric, $\mathrm{diag}(-1,+1,+1,+1)$ ; $\mathrm{d}x^\mu$ — infinitesimal displacement along the worldline; $c$ — speed of light. Reading it physically: sitting still ( $\mathrm{d}x^i=0$ ) gives $\mathrm{d}\tau=\mathrm{d}t$ , and any motion strictly shrinks $\mathrm{d}\tau$ below $\mathrm{d}t$ . To get the total age accumulated between two events, integrate the increment along the whole worldline against any convenient parameter $\lambda$ . The result depends only on the geometric path, not on the labelling — reparametrise $\lambda$ and the Jacobian cancels — which is exactly what makes proper time an arc length rather than a coordinate.

\tau[\gamma]=\frac{1}{c}\int_{\lambda_1}^{\lambda_2}\sqrt{-\eta_{\mu\nu}\frac{\mathrm{d}x^\mu}{\mathrm{d}\lambda}\frac{\mathrm{d}x^\nu}{\mathrm{d}\lambda}}\;\mathrm{d}\lambda \qquad (2)

where: $\gamma$ — the worldline joining the two events; $\lambda$ — arbitrary monotonic parameter along it; $\tau[\gamma]$ — total elapsed proper time, a functional of the entire path. Physical reading: because $\tau$ depends on the whole curve and not just its endpoints, two worldlines between the same two events generically report different totals — this is the mathematical seed of the paradox. Specialize to an inertial home frame and parametrize by that frame's coordinate time $t$ . Factoring $\mathrm{d}t^2$ out of equation (1) and identifying the coordinate speed $v=|\mathrm{d}\mathbf{x}/\mathrm{d}t|$ gives proper time its most familiar face, the Lorentz factor.

\mathrm{d}\tau=\sqrt{1-\frac{v^2}{c^2}}\;\mathrm{d}t=\frac{\mathrm{d}t}{\gamma},\qquad \gamma=\frac{1}{\sqrt{1-v^2/c^2}} \qquad (3)

where: $v$ — the clock's speed in the home inertial frame; $t$ — that frame's coordinate time; $\gamma\ge1$ — Lorentz factor. Physical reading: a clock moving at speed $v$ advances at rate $1/\gamma<1$ against coordinate time. This is time dilation seen not as an optical effect but as the length of a worldline. Applying equation (3) to each twin turns the whole problem into two integrals over the same coordinate-time interval $T$ . Alice sits at $v=0$ ; Bob carries a speed profile $v(t)$ that must be nonzero to leave and return.

\Delta\tau_{\text{Alice}}=\int_0^T \mathrm{d}t=T,\qquad \Delta\tau_{\text{Bob}}=\int_0^T\frac{\mathrm{d}t}{\gamma(t)}\le T \qquad (4)

where: $T$ — total coordinate time of the trip, equal to Alice's ageing; $\gamma(t)$ — Bob's instantaneous Lorentz factor, $\ge1$ at every instant. Physical reading: the integrand for Bob is never larger than one, so his total can only fall short of Alice's — the inequality is built in before any turnaround is even specified. For the idealized constant-speed round trip ( $v$ the same out and back, instantaneous reversal), the integral collapses to a single factor: the whole age difference is captured by one number.

\frac{\Delta\tau_{\text{Bob}}}{\Delta\tau_{\text{Alice}}}=\frac{1}{\gamma}=\sqrt{1-\frac{v^2}{c^2}}\qquad(\text{constant-speed round trip}) \qquad (5)

where: $v$ — the constant outbound and inbound speed. Physical reading: this is the curve plotted in the second figure; at $v=0.8c$ it equals $0.6$ , at $v=0.6c$ it equals $0.8$ , and it heads to zero as $v\to c$ — the traveller can in principle return arbitrarily young. Finally, the bookkeeping that dissolves the paradox lives in the simultaneity shift at the corner of Bob's worldline. Comparing the "home time now" that Bob reckons just before and just after his velocity flips gives a finite jump.

\Delta t_{\text{jump}}=\frac{2vL}{c^2}=\frac{v^2}{c^2}\,T \qquad (6)

where: $L=v\,(T/2)$ — home-frame distance to the turnaround; $v$ — travel speed, whose sign reverses at the corner. Physical reading: at $v=0.8c$ the jump is $0.64\,T$ — exactly the green wedge of Alice's history the first figure shows Bob's now-line sweeping across, and exactly the amount his naive reciprocal count was missing.

Derivation ARound-trip age difference from the proper-time integral

The cleanest resolution invokes no observer at all: compute each twin's elapsed age directly as the integral of $\mathrm{d}\tau$ along their worldline, and simply compare the two numbers. This is a coordinate-free comparison of two path lengths in spacetime, and it settles who is younger before anyone argues about frames. Work in Alice's inertial frame; let the trip last coordinate time $T$ ; let Bob fly out at constant speed $v$ for the first half and return at speed $v$ for the second half, with a turnaround so brief its duration is negligible. We integrate equation (3) leg by leg.

Step 1. For Alice, $v=0$ throughout, so equation (3) gives $\mathrm{d}\tau_{\text{Alice}}=\mathrm{d}t$ and $\Delta\tau_{\text{Alice}}=\int_0^T\mathrm{d}t=T$ . Why: a clock at rest at the origin of an inertial frame is the unique case where proper time equals coordinate time, fixing the benchmark every other path is measured against.

Step 2. On Bob's outbound leg the speed is the constant $v$ , so the integrand $\sqrt{1-v^2/c^2}$ is constant and $\Delta\tau_{\text{out}}=\int_0^{T/2}\sqrt{1-v^2/c^2}\,\mathrm{d}t=\tfrac{T}{2}\sqrt{1-v^2/c^2}$ . Why: constant speed pulls the Lorentz factor out of the integral, turning a functional into an ordinary product.

Step 3. The inbound leg has the same speed $v$ — only the direction reversed, and only $v^2$ enters the metric — so $\Delta\tau_{\text{in}}=\tfrac{T}{2}\sqrt{1-v^2/c^2}$ , identical to the outbound leg. Why: proper time cares about the magnitude of the velocity, not its heading, so the two legs contribute equally.

Step 4. Summing, $\Delta\tau_{\text{Bob}}=\Delta\tau_{\text{out}}+\Delta\tau_{\text{in}}=T\sqrt{1-v^2/c^2}=T/\gamma<T=\Delta\tau_{\text{Alice}}$ . Why: the traveller ages less by exactly the Lorentz factor; the turnaround itself contributes zero proper time in the idealization, yet is indispensable — it is what makes Bob's worldline a bent, non-geodesic curve between the same two events. Verified: at $v=0.8c$ , $\gamma=5/3$ , so $\Delta\tau_{\text{Bob}}=0.6\,T$ ; for $T=10$ yr Bob ages 6 yr, Alice 10 yr. Cf. MTW §1.5, Hartle ch. 5.

Go deeper · Derivation B · 1 minThe inertial worldline maximises proper time

Derivation BThe inertial worldline maximises proper time

The paradox has a purely geometric heart: among all timelike worldlines connecting two fixed timelike-separated events, the straight inertial one accumulates the greatest proper time. This is the exact reverse of the Euclidean triangle inequality — where a straight line is the shortest path — and the reversal is forced by the single minus sign the metric carries on the spatial part. Establishing it shows that the stay-at-home twin ages most for a reason far deeper than any story about rockets: she simply walks the longest available path. Work in the inertial frame where the two events are $P=(0,\vec0)$ and $Q=(T,\vec0)$ , so the straight worldline is the one sitting at $\vec x=0$ .

Step 1. Parametrize an arbitrary timelike worldline from $P$ to $Q$ by the coordinate time $t$ of this frame, giving $\tau[\gamma]=\int_0^T\sqrt{1-v^2(t)/c^2}\,\mathrm{d}t$ with $v(t)=|\mathrm{d}\vec x/\mathrm{d}t|$ its instantaneous speed. Why: using $t$ as the parameter puts every candidate path on the same footing, integrated over the identical interval $[0,T]$ , so only the speed profile distinguishes them.

Step 2. Bound the integrand pointwise: $\sqrt{1-v^2(t)/c^2}\le1$ , with equality if and only if $v(t)=0$ at that instant. Why: the square root can never exceed one, and any real speed strictly lowers it — spatial motion is a penalty on accumulated proper time, never a bonus.

Step 3. Integrate the inequality across the whole trip: $\tau[\gamma]=\int_0^T\sqrt{1-v^2/c^2}\,\mathrm{d}t\le\int_0^T 1\,\mathrm{d}t=T=\tau_{\text{straight}}$ . Why: a path that leaves $\vec x=0$ must carry $v\neq0$ over some stretch to return, so its integral is strictly less than $T$ ; only the worldline that never moves attains the maximum.

Step 4. Interpret the result. In Euclidean geometry the straight line minimises length; in Lorentzian geometry the straight timelike line maximises proper time, the inequality flipped by the metric's minus sign. Why: the stay-at-home twin's worldline is precisely this maximiser, so she ages most — and the theorem holds regardless of how gently or violently the traveller turns around. Verified: consistent with Wald §9.3 and Carroll §1.7 that timelike geodesics in Minkowski space locally maximise proper time; the strict inequality follows from $\sqrt{1-v^2/c^2}<1$ for $v\neq0$ .

Go deeper · Derivation C · 2 minSymmetry resolved by the relativity of simultaneity

Derivation CSymmetry resolved by the relativity of simultaneity

The most persistent objection is reciprocal: on each inertial leg Bob is entitled to say "it is Alice's clock that runs slow", so where does her extra ageing come from? The answer is that Bob's claim is correct on each leg, but stitching his two legs together is not the same as staying in one frame — and the seam reveals a jump in what he calls "now, at home". This derivation carries out Bob's own accounting explicitly and shows the deficit is filled precisely by the simultaneity shift of equation (6). Let the turnaround occur at home distance $L=v(T/2)$ and home time $T/2$ ; examine Bob's reckoning of Alice's current age just before and just after he reverses velocity.

Step 1. On each leg Bob's own elapsed time is $T/2\gamma$ , and in his instantaneous inertial frame Alice moves at speed $v$ , so he judges her clock to advance by $1/\gamma$ of his: $\Delta\tau_{\text{Alice, per leg}}=(1/\gamma)(T/2\gamma)=T/2\gamma^2$ . Why: within a single inertial frame the reciprocal time-dilation rule is exact, and Bob applies it legitimately on each leg — this is the step the naive reasoner stops at.

Step 2. Adding both legs, Bob's naive tally of Alice's ageing is $2\times T/2\gamma^2=T/\gamma^2$ , which for $v=0.8c$ is only $0.36\,T$ — far short of the $T$ Alice actually logs. Why: summing the two legs as if they belonged to one frame double-counts nothing but silently omits what happens at the seam between them.

Step 3. At the turnaround Bob's line of simultaneity at home swings forward by $\Delta t_{\text{jump}}=2vL/c^2=v^2T/c^2$ (equation 6); adding it, $T/\gamma^2+v^2T/c^2=T(1/\gamma^2+v^2/c^2)=T$ , since $1/\gamma^2=1-v^2/c^2$ . Why: the two terms sum to Alice's full age exactly — the block of her history Bob's now-line sweeps across in the instant of reversal is the $v^2T/c^2$ his leg-by-leg count left out.

Step 4. Conclude that there is no contradiction. Each reciprocal claim is valid inside its own inertial frame, but Bob uses two frames while Alice uses one, and the relativity of simultaneity — not the magnitude of the acceleration — restores the missing time. Why: the asymmetry is which twin's worldline is straight, made quantitative by a simultaneity jump rather than by any dynamical property of the turnaround. Verified: at $v=0.8c$ , $1/\gamma^2=0.36$ , $v^2/c^2=0.64$ , sum $=1.00$ ; Bob's own elapsed time is $0.6\,T$ as in the first derivation.

Limiting cases

The differential ageing depends only on the speed profile, and reduces to familiar regimes:

Case	Result	Meaning
Non-relativistic limit	$\dfrac{\Delta\tau_{\text{Bob}}}{\Delta\tau_{\text{Alice}}}=\sqrt{1-\dfrac{v^2}{c^2}}\approx1-\dfrac{v^2}{2c^2}$	For $v\ll c$ the effect is tiny but nonzero — this is the term measured by airborne atomic clocks and corrected for in GPS.
Ultra-relativistic limit	$\Delta\tau_{\text{Bob}}\to0$ as $v\to c$	The traveller's proper time along the journey vanishes; a photon experiences no elapsed time at all, its worldline being null rather than timelike.
Instantaneous turnaround	full effect, $\Delta\tau_{\text{Bob}}=T/\gamma$	Infinite acceleration for zero duration contributes zero proper time yet gives the entire age difference — proof the effect is kinematic, not dynamical.
No turnaround (both inertial)	symmetric: neither ages less	If the twins drift apart forever and never reunite, "who is younger" is frame-dependent — there is no invariant comparison and the paradox dissolves.

Common trapThe travelling twin ages less because he accelerated, and acceleration slows clocks.

This is the single most common wrong resolution, and it fails a simple test. Acceleration is what breaks the symmetry — it is the physical marker distinguishing the bent worldline from the straight one — but it does not set the size of the age difference. An idealized instantaneous turnaround applies infinite acceleration for zero time and contributes exactly zero proper time, yet delivers the full effect $\Delta\tau_{\text{Bob}}=T/\gamma$ . The amount of differential ageing is fixed by the speeds and durations of the inertial legs, through $\int\mathrm{d}t/\gamma$ , not by how many g's the traveller pulls at the corner. Two travellers with wildly different turnaround accelerations but identical cruise speeds and leg times return exactly the same age. The acceleration is necessary to make the scenario well-posed (someone must turn around) but the lost time is geometric — the length difference of two worldlines — not a force acting on a clock.

Worked examples

1 · A round trip to a star, cross-checked by length contraction (confidence builder)

A star sits 4 light-years from Earth in Earth's rest frame. Alice stays home; Bob flies out at $v=0.8c$ , turns around instantly, and returns at $0.8c$ . How long does the trip take in Earth time, how much does each twin age, and does the traveller's own frame give a consistent answer?

Step 1. In Earth's frame the one-way coordinate time is distance over speed, $t_{\text{one-way}}=4\,\text{ly}/0.8c=5$ yr, so the round trip takes $T=10$ yr. Alice is inertial and at rest, so her proper time equals coordinate time: $\Delta\tau_{\text{Alice}}=\mathbf{10\ \text{yr}}$ .

Step 2. Bob carries $\gamma=1/\sqrt{1-0.64}=1/0.6=5/3$ , so by equation (5) $\Delta\tau_{\text{Bob}}=T/\gamma=10\times0.6=\mathbf{6\ \text{yr}}$ . On reunion Bob has aged 6 years against Alice's 10 — younger by 4 years.

Step 3. Cross-check from Bob's cabin. In his outbound frame the Earth–star distance is length-contracted to $L/\gamma=4/(5/3)=2.4$ ly, which he covers at $0.8c$ in $2.4/0.8=3$ yr; the inbound leg is another 3 yr, totalling 6 yr of his own time — identical to Step 2. The two frames disagree about distance and about simultaneity, but they agree exactly on the invariant each clock reads. T = 10 yr, Alice 10 yr, Bob 6 yr, traveller younger by 4 yr — and length contraction reproduces Bob's 6 yr.

2 · Where the missing time hides: the simultaneity ledger

Take the same $v=0.8c$ trip with $T=10$ yr, but now do Bob's naive reciprocal accounting from Derivation C and show explicitly that it recovers Alice's full age only once the turnaround jump is added.

Step 1. Bob's own elapsed time is $T/\gamma=6$ yr, split as 3 yr per leg. On each leg he judges Alice's clock to run slow by $1/\gamma=0.6$ , so per leg he counts $3\times0.6=1.8$ yr of Alice's ageing; over both legs that is $2\times1.8=T/\gamma^2=\mathbf{3.6\ \text{yr}}$ .

Step 2. That leaves $10-3.6=6.4$ yr of Alice's life unaccounted for. Equation (6) supplies exactly this: the simultaneity jump is $\Delta t_{\text{jump}}=v^2T/c^2=0.64\times10=\mathbf{6.4\ \text{yr}}$ , the block of Alice's history Bob's now-line sweeps across as he reverses.

Step 3. Add them: $3.6+6.4=\mathbf{10\ \text{yr}}$ , Alice's full age, recovered with no appeal to acceleration magnitude. The reciprocal "each sees the other slow" claim was never wrong — it was simply incomplete, because Bob's two legs live in two different frames whose "now, at home" surfaces do not line up. Naive legs 3.6 yr + simultaneity jump 6.4 yr = Alice's 10 yr.

3 · The relativistic rocket to a distant star

Instead of an instantaneous turnaround, let a ship accelerate at a constant proper acceleration $g=9.8\ \mathrm{m/s^2}$ (comfortable Earth gravity) all the way out to a star $d=10$ light-years away. Using the proper-time integral for constant proper acceleration, $\tau=(c/g)\,\mathrm{arccosh}(1+gd/c^2)$ , find the traveller's elapsed time and compare it with Earth's.

Step 1. Form the dimensionless argument. With $c=2.998\times10^8$ m/s, $g=9.8\ \mathrm{m/s^2}$ , and $d=10\times9.461\times10^{15}$ m, the group $gd/c^2=10.32$ , so $1+gd/c^2=\mathbf{11.32}$ . The natural time unit is $c/g=3.06\times10^7\ \mathrm{s}=0.969$ yr.

Step 2. Evaluate the traveller's proper time: $\tau=(c/g)\,\mathrm{arccosh}(11.32)=0.969\times3.12=\mathbf{3.0\ \text{yr}}$ . A 10-light-year crossing costs the crew only about three years of life, because their clock rate collapses as $\gamma$ climbs during the burn.

Step 3. Earth's coordinate time follows from $t=(c/g)\sinh(g\tau/c)=0.969\times\sinh(3.12)=0.969\times11.3=\mathbf{10.9\ \text{yr}}$ — just over the 10 years light itself would need, as it must be. The one-way age gap is already $10.9-3.0\approx8$ yr; a full there-and-back journey with a symmetric deceleration and return would roughly quadruple both figures, a literal twin paradox with the turnaround spread smoothly over the whole trip rather than concentrated at a corner. Crew ≈ 3.0 yr, Earth ≈ 10.9 yr for a one-way 10-ly cruise at 1 g.

Physics of the portrait

The twin paradox is not a discovery so much as a provocation — the sharpest way anyone found to force relativity's strangeness into the open. Einstein noted the bare clock effect in 1905; the memorable story, with a human voyager who leaves and returns to find the world aged, and with acceleration named as the asymmetry, belongs to one French physicist.

Origins · 2 minPaul Langevin and how this was discovered

Paul Langevin

1872–1946 · the travelling-twin thought experiment and relativistic time

Paul Langevin was the physicist who turned a line of Einstein's 1905 kinematics into the story everyone now argues about. In his 1911 essay L'évolution de l'espace et du temps, delivered at the Congress of Bologna, he imagined a traveller shot out to a distant star in a projectile at nearly light speed and brought back, returning to find that only two years had passed for him while two centuries had passed on Earth — and he put his finger squarely on the resolution, insisting that the traveller's acceleration is what makes the two histories physically distinct, so there is no genuine reciprocity. It was Langevin, more than anyone, who made the effect vivid and who framed the debate that Max von Laue (1913) and Einstein himself (1918) would sharpen. Langevin was a formidable physicist in his own right well beyond this thought experiment: a student of Pierre Curie, he built the modern theory of paramagnetism and diamagnetism from the statistics of molecular magnets, and during the First World War he pioneered piezoelectric echo-ranging — the direct ancestor of sonar — to hunt submarines. He was the leading champion of relativity in France, a close associate of the Curies, and a courageous public figure: an outspoken antifascist, he was arrested and held under house arrest by the Vichy authorities during the occupation, saw his son-in-law executed and his daughter deported, escaped to Switzerland in 1944, and returned in triumph after the Liberation. The travelling twin, though, is the image that carries his name into every relativity course.

History

1905

— Einstein's special relativity shows a clock carried around a closed loop lags one left at rest

1911

— Langevin frames the travelling-twin story and pins the asymmetry on the traveller's acceleration

1913

— von Laue stresses that the traveller occupies two inertial frames while the stay-at-home occupies one

1918

— Einstein answers the objection in a published dialogue, reading the turnaround as a pseudo-gravitational field

1971

— Hafele and Keating fly caesium clocks around the world and recover the predicted nanosecond split

1977

— the CERN muon storage ring confirms lifetime dilation at γ ≈ 29 to better than one percent

Going further

Texts. Misner, Thorne & Wheeler, Gravitation (1973), §1.5 · Wald, General Relativity (1984), §9.3 · Carroll, Spacetime and Geometry (2004), §1.7 · Hartle, Gravity (2003), ch. 5 · Schutz, A First Course in General Relativity (2009), §1.13 · Taylor & Wheeler, Spacetime Physics (1992), ch. 4 · Hafele & Keating, Science 177, 166 (1972).

Provenance. The proper-time functional $c\,\tau=\int\sqrt{-g_{\mu\nu}\mathrm{d}x^\mu\mathrm{d}x^\nu}$ and the increment $\mathrm{d}\tau=\mathrm{d}t/\gamma$ are recalled by name from aq-gr-proper-time; the simultaneity jump uses the relativity of simultaneity owned by aq-gr-relativity-of-simultaneity; the length-contraction cross-check in Worked Example 1 is owned by aq-gr-length-contraction; the maximal-proper-time theorem is the timelike-geodesic property of aq-gr-spacetime-interval and Minkowski geometry. All results are in signature $-+++$ . A verifier should confirm: (1) $\mathrm{d}\tau=\sqrt{1-v^2/c^2}\,\mathrm{d}t$ from factoring the line element (Schutz §1.13, Carroll §1.7); (2) the worked numbers — $v=0.8c\Rightarrow\gamma=5/3,\,1/\gamma=0.6$ ; $v=0.6c\Rightarrow1/\gamma=0.8$ ; (3) the simultaneity identity $1/\gamma^2+v^2/c^2=1$ and the jump $v^2T/c^2=6.4$ yr at $v=0.8c,\,T=10$ yr; (4) the length-contraction cross-check $L/\gamma=2.4$ ly giving Bob 6 yr; (5) the relativistic-rocket values $\tau=(c/g)\,\mathrm{arccosh}(1+gd/c^2)\approx3.0$ yr and Earth $t\approx10.9$ yr for $g=9.8\ \mathrm{m/s^2}$ , $d=10$ ly; (6) the GPS split (SR $\approx-7$ , GR $\approx+45$ , net $\approx+38$ μs/day) and the muon rest lifetime 2.2 μs against standard references. The Hafele–Keating result is reported qualitatively; see Hafele & Keating, Science 177, 166 (1972).

Prerequisites. Proper Time · Time Dilation · Relativity of Simultaneity · Spacetime Interval · Lorentz Transformations. Leads to. Minkowski Spacetime · Causal Structure and Light Cones · Four-Vectors.

Ask the tutor: ask me to work any twin trip you like — pick the speed and distance and I will age both twins, split the time between the inertial legs and the simultaneity jump, and cross-check it from the traveller's own frame.

If motion is relative, why isn't the paradox symmetric?

Because only one twin changes inertial frames. Alice stays in a single inertial frame for the whole trip; Bob uses one frame going out and a different one coming back, joined by an acceleration at the turnaround. The reciprocal time-dilation rule — "each sees the other's clock run slow" — is exact, but it holds only within one inertial frame, and Bob cannot patch his two legs together as though they were one frame without accounting for the jump in simultaneity at the seam. Their worldlines are not related by a single Lorentz transformation, so there is no symmetry to enforce equal ageing. The asymmetry is a genuine geometric fact: one worldline is straight and one is bent.

Does the acceleration itself cause the time difference?

No, and this is the sharpest trap in the subject. Acceleration is necessary to break the symmetry — someone has to turn around, and turning around means accelerating — but it does not set the amount of differential ageing. That amount is fixed entirely by the speeds and durations of the inertial legs through $\Delta\tau=\int\mathrm{d}t/\gamma$ . An idealized instantaneous turnaround applies infinite acceleration for zero time, contributes zero proper time, and still yields the full effect. Two travellers who cruise at the same speed for the same time but turn around with very different accelerations return exactly the same age. The lost time is the length difference between two worldlines, a geometric quantity, not the result of a force pressing on a clock.

Why is the straight worldline the longest, not the shortest?

Because the spacetime metric carries a minus sign on the time part relative to the space parts, so the geometry is Lorentzian, not Euclidean. In ordinary geometry the length element adds spatial pieces, $\mathrm{d}\ell^2=\mathrm{d}x^2+\mathrm{d}y^2$ , and detours make paths longer, so the straight line is shortest. In spacetime the proper-time element subtracts the spatial pieces, $\mathrm{d}\tau^2=\mathrm{d}t^2-\mathrm{d}x^2/c^2$ , so any spatial wandering reduces the accumulated proper time. The straight timelike worldline, which does no wandering, therefore accumulates the most. This single sign flip is why the stay-at-home twin ages most and why the intuition imported from road maps points exactly backwards.

What actually happens at the turnaround?

In the idealized problem the turnaround is instantaneous and contributes no proper time to Bob at all, yet it is where the resolution lives. As Bob's velocity reverses, his surface of simultaneity — the events he calls "now, back home" — swings around like a searchlight, and the home time it points at leaps forward by $v^2T/c^2$ . That jump is precisely the stretch of Alice's history that Bob's leg-by-leg reciprocal count omitted. If you smear the turnaround over a finite time, nothing essential changes: Bob's now-line sweeps forward smoothly instead of jumping, and integrating the sweep reproduces the same total. The corner is where the two inertial frames are stitched together, and the stitch is a simultaneity shift, not a burst of extra ageing.

Has the twin paradox actually been tested?

Directly and repeatedly. In 1971 Hafele and Keating flew caesium atomic clocks around the world on commercial airliners and found them offset from a ground clock by the predicted tens to hundreds of nanoseconds, with the sign depending on flight direction. The Global Positioning System is a permanent, operational version: its satellite clocks run a net 38 microseconds per day fast relative to the ground and are corrected continuously, or navigation would drift kilometres a day. Cosmic-ray muons reach sea level only because their proper time runs short along their fast worldlines, and the CERN muon storage ring confirmed the lifetime dilation to better than a percent at $\gamma\approx29$ . The travelling twin is one of the most thoroughly verified predictions in physics.

Could a real astronaut come back younger than people on Earth?

Yes, in principle without limit, and in tiny amounts already in fact. Cosmonaut Sergei Krikalev, from long stays on Mir and the ISS, has aged a few thousandths of a second less than he would have on the ground — a real, if minuscule, twin effect dominated by orbital speed. Scaling up is only an engineering problem, not a physics one: a ship that could sustain $1\,g$ of proper acceleration would let its crew cross tens of light-years within their own lifetimes while centuries passed at home, exactly as the relativistic-rocket calculation shows. The catch is entirely practical — the energy and reaction mass required are astronomical — but the relativistic bookkeeping is identical to the simple round-trip integral, just with a smoothly varying speed.