The twin paradox is not a paradox Image: NASA, public domain

The twin paradox is not a paradox

Two twins, both aged 20, say goodbye. One stays on Earth. The other boards a ship that leaves at \(0.6c\), flies to a star six light years away, turns round at once, and comes back at the same speed. When the ship lands, the Earth twin is 40 and the ship twin is 36. There is no disagreement between them about who is older, and there is no paradox. The problem has a reputation because of a tempting wrong argument, and the way to see through it is to work out the numbers.

The Earth twin's bookkeeping

In the Earth frame the trip is easy. Six light years at \(0.6c\) takes 10 years each way, so the ship is away for 20 years. The Lorentz factor at \(0.6c\) is

\[ \gamma = \frac{1}{\sqrt{1 - 0.36}} = 1.25, \]

and a moving clock runs slow by this factor, so the ship's clock advances 8 years on the way out and 8 years on the way back. The travelling twin comes home having aged 16 years. So far this is ordinary time dilation, the subject of the light clock derivation.

The wrong argument

Now the tempting step. From the ship, the Earth is the thing that moves at \(0.6c\), first away and then towards. The Earth clock should run slow by the same factor 1.25. The ship twin ages 16 years, so the Earth twin should age \(16/1.25 = 12.8\) years. Each twin claims the other is younger, and when they meet only one of them can be right. That is the supposed paradox.

The error is the phrase "from the ship". There is no single inertial frame in which the ship is at rest for the whole trip. There is one frame for the outbound leg and a different one for the return leg, and the ship twin changes from one to the other at the star. The Earth twin stays in one inertial frame throughout. The situation is not symmetric, and the arithmetic of time dilation, which is a statement about one inertial frame looking at a clock moving uniformly through it, cannot simply be applied by the ship twin to the whole trip.

What the turnaround does

The clean way to see this is to ask, at each moment of the ship's trip, what time it is on Earth "now" according to the ship. The word "now" for events far away depends on the frame, which is the relativity of simultaneity, the main subject of Unit 5.

On the outbound leg the ship's frame does see the Earth clock running slow. When the ship's clock reads 8 years, at the star, the ship's frame says the Earth clock reads \(8/1.25 = 6.4\) years. On the return leg the ship's frame again sees the Earth clock running slow, and in the 8 years of the return the Earth clock advances another 6.4 years, to a total of 12.8 at landing. That is the number the wrong argument produced, and it leaves out one thing.

The return frame does not agree with the outbound frame about what time it is on Earth at the moment of turnaround. In the outbound frame, the Earth event simultaneous with the turnaround is at Earth time 6.4 years. In the return frame, by the symmetry of the two legs, the Earth event simultaneous with the turnaround is at Earth time \(20 - 6.4 = 13.6\) years. When the ship changes frames at the star, its definition of "now on Earth" jumps forward by 7.2 years. Nothing happens on Earth in that moment; what changes is which distant events the ship calls simultaneous with itself. Add the three pieces, \(6.4 + 7.2 + 6.4\), and the ship's own accounting gives 20 years for the Earth clock, which is what the Earth twin said all along.

On a spacetime diagram the picture is one straight vertical line for the Earth, a bent line for the ship, and lines of simultaneity through the ship that tilt one way on the outbound leg and the other way on the return. At the corner, the tilt flips and a wedge of the Earth's world line, 7.2 years of it, is skipped over. The longer the trip, the bigger the wedge.

A check with light signals

If the lines of simultaneity feel like a bookkeeping device, there is a check that uses only things each twin can see. Let each twin send a flash of light on every birthday. A flash sent by someone moving away at \(0.6c\) arrives at a rate reduced by the relativistic Doppler factor \(\sqrt{(1-0.6)/(1+0.6)} = 0.5\), and from someone approaching at \(0.6c\) at a rate increased by the factor 2.

The ship twin, during the 8 years outbound, receives Earth flashes at one every two years: 4 flashes. During the 8 years back, at two a year: 16 flashes. Total 20, the Earth twin's age at landing minus 20. The Earth twin receives the ship's flashes at one every two years until the light from the turnaround arrives, which happens 10 years after departure plus 6 years for the light to come back, at Earth year 16: 8 flashes. Then, for the final 4 years, two a year: 8 more. Total 16. Each twin counts the other's birthdays by direct observation and gets the right answer. The asymmetry is visible in the counting: the ship twin sees the switch from slow to fast at the halfway point of their own time, the Earth twin sees it only after 16 of their 20 years.

Acceleration and the real experiment

Does the ship twin's acceleration at the turnaround cause the difference? It is what breaks the symmetry, and in that sense yes, but the amount of ageing does not depend on how hard the brakes are applied. A ship that turns round in a second and one that turns round in a day age almost identically; the difference comes from the length of the two legs. Proper time is the length of a world line in spacetime, measured with the interval, and a bent line between two events is shorter than the straight one. Unit 6 of the masterclass computes the proper time along a smoothly accelerated round trip and shows that the straight line always wins.

The experiment has been done, with atomic clocks instead of twins. In October 1971 Joseph Hafele and Richard Keating flew four caesium clocks around the world on commercial airliners and compared them with clocks that stayed at the United States Naval Observatory. The flying clocks lost or gained tens of nanoseconds, in agreement with a prediction that combined special relativity with the gravitational effect of altitude. That story has its own article. The muons that reach the ground, the clocks on the GPS satellites, and the particle lifetimes measured in every accelerator laboratory are the same physics, and none of them has ever shown a sign of a paradox.

Study relativity properly

The ToR Masterclass: weekly units, quizzes, a final exam, and a certificate.
About the program