Time Dilation

Time is the quantity Newton was most sure of: a single river $t$ flowing "equably, without relation to anything external," shared by all observers. Special relativity drains that river. The moment you demand — as every experiment since Michelson and Morley has forced you to — that the speed of light $c$ be the same in every inertial frame, absolute time becomes impossible. Two observers in relative motion cannot agree on how much time passes between a given pair of events, because the light carrying information between them travels different path-lengths for each. Time dilation is the sharpest edge of that collapse: a clock moving at speed $v$ through your frame ticks slow by the Lorentz factor $\gamma = 1/\sqrt{1 - v^2/c^2}$ , and there is nothing you can do about it. The honest statement is geometric rather than kinematic. Every clock reads its own proper time $\tau$ , the Minkowski "length" of the worldline it traces through spacetime, and the invariant interval $\mathrm{d}s^2 = -c^2\,\mathrm{d}t^2 + \mathrm{d}x^2 + \mathrm{d}y^2 + \mathrm{d}z^2$ ties that length to the coordinates of any frame. Because the interval is the same for everyone while the coordinates are not, a moving clock's proper time falls behind coordinate time by exactly $\gamma$ . This is not a trick of signal delays — it is a fact about the shape of spacetime, and it is reciprocal: each of two inertial observers finds the other's clock slow, with no contradiction, once relativity of simultaneity is taken seriously. This article takes the invariant interval and the Lorentz transformation as given and asks what the slowing of a moving clock actually means, how it is derived three independent ways, and why it is among the best-tested predictions in physics.

Key takeaways

A clock moving at speed $v$ ticks slow by the Lorentz factor $\gamma = 1/\sqrt{1 - v^2/c^2} \ge 1$ : a coordinate interval $\Delta t$ corresponds to proper time $\Delta\tau = \Delta t/\gamma$ on the moving clock.
It follows from the two postulates plus the invariance of the spacetime interval; the moving clock reads proper time, the Minkowski arc length of its worldline.
The effect is reciprocal — each inertial observer sees the other's clock slow — yet consistent, because the two comparisons use different sets of simultaneous events.
Among timelike worldlines joining two events, the straight inertial one accumulates the most proper time; a spatial detour always costs elapsed time. This resolves the twin "paradox."
Confirmed across roughly nineteen orders of magnitude in $\gamma - 1$ : atmospheric muons, Ives–Stilwell Doppler, Hafele–Keating flying clocks, storage-ring muons, and GPS.
Key idea: a moving clock measures proper time $\Delta\tau = \Delta t/\gamma$ , so with $\gamma = 1/\sqrt{1 - v^2/c^2} \ge 1$ the moving clock always lags — a geometric statement about the length of its worldline, not an illusion of signalling.

The core insight

A moving clock measures proper time $\Delta\tau$ , related to the coordinate time $\Delta t$ of the frame in which it moves at speed $v$ by $\Delta t = \gamma\,\Delta\tau$ with $\gamma = 1/\sqrt{1 - v^2/c^2} \ge 1$ — so the moving clock always lags.

Where this lives today

Time dilation long ago stopped being a curiosity and became infrastructure. Every global navigation satellite carries a clock pre-tuned to the "wrong" rate so that, once the $+38\ \mu\mathrm{s/day}$ net drift is folded in, it keeps true ground time — without it, positioning smears by kilometres within a day. At the measurement frontier, optical lattice clocks now reach fractional uncertainties near $10^{-18}$ , precise enough that Chou and colleagues at NIST measured the kinematic dilation of a single aluminium ion moving at a few metres per second, and the gravitational shift over a $33\ \mathrm{cm}$ height difference — relativity tested at walking pace. In particle physics it is a working tool: muons in storage rings and heavy-ion products are routinely dilated by factors in the thousands, and $g\!-\!2$ experiments exploit that dilation while confirming the clock hypothesis under enormous acceleration. Open questions sit where kinematic and gravitational dilation blur together — relativistic geodesy mapping the geoid from clock-rate differences, and searches for varying fundamental constants across a gravitational potential.

The physical picture

Strip a clock down to its physics and you get a "light clock": two parallel mirrors a distance $L$ apart with a single photon bouncing between them, one tick per round trip. At rest, that tick lasts $\Delta\tau = 2L/c$ . Now set the clock gliding sideways at $v$ . In your frame the photon can no longer go straight up and down — between leaving the bottom mirror and returning, the apparatus has slid forward, so the photon runs a longer diagonal just to stay between the mirrors. Light will not speed up to compensate, so a longer path means a longer tick. The moving clock runs slow, and the right triangle formed by the diagonal, the mirror separation, and the displacement delivers the factor $\gamma$ exactly — with no jargon beyond "the speed of light is the same for everyone."

Figure

A light clock at rest ticks every Δτ = 2L/c; set it moving at v = 0.8c and the photon must cover the longer diagonal cΔt/2, so one tick stretches to Δt = γΔτ = 1.667 Δτ — the moving clock runs slow

There is a deeper way to see it that pays off across relativity: proper time is arc length through spacetime, but with a minus sign in front of the time term. In Euclidean geometry a straight line is the shortest path between two points; in Minkowski geometry the straight (inertial) worldline between two timelike-separated events is the longest in proper time, and every wiggle into space trades it away. That single sign flip is the whole content of the twin "paradox": the twin who stays home traces the straight worldline and ages most; the twin who rockets away and turns back takes a bent, shorter path and returns younger. Acceleration does not slow the traveller's clock; it only marks which worldline is the bent one. The reciprocity that troubles newcomers dissolves once you ask how a clock rate is measured. "How fast is your single clock ticking?" is not answerable by one observer alone; you compare your one moving clock against a whole row of my clocks as you pass them, synchronized in my frame. Because we disagree about which distant clocks read the same time (relativity of simultaneity), we can each honestly find the other's clock slow. The second figure shows how violent the slowing becomes: for everyday speeds it is invisible, $\gamma - 1 \approx v^2/2c^2$ giving a jet at $250\ \mathrm{m/s}$ only $3\times10^{-13}$ , yet the curve turns into a cliff as $v \to c$ .

Figure

The Lorentz factor γ = 1/√(1 − v²/c²) versus speed: essentially 1 at everyday speeds, γ = 1.005 at 0.1c, γ = 1.667 at 0.8c, a runaway climb to γ = 7.09 at 0.99c, and γ → ∞ as v → c where proper time freezes

The core equations

Everything begins with the one quantity all inertial observers agree on. The spacetime interval between two neighbouring events is frame-independent, and for a timelike worldline it defines the proper time the clock reads. This equation contains time dilation whole; the rest is bookkeeping.

-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: $\mathrm{d}\tau$ — proper-time interval read by a clock carried between the events, real and positive for a timelike worldline; $\mathrm{d}t,\mathrm{d}x,\mathrm{d}y,\mathrm{d}z$ — coordinate displacements in some inertial frame; $c$ — vacuum speed of light, $299\,792\,458\ \mathrm{m/s}$ . The left side is the same in every frame because $\tau$ belongs to the worldline, not to any coordinate system; the right side is the Minkowski metric. The amount a moving clock lags is packaged into one dimensionless number, the Lorentz factor — $1$ at rest, imperceptible at ordinary speeds, divergent as $v \to c$ .

\gamma = \frac{1}{\sqrt{1 - \dfrac{v^2}{c^2}}} \qquad (2)

where: $\gamma$ — Lorentz (time-dilation) factor, $\gamma \ge 1$ ; $v$ — speed of the clock relative to the chosen inertial frame. Physically, $\gamma$ is the number of coordinate seconds per proper second of the moving clock — the factor by which its rate is reduced in your frame. Dividing coordinate time by $\gamma$ gives the headline result: a duration $\Delta t$ on your synchronized lab clocks corresponds to a shorter $\Delta\tau$ ticked off by a single clock that moved between the events.

\Delta t = \gamma\,\Delta\tau \qquad\Longleftrightarrow\qquad \Delta\tau = \frac{\Delta t}{\gamma} \qquad (3)

where: $\Delta t$ — coordinate time in the frame where the clock moves; $\Delta\tau$ — proper time read on the moving clock. Because $\gamma \ge 1$ , proper time is always the smaller — the moving clock records less, never more. When the speed varies you cannot use one $\gamma$ ; you integrate the instantaneous proper-time increments — the general definition of elapsed time on any worldline, inertial or accelerated.

\tau = \int \frac{\mathrm{d}t}{\gamma(t)} = \int \sqrt{1 - \frac{v(t)^2}{c^2}}\;\mathrm{d}t \qquad (4)

where: $v(t)$ — instantaneous speed; $\gamma(t)$ — instantaneous Lorentz factor. The integrand is at most $1$ and drops below it whenever the clock moves, so any round trip accumulates less proper time than a clock that stayed put — the geometric root of the twin result. A change of variables turns $\gamma$ into a hyperbolic cosine. The rapidity $\phi$ is the boost "angle," and unlike velocities it adds under collinear boosts.

\gamma = \cosh\phi, \qquad \frac{v}{c} = \tanh\phi, \qquad \frac{v}{c}\,\gamma = \sinh\phi \qquad (5)

where: $\phi$ — rapidity. As $v \to c$ , $\phi \to \infty$ : no finite stack of boosts reaches light speed, which is why $\gamma$ has no upper bound. Recalled from the rapidity article; it makes the $\gamma \to \infty$ limit transparent. Finally, differentiating position by proper time rather than coordinate time defines the four-velocity, which carries $\gamma$ into relativistic dynamics; its fixed Minkowski norm is the covariant restatement of time dilation.

u^\mu = \frac{\mathrm{d}x^\mu}{\mathrm{d}\tau} = \gamma\,(c,\,\vec{v}), \qquad u^\mu u_\mu = -c^2 \qquad (6)

where: $u^\mu$ — four-velocity; $x^\mu = (ct,\vec{x})$ — spacetime position. The $\gamma$ appears because we differentiate by the clock's own time $\tau$ ; every four-velocity has "length" $c$ , so moving through space means slowing through time.

Derivation ATime dilation from the invariant interval

The goal here is to obtain $\Delta t = \gamma\,\Delta\tau$ directly from the frame-independence of the spacetime interval, without writing a single Lorentz transformation, so that time dilation appears as a pure consequence of the metric rather than of coordinate machinery. The strategy is to evaluate the same invariant interval $\mathrm{d}s^2$ twice — once in the clock's rest frame, where only time flows, and once in the lab, where the clock also moves through space — then equate them because the interval is the same for everyone.

Step 1. In the clock's rest frame it does not move, so $\mathrm{d}x = \mathrm{d}y = \mathrm{d}z = 0$ and the interval between two ticks is purely temporal: $\mathrm{d}s^2 = -c^2\,\mathrm{d}\tau^2$ . Why: proper time is defined as the interval read by a clock at rest in its own frame, so this is the definition of $\tau$ .

Step 2. In the lab the same clock moves at $v$ , advancing $\mathrm{d}x = v\,\mathrm{d}t$ in coordinate time $\mathrm{d}t$ , so $\mathrm{d}s^2 = -c^2\,\mathrm{d}t^2 + \mathrm{d}x^2 = -c^2\,\mathrm{d}t^2\left(1 - v^2/c^2\right)$ . Why: substituting the trajectory $\mathrm{d}x = v\,\mathrm{d}t$ encodes "this clock is moving" into the geometry.

Step 3. The interval is invariant, so the two evaluations are equal: $c^2\,\mathrm{d}\tau^2 = c^2\,\mathrm{d}t^2\left(1 - v^2/c^2\right)$ . Why: the same geometric quantity computed in two frames forces a relation between $\mathrm{d}\tau$ and $\mathrm{d}t$ — this is the whole physics.

Step 4. Take the positive square root: $\mathrm{d}t = \gamma\,\mathrm{d}\tau$ , and integrating a constant-velocity segment gives $\Delta t = \gamma\,\Delta\tau$ . Why: since $\gamma \ge 1$ the coordinate interval exceeds the proper interval — the moving clock reads less. *Verified: reduces to $\Delta t = \Delta\tau$ at $v = 0$ ; gives $\gamma(0.8c) = 5/3$ ; agrees with the two derivations below and Schutz §1.6.*

Go deeper · Derivation B · 1 minThe light-clock derivation from Pythagoras alone

Derivation BThe light-clock derivation from Pythagoras alone

The aim here is to recover the exact Lorentz factor from nothing but the Pythagorean theorem and the postulate that light moves at $c$ in every frame, so that a reader who has not met the metric can still see why the factor is $\gamma$ and not something else. We build a transverse light clock, compute one tick at rest, then watch the photon trace a diagonal in the lab and demand it still move at $c$ along that longer path.

Step 1. In the rest frame the photon goes straight up and back, a round-trip distance $2L$ at speed $c$ , so one tick is $\Delta\tau = 2L/c$ . Why: with the mirrors perpendicular to the motion, transverse distances are unaffected by the boost, so $L$ is a shared ruler in both frames.

Step 2. In the lab the clock drifts at $v$ , moving $v\,\Delta t$ in one tick $\Delta t$ , and the photon covers a lab distance $c\,\Delta t$ along two diagonals; each half-tick is the hypotenuse of a right triangle with vertical leg $L$ and horizontal leg $v\,\Delta t/2$ . Why: the photon must still reach the moving mirror, so its lab path genuinely exceeds $2L$ , and $c$ cannot increase to compensate.

Step 3. Pythagoras on one half-tick: $\left(c\,\Delta t/2\right)^2 = L^2 + \left(v\,\Delta t/2\right)^2$ , so $\dfrac{c^2\,\Delta t^2}{4}\left(1 - \dfrac{v^2}{c^2}\right) = L^2$ . Why: isolating $\Delta t$ is where the factor $1 - v^2/c^2$ is born — the shortfall between the photon's speed and its useful vertical progress.

Step 4. Solve: $\Delta t = \dfrac{2L/c}{\sqrt{1 - v^2/c^2}} = \gamma\,\dfrac{2L}{c} = \gamma\,\Delta\tau$ . Why: substituting $\Delta\tau = 2L/c$ matches the interval derivation — a moving light clock ticks slow by exactly $\gamma$ , from geometry alone. *Verified: at $v = 0$ gives $\Delta t = 2L/c = \Delta\tau$ ; matches equation (3) and Taylor & Wheeler, Spacetime Physics, Ch. 1; the transverse- $L$ assumption is confirmed by the length-contraction result.*

Go deeper · Derivation C · 1 minTime dilation and reciprocity from the Lorentz transformation

Derivation CTime dilation and reciprocity from the Lorentz transformation

This derivation reproduces $\Delta t = \gamma\,\Delta\tau$ from the explicit Lorentz boost and, more importantly, makes reciprocity manifest and shows why it is consistent rather than contradictory. The plan is to place a clock at rest at the origin of a moving frame $S'$ , transform two of its ticks into the lab $S$ , then swap the frames — exposing that each observer's "the other clock is slow" rests on events simultaneous in different frames.

Step 1. Let $S'$ move at $v$ along the $x$ -axis of $S$ , inverse boost $t = \gamma\left(t' + v x'/c^2\right)$ . Put a clock at rest at $x' = 0$ ; its ticks are $(t'_1, 0)$ and $(t'_2, 0)$ , and since it does not move in $S'$ its proper time is $\Delta\tau = \Delta t' = t'_2 - t'_1$ . Why: a clock at rest at $x'=0$ is exactly the situation equation (3) describes, ready for the boost to convert.

Step 2. Transform each tick to $S$ . Because $x' = 0$ , the spatial term vanishes and $t_i = \gamma\,t'_i$ . Why: the vanishing $x'$ term is why the answer is clean — the special feature of "one clock, same place in its own frame."

Step 3. Subtract: $\Delta t = \gamma\,(t'_2 - t'_1) = \gamma\,\Delta\tau$ , so the lab separates the two ticks by the longer interval $\gamma\,\Delta\tau$ and judges the clock slow. Why: time dilation again, now from the transformation instead of the metric — a nontrivial consistency check.

Step 4. Reverse roles: a clock at rest at $x = 0$ in $S$ , boosted into $S'$ , gives $\Delta t' = \gamma\,\Delta t$ — so $S'$ finds the $S$ clock slow too. No contradiction: $S$ compares against events at one point in $S$ but two in $S'$ ; not the same comparison. Why: the asymmetry lives in which events count as "simultaneous," so reciprocity is about simultaneity conventions, not a logical fault. *Verified: consistent with the inverse boost and with Carroll §1.2 and Schutz §1.8; the reciprocal $\Delta t' = \gamma\,\Delta t$ follows by identical algebra with $x=0$ .*

Limiting cases

Case	Result	Meaning
Non-relativistic ( $v \ll c$ )	$\gamma \approx 1 + \tfrac{1}{2}\dfrac{v^2}{c^2}$	Quadratic, minute correction; recovers Newton's absolute time as $v/c \to 0$ . A jet at $250\ \mathrm{m/s}$ has $\gamma - 1 \approx 3\times10^{-13}$ .
Mildly relativistic ( $v = 0.1c$ )	$\gamma \approx 1.005$	Half-a-percent slowing; already resolvable with modern optical-clock comparisons.
Highly relativistic ( $v = 0.8c$ )	$\gamma = 5/3 \approx 1.667$	Clock runs two-thirds speed; the twin-paradox regime.
Ultra-relativistic ( $v = 0.99c$ )	$\gamma \approx 7.09$	Clock runs about seven times slow; cosmic-ray muons and storage-ring beams.
Photon limit ( $v \to c$ )	$\gamma \to \infty,\ \mathrm{d}\tau \to 0$	Proper time per coordinate time vanishes; a light ray has a null worldline and no rest frame.

Common trapReciprocal time dilation already settles the twin paradox, so acceleration must be doing something extra

The statements "B's clock runs slow in A's frame" and "A's clock runs slow in B's frame" are both true and not contradictory — but they are not enough to say who is younger when the twins reunite, and it is a mistake to think the story is symmetric until acceleration mysteriously "breaks" it. The clean way to see it: each twin's ageing is the proper time along their worldline, $\tau = \int \mathrm{d}t/\gamma$ , an invariant every frame agrees on. The stay-home twin traces one straight worldline; the traveller a bent one, and in Minkowski geometry the straight worldline is longest in proper time — so the traveller unambiguously ages less, computable in any single inertial frame without invoking the acceleration. Acceleration merely marks which twin took the bent path; it does not slow any clock (the clock hypothesis, confirmed near $10^{18}\,g$ ). Whenever you compare rates, name which events and whose simultaneity first.

Worked examples

1 · The travelling twin (fully numeric)

One twin flies to a star $4$ light-years away and back at $v = 0.8c$ ; the other stays on Earth. How much does each age, and does the length-contracted view of the traveller agree?

Step 1. The factor: $\gamma = 1/\sqrt{1 - 0.8^2} = 1/\sqrt{0.36} = 1/0.6 = \mathbf{5/3 \approx 1.667}$ , the ratio of Earth time to traveller proper time per leg.

Step 2. Earth-frame time: each leg is $4$ ly at $0.8c$ , taking $5$ years, so the round trip is $\Delta t = \mathbf{10\ \text{years}}$ . The stay-home twin ages $10$ years.

Step 3. Traveller proper time: $\Delta\tau = \Delta t/\gamma = 10/1.667 = \mathbf{6.0\ \text{years}}$ . The traveller returns 4 years younger.

Step 4. Cross-check from the traveller's frame: the distance is contracted to $4/\gamma = 2.4$ ly each way, covered at $0.8c$ in $3.0$ years per leg, again $6.0$ years total. Both frames agree: 10 years on Earth, 6 for the traveller — time dilation and length contraction give one invariant answer.

2 · Atmospheric muons reaching the ground (fully numeric)

Cosmic-ray muons are produced $15\ \mathrm{km}$ up with rest-frame lifetime $\tau_0 = 2.2\ \mu\mathrm{s}$ and speed $v = 0.995c$ ( $\gamma = 10.0$ ). What fraction survive to sea level, versus a world without time dilation?

Step 1. Classical (no dilation): in one lifetime a muon travels only $v\,\tau_0 = 0.995 \times 3\times10^8 \times 2.2\times10^{-6} \approx \mathbf{657\ \mathrm{m}}$ , far short of $15\ \mathrm{km}$ . The lab flight time is $t = 15000/(0.995 \times 3\times10^8) = 50.3\ \mu\mathrm{s}$ , so a naive survival fraction is $e^{-t/\tau_0} = e^{-22.8} \approx \mathbf{1.2\times10^{-10}}$ — essentially none.

Step 2. Relativistically, decay runs on the muon's own clock, $N = N_0 e^{-\Delta\tau/\tau_0}$ , which advances only $\Delta\tau = t/\gamma = 50.3/10.0 = \mathbf{5.02\ \mu\mathrm{s}}$ during the descent.

Step 3. Survival: $N/N_0 = e^{-5.02/2.2} = e^{-2.28} \approx \mathbf{0.10}$ — about one in ten reaches the ground, an enhancement of $\sim 10^{9}$ over the classical prediction. Equivalently, in the muon's frame the atmosphere is contracted to $15/10 = 1.5\ \mathrm{km}$ , giving the identical answer. This is the Rossi–Hall (1941) and Frisch–Smith (1963) result.

3 · GPS: the kinematic clock correction (fully numeric)

A GPS satellite orbits at $v \approx 3.87\ \mathrm{km/s}$ . How fast do its clocks run slow from special-relativistic dilation, how does gravity combine with it, and what positioning error follows if both are ignored?

Step 1. SR rate: for $v \ll c$ , $\gamma - 1 \approx \tfrac{1}{2}v^2/c^2 = \tfrac{1}{2}(3870/2.998\times10^8)^2 = 8.3\times10^{-11}$ . Over one day ( $86\,400\ \mathrm{s}$ ) the orbiting clock loses $8.3\times10^{-11}\times 86\,400 \approx \mathbf{-7.2\ \mu\mathrm{s/day}}$ .

Step 2. Gravitational contribution: clocks higher in Earth's potential run faster, adding about $\mathbf{+45.6\ \mu\mathrm{s/day}}$ at GPS altitude (from the gravitational-time-dilation article). Opposite sign to the kinematic term.

Step 3. Net drift and error: $-7.2 + 45.6 \approx \mathbf{+38\ \mu\mathrm{s/day}}$ . Since positions come from light travel time, this maps to $c\,\delta t \approx \mathbf{11\ \mathrm{km/day}}$ . GPS works only because both corrections are built into the satellite clocks.

Physics of the portrait

Time dilation made "now" a local rather than a universal word — and it came not from new data but from one man taking the constancy of light more seriously than anyone before.

Origins · 2 minAlbert Einstein and how this was discovered

Albert Einstein

1879–1955 · special relativity and the physics of time

By 1905 the machinery was almost all on the table. Lorentz had written down the transformation that bears his name and a "local time" $t'$ he regarded as a mathematical auxiliary; FitzGerald and Lorentz had proposed length contraction to save the ether from the Michelson–Morley null result; Poincaré had questioned whether distant simultaneity meant anything absolute. What no one had done was throw away the ether and read the equations as physics. In his June 1905 paper "On the Electrodynamics of Moving Bodies," Albert Einstein — a $26$ -year-old patent examiner in Bern with no university post — took just two postulates, the principle of relativity and the invariance of $c$ , and derived that moving clocks run slow and moving rods contract, not as dynamical effects of motion through a medium but as properties of time and space themselves. The same short paper contains the relativity of simultaneity that resolves the reciprocity puzzle, and the remark that a clock carried around a closed path returns having lost time — the twin effect, stated before it was called a paradox. Three years later Hermann Minkowski recast the theory as the geometry of four-dimensional spacetime, giving time dilation its modern reading as the Minkowski length of a worldline: "Henceforth space by itself, and time by itself, are doomed to fade away into mere shadows." Einstein, who first dismissed that geometry as "superfluous learnedness," would build general relativity on exactly that spacetime — and would live to see his slowing clocks confirmed in muons, in flown atomic clocks, and in the satellites overhead.

History

1887

— Michelson and Morley find no ether drift, setting the stage for a new kinematics

1892

— Lorentz introduces a contraction and a mathematical local time to explain the null result

1905

— Einstein derives time dilation and the relativity of simultaneity from two postulates

1908

— Minkowski recasts relativity as the geometry of four-dimensional spacetime

1938

— Ives and Stilwell measure the transverse Doppler shift, the first lab test of time dilation

1941

— Rossi and Hall confirm the dilated lifetime of cosmic-ray muons

1971

— Hafele and Keating fly atomic clocks around the world and read off the predicted offset

Going further

References. Misner, Thorne & Wheeler, Gravitation (1973), §1.5, §6 · Wald, General Relativity (1984), §4.2 · Carroll, Spacetime and Geometry (2004), §1.2 · Hartle, Gravity (2003), Ch. 4 · Schutz, A First Course in General Relativity (2009), §1.6, §1.8 · Taylor & Wheeler, Spacetime Physics (1992), Ch. 1–3.

Provenance. The Lorentz factor and $\Delta t = \gamma\,\Delta\tau$ are standard (Carroll §1.2; Schutz §1.8; MTW §1.5); the invariant interval and rapidity forms are recalled by name from the spacetime-interval and Lorentz-boost articles. A verifier should recompute: (1) the interval and light-clock derivations of $\gamma$ ; (2) the Lorentz-boost derivation at $x' = 0$ with its reciprocity check; (3) $\gamma(0.8c) = 5/3$ and the twin ages $10$ yr vs $6$ yr; (4) muon survival $e^{-5.02/2.2} \approx 0.10$ with $\tau_0 = 2.2\ \mu\mathrm{s}$ (PDG); (5) the GPS split $-7$ (SR) $+46$ (GR) $\approx +38\ \mu\mathrm{s/day}$ net (Ashby, Living Rev. Relativity 2003). Speed of light $c = 299\,792\,458\ \mathrm{m/s}$ exact.

Prerequisites. Postulates of Special Relativity · Lorentz Transformations · Spacetime Interval · Relativity of Simultaneity. Leads to. Proper Time · Length Contraction · Lorentz Boosts and Rapidity · Causal Structure and Light Cones.

Ask the tutor: ask me to walk through the twin paradox on a spacetime diagram, or to work out the GPS clock offset from orbital and gravitational data.

Is time dilation a real physical effect or just an illusion of signal delay?

It is completely real and has nothing to do with light-travel time. A common first guess is that a moving clock only looks slow because its light takes longer to reach you, but time dilation survives after every such delay is subtracted out. The proper measurement — comparing the moving clock against a row of clocks synchronized in your frame as it passes them — removes signalling delays entirely and still finds it behind by $\gamma$ . The muons that reach the ground genuinely decay less, and the flown clocks in Hafele–Keating genuinely read a different number on return. It is a fact about elapsed proper time, not about when you see the ticks.

If motion is relative, which clock actually runs slow?

Each of two inertial observers correctly measures the other's clock to run slow, and both are right because they answer different questions. "Is your clock slow?" only has an answer once you fix a set of simultaneous events to compare against, and the two frames disagree about which distant events are simultaneous. So while both move inertially there is no frame-independent fact of the form "clock A is the slow one." An absolute answer appears only when the clocks meet twice — depart and reunite — because then you compare accumulated proper times directly, and the clock that took the bent worldline reads less. Relativity of motion is exactly why the single-clock comparison has no absolute winner.

Does acceleration itself slow a clock, as in the twin paradox?

No — this is the clock hypothesis, and it is very well tested. A clock's rate depends only on its instantaneous speed through $\gamma(v)$ , not on its acceleration; muons in the CERN storage rings feel accelerations near $10^{18}\,g$ yet their dilated lifetimes match the speed-based $\gamma\tau_0$ . In the twin scenario the acceleration matters only because it makes the traveller's worldline bent instead of straight, and it is the bent shape — the extra journey through space — that costs proper time. You can even get the age difference in a purely inertial frame where the acceleration never appears, just integrating $\int \mathrm{d}t/\gamma$ along each path. Acceleration selects the path; speed sets the rate.

Why does a moving clock reading less time correspond to the longest worldline?

Because Minkowski geometry carries a minus sign Euclidean geometry does not. Proper time is the "length" of a worldline in the metric $-c^2\,\mathrm{d}\tau^2 = -c^2\,\mathrm{d}t^2 + \mathrm{d}x^2$ , where the spatial terms subtract. Adding motion (nonzero $\mathrm{d}x$ ) therefore reduces $\mathrm{d}\tau$ — the opposite of Euclidean space, where wandering sideways adds length. So among all worldlines joining two events, the straight inertial one has the greatest proper time, and any detour accumulates less. "Moving clock runs slow" and "straight worldline is longest" are the same statement read two ways.

What sets the scale of the effect, and why don't we notice it day to day?

The scale is $v/c$ , and for anything human it is tiny. For small speeds $\gamma - 1 \approx \tfrac{1}{2}v^2/c^2$ , so the fractional slowing grows only as the square of speed: a car at highway speed has $\gamma - 1 \sim 10^{-16}$ , a jet at $250\ \mathrm{m/s}$ only $\sim 3\times10^{-13}$ . That is why nothing in ordinary life betrays it, and why absolute Newtonian time was a superb approximation for three centuries. It becomes conspicuous only when $v$ is a real fraction of $c$ (cosmic-ray muons, accelerator beams) or when clocks are stable enough ( $10^{-16}$ and beyond) to see nanoseconds, as with atomic clocks and GPS.

Is a photon's frame a valid place to ask about time dilation?

No — there is no inertial rest frame for light, so "the photon's point of view" is not a legitimate frame. As $v \to c$ the Lorentz factor diverges and the boost becomes singular; no allowed transformation brings a photon to rest. What is well defined is that a photon's worldline is null, $\mathrm{d}s^2 = 0$ , so the proper time along it is exactly zero. "A photon experiences no time" is a heuristic gloss on that null-interval fact, not a description of a photon frame. To reason about light, stay in a legitimate inertial frame and use the null condition; the photon-frame limit does not exist.

How is time dilation related to gravitational time dilation?

They are two faces of the same clocks-read-worldline-length principle, with different sources. This article's dilation is kinematic — it depends on relative velocity through $\gamma(v)$ in flat Minkowski spacetime. Gravitational time dilation depends instead on position in a potential: clocks deeper in the potential run slow, set by the time–time part of the curved-spacetime metric rather than by $\gamma$ . In real systems both act at once and add, as in GPS ( $-7\ \mu\mathrm{s/day}$ from speed, $+46$ from the higher potential). General relativity unifies them: both say a clock reads the proper time $\int\sqrt{-\mathrm{d}s^2}/c$ along its worldline, whatever the metric.