Proper Time

Send a traveller to a star $4$ light-years away at $\beta = 0.8$ and bring her home. Earth's calendars will have turned over $10$ years. Her wristwatch will read $6$ . Not because the watch malfunctioned, and not because anyone measured it badly — she really did live $6$ years while her twin lived $10$ , and every observer in the universe, whatever coordinates they use, computes the same $6$ . That number has a name: it is the proper time $\tau$ along her worldline, the time a clock reads when carried between two events. It is not a coordinate. Coordinate time depends on who is watching; proper time belongs to the path, and it is an invariant — built from the spacetime interval, $c^{2}\mathrm{d}\tau^{2} = c^{2}\mathrm{d}t^{2} - \mathrm{d}\mathbf{x}^{2}$ , which every inertial frame agrees on. Along a worldline of speed $\beta(t)$ it accumulates as $\tau = \int\sqrt{1-\beta^{2}(t)}\,\mathrm{d}t = \int \mathrm{d}t/\gamma$ . And now the theorem this article exists to prove. Among all timelike worldlines joining two given events, the inertial one carries the greatest proper time. Every detour into space costs you time on your own clock. In Euclid's geometry the straight line is the shortest path between two points; in spacetime, with its minus sign, the straight worldline is the longest. That single reversal is the whole content of the twin paradox — and The Twin Paradox is where the drama belongs. Here we prove the theorem.

Key takeaways

Proper time $\tau$ is the time elapsed on a clock carried along a worldline. It is a Lorentz invariant: every inertial frame, using its own coordinates, computes the same value.
It comes straight from the interval: $c^{2}\mathrm{d}\tau^{2} = c^{2}\mathrm{d}t^{2} - \mathrm{d}\mathbf{x}^{2}$ , so $\mathrm{d}\tau = \sqrt{1-\beta^{2}}\,\mathrm{d}t = \mathrm{d}t/\gamma$ and $\tau = \int \mathrm{d}t/\gamma$ along any path.
Time dilation is the special case of a straight worldline: $\Delta t = \gamma\,\Delta\tau$ .
The maximal-proper-time theorem: among all timelike worldlines joining two fixed events, the inertial one has the GREATEST proper time. Proof: work in the frame where the events share a place; then $\tau = \int\sqrt{1-\beta^{2}}\,\mathrm{d}t \leq \int\mathrm{d}t = \Delta t$ , with equality only if $\beta \equiv 0$ — which is exactly the inertial worldline.
The integrand depends on speed alone, never on acceleration. That is the clock hypothesis, and it is an empirical claim: CERN's muons survive $64.4\ \mu\mathrm{s}$ in the lab at $\gamma = 29.33$ while enduring about $10^{18}g$ .
Key idea: proper time is the invariant arc length of a worldline — the reading on the traveller's own watch. In spacetime the straight path is the LONGEST, the exact reverse of Euclid's triangle inequality, and that reversal is why the travelling twin comes home young.

The core insight

A wristwatch knows nothing of frames, $\gamma$ factors or simultaneity conventions. It simply ticks, and the number it shows between two events on its own worldline is the proper time between them — a quantity every observer computes identically from their own coordinates. Because the metric carries a minus sign, wandering off in space reduces what the watch accumulates. Straight is longest. The traveller who leaves and returns has taken a shortcut through time, and pays for it in years.

Where this lives today

Proper time is now something engineers budget for. Every GPS satellite carries an atomic clock whose accumulated proper time is what the receiver in your pocket ultimately reads: at an orbital speed of $3.874\ \mathrm{km/s}$ the special-relativistic term alone is $\gamma - 1 = 8.35\times10^{-11}$ , which over a day is $8.35\times10^{-11} \times 86\,400\ \mathrm{s} = 7.2\ \mu\mathrm{s}$ of clock loss. A larger gravitational term of the opposite sign belongs to another article; what matters here is that the SR piece is a proper-time integral along the satellite's worldline, and it is not optional. The clock hypothesis, meanwhile, is still being pushed. The Fermilab Muon $g-2$ experiment stores muons at $\gamma = 29.3$ exactly as CERN did, and its precision measurements depend on the dilated lifetime being $\gamma\tau_0$ with no acceleration correction. Optical clocks now resolve fractional frequency shifts near $10^{-18}$ — fine enough that a clock walked across a room measurably changes its proper time. Wheeler's wristwatch has become laboratory hardware.

The physical picture

Think of proper time as the odometer of a spacetime traveller — but an odometer that runs backwards against detours. Two clocks start together and are reunited later. Between those two events, whichever clock wandered furthest through space has ticked the fewest times. The clock that never moved wins.

Figure

Two worldlines between the SAME two events. The inertial twin sits at x = 0 for 10 years of coordinate time, and her watch reads 10.0 y. The traveller runs out to 4.0 light-years at β = 0.8, turns, and comes back: 5 years out and 5 years back on the coordinate clock, but her watch reads 10/γ = 6.0 y. She returns 4 years younger — and the bent path is the SHORT one in proper time

The two watches disagree because they took different paths, not because either was mistreated. And there is nothing special about $\beta = 0.8$ : sweep the traveller's speed and the proper time she brings home is a smooth function of how hard she detoured. It has a maximum, and the maximum sits exactly where the bend disappears.

Figure

PUNCHLINE — proper time against bentness. For an out-and-back trip fixed to Δt = 10 y of coordinate time, the traveller's watch reads τ = 10√(1−β²): 10.0 y at β = 0, then 9.80, 9.17, 8.0, 6.0, 4.36, and 1.41 y at β = 0.99, falling to zero as β → 1. The maximum is at β = 0 — the straight, inertial worldline. In spacetime the straight path is the LONGEST, the exact reverse of Euclid

The core equations

Everything begins with the interval, the one combination of coordinate differences that every inertial frame agrees on. Signature $(+\,-\,-\,-)$ , so a timelike separation has $s^{2} > 0$ .

c^{2}\,\mathrm{d}\tau^{2} = c^{2}\,\mathrm{d}t^{2} - \mathrm{d}x^{2} - \mathrm{d}y^{2} - \mathrm{d}z^{2} \qquad (1)

where: $\mathrm{d}\tau$ — the proper time along an infinitesimal step of the worldline; $(\mathrm{d}t, \mathrm{d}\mathbf{x})$ — the coordinate separation of its ends in some inertial frame; $c = 299\,792\,458\ \mathrm{m/s}$ . Physical reading: in the clock's own momentary rest frame $\mathrm{d}\mathbf{x} = 0$ , so $\mathrm{d}\tau$ is literally what that clock reads — and because the left-hand side is invariant, any frame's coordinates give the same number. Divide through by $\mathrm{d}t^{2}$ and the working formula appears.

\mathrm{d}\tau = \sqrt{1-\beta^{2}(t)}\;\mathrm{d}t = \frac{\mathrm{d}t}{\gamma(t)}, \qquad \tau = \int_{A}^{B}\sqrt{1-\beta^{2}(t)}\;\mathrm{d}t \qquad (2)

where: $\beta(t) = |\mathrm{d}\mathbf{x}/\mathrm{d}t|/c$ — the instantaneous speed in the chosen frame; $\gamma = 1/\sqrt{1-\beta^{2}}$ . Physical reading: the integrand is at most $1$ , and equals $1$ only at rest, so a moving clock always accumulates less than the coordinate time — moment by moment, locally. Nothing requires the motion to be uniform: the integral handles any timelike path. For a straight worldline $\beta$ is constant and the integral collapses. Time dilation is a corollary, not a separate law.

\Delta\tau = \frac{\Delta t}{\gamma} \quad\Longleftrightarrow\quad \Delta t = \gamma\,\Delta\tau, \qquad c\,\Delta\tau = \sqrt{c^{2}\Delta t^{2} - \Delta x^{2}} \qquad (3)

where: $\Delta t, \Delta x$ — the coordinate separation of the two events; $\Delta\tau$ — the reading of a clock travelling inertially between them. Physical reading: the second form is the useful one, needing no frame at all — hand it any frame's $(\Delta t, \Delta x)$ and it returns the same $\Delta\tau$ . Time Dilation and The Light Clock own the derivation of $\gamma$ ; this article owns what $\tau$ is. Finally, the claim that lets equation (2) be applied to a rocket, a storage ring, or a twin who turns around.

\frac{\mathrm{d}\tau}{\mathrm{d}t} = \sqrt{1-\beta^{2}(t)} \qquad\text{— and nothing else: no dependence on } \dot\beta \qquad (4)

where: $\dot\beta$ — the acceleration. Physical reading: this is the clock hypothesis — not a theorem of special relativity but an empirical claim about real clocks: an ideal clock's rate depends on its instantaneous speed alone. It is why an accelerated worldline can be chopped into inertial slivers and integrated, and it has been tested to about $10^{18}g$ .

Where the results come from

Derivation AProper time is an invariant — and it does not care that the clock accelerates

The goal is to show that $\mathrm{d}\tau$ deserves the name "the time the clock reads", and that its value is the same computed in any inertial frame. Two moves do it. First, in the clock's instantaneous rest frame equation (1) collapses to $\mathrm{d}\tau = \mathrm{d}t'$ — the clock's own tick, by definition. Second, the interval is invariant under boosts, so every other frame's coordinates return that same number. The second move is where the algebra lives, and the cross terms are the whole story.

Step 1. Anchor the meaning. At each moment go to the inertial frame momentarily co-moving with the clock, where $\mathrm{d}x' = \mathrm{d}y' = \mathrm{d}z' = 0$ . Equation (1) then gives $c^{2}\mathrm{d}\tau^{2} = c^{2}\mathrm{d}t'^{2}$ , so $\mathrm{d}\tau = \mathrm{d}t'$ . Why this definition is worth making: it identifies an abstract algebraic combination with a physical reading on a physical dial.

Step 2. Boost, and watch the cross terms die. With $\mathrm{d}t' = \gamma(\mathrm{d}t - V\mathrm{d}x/c^{2})$ and $\mathrm{d}x' = \gamma(\mathrm{d}x - V\mathrm{d}t)$ , expand $c^{2}\mathrm{d}t'^{2} - \mathrm{d}x'^{2}$ . The $\mathrm{d}t\,\mathrm{d}x$ terms come with coefficient $\gamma^{2}(-2V + 2V) = 0$ and cancel exactly, leaving $\gamma^{2}(1 - V^{2}/c^{2})(c^{2}\mathrm{d}t^{2} - \mathrm{d}x^{2}) = c^{2}\mathrm{d}t^{2} - \mathrm{d}x^{2}$ . Why the cancellation is the point: without it $\mathrm{d}\tau$ would depend on the frame, and a wristwatch reading would be a matter of opinion.

Step 3. Integrate, and notice what was never assumed. Proper time along a whole worldline is $\tau = \int\mathrm{d}\tau$ , a sum of invariants and hence itself an invariant: every inertial observer, using their own coordinates, computes the same total. Nowhere did the clock have to move uniformly — Step 1 used only its instantaneous velocity. That the real rate depends on speed alone, never on $\dot\beta$ , is the clock hypothesis of equation (4): an empirical claim, and a strong one, since a clock in a storage ring is wrenched around at $\sim 10^{18}g$ . Why it must be said aloud: it is the bridge from the inertial theory to bent worldlines, and without it the twin's turnaround could not be handled at all. *As a concrete check: boosting $(\mathrm{d}t, \mathrm{d}x)$ by an arbitrary $V$ , expanding $c^{2}\mathrm{d}t'^{2} - \mathrm{d}x'^{2}$ , and collecting terms shows the $\mathrm{d}t\,\mathrm{d}x$ coefficient is exactly $0$ and the result equals $c^{2}\mathrm{d}t^{2} - \mathrm{d}x^{2}$ identically in $V$ . Numerically, the interval $\Delta t = 5\ \mathrm{s}$ , $\Delta x = 3$ light-seconds gives $\tau = 4\ \mathrm{s}$ in every boosted frame.*

Go deeper · Derivation B · 2 minTHE THEOREM — the inertial worldline has the greatest proper time

Derivation BTHE THEOREM — the inertial worldline has the greatest proper time

Here is the result the article exists for, and it is sharper than "the traveller ages less". Take any two timelike-separated events and any two timelike worldlines joining them, one inertial and one not. Claim: the inertial one accumulates strictly more proper time. The proof takes three lines once you choose the frame well — and the choice is the trick, because for a timelike separation there always exists a frame in which the two events happen at the same place. There the inertial worldline just sits still, and everything else is a detour.

Step 1. Choose the frame: the events are timelike-separated, so a frame exists in which they occur at the same spatial point, separated only by $\Delta t$ in time. It is the rest frame of the inertial worldline joining them — which is exactly why that worldline is the natural benchmark. Why this is legitimate rather than a dodge: proper time is an invariant, so whatever we prove in this frame holds in every frame.

Step 2. Compute both. The inertial path has $\beta \equiv 0$ here, so $\tau_{\text{inertial}} = \int_{0}^{\Delta t}\!\mathrm{d}t = \Delta t$ . Any other timelike path from $A$ to $B$ must have $\beta(t) \neq 0$ somewhere — it has to leave and come back — and gives $\tau = \int_{0}^{\Delta t}\!\sqrt{1-\beta^{2}(t)}\;\mathrm{d}t$ . Why the integrals share limits: both paths join the same events, so both take exactly $\Delta t$ of this frame's coordinate time.

Step 3. Apply the pointwise inequality. For every $t$ , $\sqrt{1-\beta^{2}(t)} \leq 1$ , with equality iff $\beta(t) = 0$ . Integrating preserves it: $\tau \leq \Delta t = \tau_{\text{inertial}}$ , and the inequality is strict unless $\beta \equiv 0$ everywhere — that is, unless the path is the inertial one. $\boxed{\tau_{\text{inertial}} > \tau_{\text{any other}}}$ . Why this is the entire proof: the minus sign in the metric put $\sqrt{1 - \beta^{2}}$ into the integrand, and that square root is a penalty for moving. There is no way to make it a bonus.

Step 4. Read the geometry. In Euclidean space a detour is longer; here the triangle inequality reverses, and the straight worldline is the longest curve between two timelike-separated events. A traveller who leaves and returns has taken a shortcut through time and comes home young. Why the paradox is not a paradox: the two worldlines are not related by any symmetry — one is straight and one is bent, and "bent" is a frame-independent fact. The Twin Paradox is where that story is told. *The inequality $\sqrt{1-\beta^{2}} \leq 1$ holds with equality only at $\beta = 0$ , and integrating over the constant- $\beta$ family gives $\int_0^{T}\sqrt{1-\beta^2}\,\mathrm{d}t = T\sqrt{1-\beta^2} < T$ for any $\beta > 0$ . Evaluating the straight worldline ( $10.0\ \mathrm{y}$ ) against out-and-back paths at $\beta = 0.2,\ 0.4,\ 0.6,\ 0.8,\ 0.95,\ 0.99$ gives $9.80,\ 9.17,\ 8.00,\ 6.00,\ 3.12,\ 1.41\ \mathrm{y}$ respectively; a three-leg path with different out and back speeds, and a smoothly-varying $\beta(t)$ integrated numerically, likewise come in strictly under $10.0\ \mathrm{y}$ .*

Go deeper · Derivation C · 1 minConstant proper acceleration, and the one variable that stays linear

Derivation CConstant proper acceleration, and the one variable that stays linear

One card, and then this is handed on. If a rocket burns so that its crew feel a constant $a$ — an accelerometer bolted to the floor reads the same value forever — its worldline is a hyperbola and its proper time grows only like a logarithm of coordinate time. The elegant fact underneath is that the rapidity $\theta = \operatorname{artanh}\beta$ is exactly linear in the crew's own clock. The accelerated-frames articles develop the geometry; this card is here because the proper-time integral is what makes any of it computable.

Step 1. Set up. Constant proper acceleration means $\mathrm{d}(\gamma\beta c)/\mathrm{d}t = a$ , so $\gamma\beta c = at$ from rest, giving $1/\gamma = 1/\sqrt{1 + (at/c)^{2}}$ . Why $\gamma\beta$ and not $\beta$ : it is $\gamma\beta$ that grows without bound, which is how the rocket accelerates forever without reaching $c$ .

Step 2. Integrate equation (2): $\tau = \int_{0}^{t}\dfrac{\mathrm{d}t'}{\sqrt{1 + (at'/c)^{2}}} = \boxed{\dfrac{c}{a}\operatorname{arcsinh}\!\left(\dfrac{at}{c}\right)}$ , equivalently $t = (c/a)\sinh(a\tau/c)$ . For $at \gg c$ this is $\tau \approx (c/a)\ln(2at/c)$ — logarithmic, which is why a crew at $a = g$ could cross the galaxy inside a human lifetime of their time. Why the scale is $c/a$ : at $a = g$ that is $0.97$ years, so a year of burn is where relativity takes over.

Step 3. Read off the linear variable. Since $\beta = \tanh(a\tau/c)$ , the rapidity is $\theta = \operatorname{artanh}\beta = a\tau/c$ — exactly linear in proper time. Why that is the right way to see it: rapidities add and speeds do not, so constant proper acceleration is "constant rapidity gain per second of your own clock". Rapidity and the Hyperbolic Form of Boosts develops this. *Direct integration gives $\int \mathrm{d}t/\sqrt{1+(at/c)^2} = (c/a)\operatorname{arcsinh}(at/c)$ ; differentiating confirms $\mathrm{d}\theta/\mathrm{d}\tau = a/c$ exactly; and numerically $c/g = 0.97\ \mathrm{yr}$ .*

Limiting cases

Case	Result	Meaning
At rest, $\beta = 0$	$\mathrm{d}\tau = \mathrm{d}t$	proper time equals coordinate time — and this path is the maximum
Slow, $\beta \ll 1$	$\mathrm{d}\tau \approx (1 - \tfrac12\beta^{2})\,\mathrm{d}t$	second order: a GPS satellite at $3.874\ \mathrm{km/s}$ loses only $7.2\ \mu\mathrm{s}$ per day
Constant $\beta$	$\Delta\tau = \Delta t/\gamma$	ordinary time dilation — the straight-worldline special case
Ultra-relativistic, $\beta \to 1$	$\mathrm{d}\tau \to 0$	at $\beta = 0.99$ a $10$ -year round trip reads $1.41$ years
Lightlike, $\beta = 1$	$\mathrm{d}\tau = 0$ exactly	a photon accumulates no proper time and has no rest frame
Galilean limit, $c \to \infty$	$\mathrm{d}\tau \to \mathrm{d}t$	proper time collapses into absolute time, and every worldline between two events ties

Common trapthe travelling twin ages less BECAUSE of the acceleration

This is the most popular wrong answer in relativity, and it survives because it points at the right asymmetry for the wrong reason. Acceleration has no direct effect on a clock's rate — that is exactly what the clock hypothesis says, and CERN's muons make the point brutally: circulating at $\gamma = 29.33$ under about $10^{18}g$ , they live precisely as long as the naive $\gamma\tau_0$ formula predicts, with no extra term for the wrenching. The rate depends on speed, full stop. What acceleration actually does is more interesting: it is the only way a worldline can be bent — the only way to leave, turn, and come back to the same event. Bending is what costs proper time, and $\int\sqrt{1-\beta^{2}}\,\mathrm{d}t$ never asks how the bending was achieved.

Worked examples

1 · Four years younger

A traveller leaves Earth at $\beta = 0.8$ , flies to a star $4.0$ light-years away, turns around instantly, and comes home. How much does each twin age, and what is the largest possible reading her watch could have shown?

Step 1. Coordinate time. Each leg covers $4.0$ light-years at $0.8c$ , taking $4.0/0.8 = 5.0$ years, so the round trip is $\Delta t = \boxed{10.0\ \mathrm{years}}$ on Earth's clocks. The stay-at-home twin is inertial and at rest here, so her watch also reads $10.0$ years.

Step 2. The traveller's watch. $\gamma = 1/\sqrt{1-0.64} = 1/0.6 = 5/3$ , and her speed is constant on each leg, so equation (2) gives $\tau = \Delta t/\gamma = 10.0 \times 0.6 = \boxed{6.0\ \mathrm{years}}$ . She comes home $\boxed{4.0\ \mathrm{years}}$ younger.

Step 3. Check it with the interval, using no $\gamma$ at all. The outbound leg joins $(t, x) = (0, 0)$ to $(5\ \mathrm{y},\ 4\ \mathrm{ly})$ , so $c\,\Delta\tau = \sqrt{(5)^{2} - (4)^{2}} = \sqrt{9} = 3.0$ years of proper time; the return leg is the mirror image and gives $3.0$ more. Total $\boxed{6.0\ \mathrm{years}}$ — the same answer from the invariant, as it must be.

Step 4. Now bound it. By the theorem, no worldline from departure to reunion can beat the inertial one's $10.0$ years — every route she could have flown, faster, slower or wiggly, comes in under it. At $\beta = 0.6$ she would have brought home $8.0$ years, at $\beta = 0.95$ only $3.12$ , at $\beta = 0.99$ just $1.41$ .

2 · The muon that survives $10^{18}g$

In the CERN muon storage ring (Bailey et al., 1977) muons circulate at $\gamma = 29.33$ in a ring of radius $\approx 7\ \mathrm{m}$ . Their rest lifetime is $\tau_0 = 2.197\ \mu\mathrm{s}$ . Predict the lab-frame lifetime, and compute the proper acceleration they endure.

Step 1. The prediction is pure proper time: the muon decays after $\tau_0$ on its own clock, and the lab watches that clock through equation (3). So the lab-frame lifetime is $\gamma\tau_0 = 29.33 \times 2.197 = \boxed{64.4\ \mu\mathrm{s}}$ . Bailey et al. measured $\tau^{+} = 64.419 \pm 0.058\ \mu\mathrm{s}$ .

Step 2. Now the acceleration. At $\gamma = 29.33$ the speed is $\beta = \sqrt{1 - 1/\gamma^{2}} = 0.99942$ , and for circular motion the proper acceleration is $a = \gamma^{2}v^{2}/r = (29.33)^{2}(0.99942c)^{2}/7.0\ \mathrm{m} = \boxed{1.1\times10^{19}\ \mathrm{m/s^{2}}}$ , which is $\boxed{1.1\times10^{18}\,g}$ .

Step 3. Read the verdict. The measured lifetime matches $\gamma\tau_0$ — a formula containing no acceleration term whatsoever — to a fractional precision of about $2\times10^{-3}$ . A clock flung around at $10^{18}g$ keeps time by its speed alone, so equation (4) is not an assumption of convenience but a measured fact.

Step 4. Note the scale. Everyday accelerations are of order $g$ and laboratory centrifuges reach $10^{6}g$ ; this tests the hypothesis twelve orders of magnitude beyond that, and it holds.

Physics of the portrait

The portrait here is a quantity, not a person — and the reason is instructive. Proper time was named in 1908, proved maximal in 1911, measured in 1963, and only then made the foundation of the subject. No single physicist owns it; what it has instead is a biography.

Origins · 2 minProper time and how this was discovered

Proper time

1908–1992 · from Minkowski's Eigenzeit to the wristwatch reading

Hermann Minkowski gave the quantity its name — Eigenzeit, "own time" — in 1908, in the paper that recast relativity as four-dimensional geometry. Almost at once it acquired its central theorem: in 1911 Max von Laue, working in Minkowski's language, showed that among all timelike worldlines joining two events the straight one carries the greatest proper time — the reverse of Euclid's triangle inequality, and the whole content of the twin story Paul Langevin told at Bologna that same year. For half a century it remained a quantity in a formalism. Then Frisch and Smith (1963) counted muons falling down Mount Washington and demonstrated that a particle's decay clock reads its own proper time and nothing else — the integral $\int\mathrm{d}\tau$ became a thing you could photograph. What remained was to make it the starting point rather than a consequence, and that was the work of Edwin Taylor and John Archibald Wheeler, whose Spacetime Physics (first edition 1966) refused to open with $\gamma$ factors and moving rods, and built the subject on the invariant interval and the time a clock reads along its own worldline. Their name for that reading — "wristwatch time", which appears in the second edition (1992) — is the best three-word definition anyone has written: not a coordinate, not a convention, not a frame's opinion, but the number on the dial of the watch that actually went. The moral of the biography is the moral of the article: when a quantity is invariant, build the theory on it and the paradoxes never arise.

History

1908

— Minkowski introduces Eigenzeit, proper time, in "Die Grundgleichungen für die elektromagnetischen Vorgänge in bewegten Körpern"; his Cologne lecture Raum und Zeit, delivered on 21 September, makes the spacetime geometry public.

1911

— Langevin, lecturing at Bologna, tells the story of the traveller who returns younger than the twin who stayed.

1911

— Von Laue, using Minkowski's geometry, shows that among all timelike worldlines joining two events the straight one carries the greatest proper time.

1963

— Frisch and Smith measure the survival of cosmic-ray muons and confirm that the decay clock is the muon's own proper time.

1966

— Taylor and Wheeler's Spacetime Physics rebuilds the subject around the invariant interval and the clock carried along a worldline.

1977

— Bailey and co-workers measure muon lifetimes in the CERN storage ring at gamma = 29.33, finding 64.4 microseconds and confirming the clock hypothesis under about 10¹⁸ g.

1992

— The second edition of Spacetime Physics names the interval read by a carried clock the "wristwatch time".

Going further

Texts. Taylor and Wheeler, Spacetime Physics, 2nd ed., Chs. 1–4 (wristwatch time) · Rindler, Introduction to Special Relativity, 2nd ed., §§10–14 · Griffiths, Introduction to Electrodynamics, 4th ed., Ch. 12 · Hartle, Gravity, Chs. 4–5 · J. Bailey et al., "Measurements of relativistic time dilatation for positive and negative muons in a circular orbit", Nature 268 (1977) 301. The key results include: the boost-invariance of $c^{2}\mathrm{d}t^{2} - \mathrm{d}x^{2}$ (the $\mathrm{d}t\,\mathrm{d}x$ coefficient vanishes identically); the inequality $\sqrt{1-\beta^{2}} \leq 1$ with equality iff $\beta = 0$ ; the hyperbolic-motion integral giving $(c/a)\operatorname{arcsinh}(at/c)$ and $\mathrm{d}\theta/\mathrm{d}\tau = a/c$ ; and numerically the interval check $\sqrt{25-9} = 4$ , the CERN result $64.4\ \mu\mathrm{s}$ under $1.1\times10^{18}g$ , $c/g = 0.97\ \mathrm{yr}$ , and the GPS SR term $7.2\ \mu\mathrm{s/day}$ . The phrase "wristwatch time" is documented in the second edition of Spacetime Physics (1992); the 1966 first edition is credited with making proper time the operational spine of the subject, and the coinage is not claimed for it. Minkowski's Eigenzeit (1908) is the name's origin, as the history card notes.

Builds on. The Invariant Spacetime Interval · The Lorentz Factor Gamma · Timelike, Spacelike, and Lightlike Separations · The Light Clock and Time Dilation. Leads to. Time Dilation · Proper Length · The Twin Paradox · Proper Time and Proper Length · Four-Vectors.

Is proper time the shortest time or the longest? I have seen both claims.

Both — they are different statements, and keeping them apart is essential. (i) For two events on one clock's worldline, that clock's proper time is the smallest elapsed time any inertial frame assigns to the pair, since every other frame says $\Delta t = \gamma\,\Delta\tau > \Delta\tau$ . That is time dilation. (ii) Among all worldlines joining two fixed events, the inertial one has the largest proper time — this article's theorem. The first compares one path across many frames; the second compares many paths within one frame. Confusing them is the commonest way to get the twin paradox backwards.

Does the accelerating twin age less because acceleration slows her clock?

No. The clock hypothesis — tested to about $10^{18}g$ in CERN's muon ring — says a clock's rate depends only on its instantaneous speed, with no term for $\dot\beta$ . Acceleration matters for exactly one reason: it is the only way to bend a worldline, and hence the only way to leave and return. The ageing is bought by the bend, and $\int\sqrt{1-\beta^{2}}\,\mathrm{d}t$ never asks how violently you turned. A gentle year-long turn and a brutal millisecond one give the same answer if the speeds match.

Why is the straight worldline the LONGEST? Every geometric instinct says the opposite.

Because the metric has a minus sign, and your instincts were trained on a plus. In Euclidean geometry $\mathrm{d}s^{2} = \mathrm{d}x^{2} + \mathrm{d}y^{2}$ , so wandering in $y$ adds to the length. In spacetime $c^{2}\mathrm{d}\tau^{2} = c^{2}\mathrm{d}t^{2} - \mathrm{d}x^{2}$ , so wandering in $x$ subtracts. The factor measuring your path is $\sqrt{1 - \beta^{2}}$ , which can never exceed one. The triangle inequality does not merely fail — it reverses, and the reversal is exactly one minus sign deep.

Is proper time the same thing as the spacetime interval?

Very nearly; the difference is a convention. For a timelike separation the invariant $s^{2} = c^{2}\Delta t^{2} - \Delta x^{2}$ is positive and $\Delta\tau = \sqrt{s^{2}}/c$ — so proper time is the interval, in seconds rather than metres, read as the time on a clock that travels between the events. For a spacelike separation $s^{2} < 0$ , no clock can make the trip, and the natural reading is a proper length. Same invariant, two physical readings, chosen by the sign. The Invariant Spacetime Interval owns the general object; here we use its timelike face.

What is the proper time "in frame S"? I keep being told the question is wrong.

It is wrong, and gently so: the question contains a category error. Proper time is not a property of a frame but of a worldline — as the length of a road is a property of the road, not of the map you drew it on. Ask "what proper time elapses along this clock's path between these two events" and every frame in the universe returns the same number. Ask "what is the proper time in $S$ " and there is nothing to compute until you say whose clock. When a textbook says "the proper time in the rocket frame", it means "the proper time along the rocket's worldline".

If a photon accumulates zero proper time, what does the universe look like to a photon?

The question has no answer, and that is the point rather than a dodge. Proper time is defined by $c^{2}\mathrm{d}\tau^{2} = c^{2}\mathrm{d}t^{2} - \mathrm{d}\mathbf{x}^{2}$ , and for a light ray this vanishes identically: $\mathrm{d}\tau = 0$ at every step. There is no frame in which a photon is at rest — a boost can bring you arbitrarily close to $c$ but never to it, and the limit is singular, not a destination. So there is no clock to carry and nothing for it to read. What survives is a statement about the path: it is a null curve, of zero interval, joining events that light can connect.

Proper Time

Where this lives today

The physical picture

The core equations

Where the results come from

Limiting cases

Worked examples

1 · Four years younger

2 · The muon that survives $10^{18}g$

Physics of the portrait

Going further

Before this

Where this leads

Proper Time

Where this lives today

The physical picture

The core equations

Where the results come from

Limiting cases

Worked examples

1 · Four years younger

2 · The muon that survives 1018g10^{18}g1018g

Physics of the portrait

Going further

Before this

Where this leads

2 · The muon that survives $10^{18}g$