Length Contraction

Ask what it means to say a moving train "is" twelve metres long, and you are forced into a surprisingly deep operation: you must record where its front is and where its back is at the same instant, and subtract. The "at the same instant" is the whole story. Special relativity's second pillar — the relativity of simultaneity, established in this chapter's preceding article — says that two events which are simultaneous in one inertial frame are generally not simultaneous in another. So two observers in relative motion, each doing the honest thing and marking both ends of the object at their own single instant, are marking different pairs of spacetime events. They get different separations. That disagreement, and nothing more mysterious, is length contraction. Quantitatively the answer is clean. An object of proper length $L_0$ — its length in the frame where it sits at rest — is measured, in any frame through which it moves at speed $v$ , to be $L = L_0\sqrt{1 - v^2/c^2} = L_0/\gamma$ along the direction of motion. Since the Lorentz factor $\gamma \ge 1$ , the moving length is always the shorter one, and directions perpendicular to the motion are untouched. The effect is perfectly reciprocal — each observer finds the other's ruler contracted — which is not a paradox but a direct consequence of the two frames slicing spacetime differently. Length contraction is the spatial face of the same Minkowski geometry whose temporal face is time dilation; both are carried by the single factor $\gamma$ , and both are geometry, not a force squeezing the atoms of the moving body.

Key takeaways

Length contraction is the shortening of a moving object along its direction of motion: proper length $L_0$ is measured as $L = L_0/\gamma$ , with $\gamma = 1/\sqrt{1-v^2/c^2}$ .
The proper length $L_0$ — the length in the object's rest frame — is the largest length any inertial observer assigns; every moving observer measures less.
The effect follows from the Lorentz transformation and hinges on the relativity of simultaneity: a "length" is the gap between two simultaneously marked endpoints, and simultaneity is frame-dependent.
Only the dimension parallel to the boost contracts; transverse dimensions are exactly unchanged, which a symmetry argument forces.
Contraction is reciprocal and paradox-free — $A$ sees $B$ 's ruler short and $B$ sees $A$ 's ruler short, because they mark different pairs of events.
Key idea: a length is the spatial separation of two simultaneous endpoint events; because simultaneity is relative, the moving observer's "same-time" events differ from the rest observer's, and Minkowski geometry yields $L = L_0/\gamma$ along the motion.

The core insight

A length measurement is the spatial separation of two simultaneous events marking the endpoints. Because simultaneity is relative, the moving observer's "same-time" endpoint events differ from the rest observer's, and the geometry of Minkowski spacetime yields $L = L_0/\gamma$ along the direction of motion.

Where this lives today

Length contraction is a daily working assumption in relativistic engineering and a visible effect in the largest machines ever built. At the Relativistic Heavy Ion Collider and the LHC, gold and lead nuclei are accelerated to $\gamma \sim 100$ (RHIC) and beyond $\gamma \sim 2500$ (LHC), so each spherical nucleus arrives in the lab frame as an ultra-thin "pancake" flattened by that factor along the beam. This is no fiction: the geometry of the flattened overlap region, the timing of the collision, and the initial energy density of the quark–gluon plasma that forms are all computed from the contracted longitudinal size, and the models fail without it. The same physics reshapes the field of any fast charge — the Coulomb field of a relativistic electron is compressed into a transverse disk, the field-theoretic image of a contracted charge cloud, underlying synchrotron radiation and the Weizsäcker–Williams treatment of a fast charge's field as a beam of photons. Even the muon detector at a mountain's base is a standing verification: the flux that arrives is the flux only a contracted atmosphere (equivalently, a dilated clock) allows. And as geometry, the tilted-slice picture is the gateway to general relativity, where "which spatial slice you cut" becomes a genuine freedom of curved spacetime rather than a mere choice of inertial frame.

The physical picture

Picture the object not as a thing sitting in space but as a worldtube threading through spacetime — a four-dimensional slab traced out as its two ends persist through time. Each endpoint sweeps a worldline; the pair of worldlines bounds a band. To "measure the length" is to cut across that band with a surface of simultaneity — a set of events you declare to be "now" — and read off the width of the slice. The catch is that different observers use differently tilted "now" surfaces, because moving clocks are synchronized differently. A tilt changes where the cut meets the two worldlines, and so it changes the width. Length contraction is that geometrical fact: the same worldtube, sliced two ways, has two widths.

Figure

A rod moving at β = 0.6 in the lab traces a worldtube of two parallel worldlines. The lab cuts it at one lab-time (horizontal) and measures the contracted L = 80 m; the rod's own simultaneity, the tilted x′ line, measures the proper L₀ = 100 m — longer by γ = 1.25.

The index changes with the tilt, and the tilt is set by speed, so the natural summary is a curve: how much does a metre stick shrink as its speed climbs toward $c$ ? The answer is deceptively lazy at first — at everyday speeds the factor $\sqrt{1-\beta^2}$ is indistinguishable from one — and then vicious near the light barrier, plunging toward zero. The second figure is that curve, the quantitative payoff of the whole article, with the three speeds we return to in the worked examples marked on it.

Figure

The contraction factor L/L₀ = √(1 − β²) versus speed β = v/c: almost flat until β ≈ 0.5, then plunging. At β = 0.6 it is 0.80 (γ = 1.25), at β = 0.866 it is 0.50 (γ = 2), and at β = 0.995 it is 0.10 (γ = 10).

Two features are worth naming before the algebra. The tilt of the "now" line and the drop of the curve are one fact seen twice — the more the simultaneity surface tilts (higher $\beta$ ), the shorter the slice. And contraction is inseparable from time dilation: boosting into a moving frame mixes the space and time axes, so the projection onto the space axis (contraction) and onto the time axis (dilation) are governed by the identical $\gamma$ . There is one relativistic effect here, wearing two costumes.

The core equations

The operational result is a single formula: a length parallel to the motion is divided by $\gamma$ . Everything else in this section either defines the ingredients or shows where the formula comes from.

L = \frac{L_0}{\gamma} = L_0\sqrt{1 - \frac{v^2}{c^2}} \qquad (1)

where: $L$ — length measured in the frame where the object moves at speed $v$ ; $L_0$ — proper length (length in the object's rest frame); $v$ — relative speed along the direction of the measured length; $c$ — speed of light in vacuum; $\gamma$ — Lorentz factor. Because $\gamma \ge 1$ , the measured $L$ never exceeds $L_0$ , and the two coincide only at $v = 0$ . The Lorentz factor is the universal relativistic amplifier: one at rest, growing without bound as $v \to c$ . It is the same $\gamma$ that dilates time, boosts momentum, and blueshifts a Doppler signal.

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

where: $\gamma$ — Lorentz factor; equals $1$ at rest and diverges as $v \to c$ ; $v/c$ — speed as a fraction of $c$ . The square root and the " $1-$ " inside make the contraction second-order, hence invisible at ordinary speeds. It is cleaner to work with the dimensionless velocity $\beta$ , which is bounded by the light barrier $\beta < 1$ .

\beta = \frac{v}{c}, \qquad \gamma = \frac{1}{\sqrt{1-\beta^2}} \qquad (3)

where: $\beta$ — dimensionless velocity, $0 \le \beta < 1$ ; the contraction factor is simply $\sqrt{1-\beta^2}$ , the curve plotted above. The formula descends from the Lorentz boost relating the coordinates of one inertial frame $S$ to another $S'$ gliding at $v$ along $x$ — the machinery that carries all of the kinematics.

\begin{aligned} t' &= \gamma\left(t - \frac{v x}{c^2}\right)\\ x' &= \gamma\left(x - v t\right) \end{aligned} \qquad (4)

where: $(t,x)$ — coordinates of an event in $S$ ; $(t',x')$ — coordinates of the same event in $S'$ . The $-vx/c^2$ term in $t'$ is the relativity of simultaneity in disguise — it is why "same $t$ " is not "same $t'$ ", and hence why contraction happens at all. Applied to two endpoint events marked at one instant ( $\Delta t = 0$ ) in the measuring frame, the boost gives the pivotal relation: the rest-frame separation is $\gamma$ times the measured one — the equation the derivation below turns into $L = L_0/\gamma$ .

\Delta x' = \gamma\,\Delta x \quad (\Delta t = 0 \text{ in } S) \qquad (5)

where: $\Delta x$ — endpoint separation measured simultaneously in the frame $S$ through which the object moves; $\Delta x'$ — separation in the rest frame $S'$ , which equals $L_0$ . Reading it as $\Delta x = \Delta x'/\gamma$ gives the contraction directly. Geometrically the proper length is an invariant: it is the length of the spacelike interval joining the two endpoint worldlines along the rest-frame slice, and every frame computes the same number from its own coordinates.

L_0^2 = (\Delta x)^2 - (c\,\Delta t)^2 \qquad (6)

where: $\Delta x, \Delta t$ — coordinate separations of the two endpoint events in any frame; $L_0$ — invariant proper length. In the rest frame $\Delta t = 0$ and $L_0 = \Delta x$ ; in a moving frame both terms are nonzero but combine to the same $L_0$ . Finally, packaging the boost in terms of rapidity $\phi$ makes the geometry explicit: a boost is a hyperbolic rotation, and the contraction is just $1/\cosh\phi$ .

L = L_0\sqrt{1-\tanh^2\!\phi} = \frac{L_0}{\cosh\phi} \qquad (7)

where: $\phi$ — rapidity, defined by $\tanh\phi = \beta$ ; $\cosh\phi = \gamma$ . Rapidities add linearly under successive boosts, which is why $\phi$ , not $v$ , is the natural angle-like variable of Minkowski geometry.

Derivation ALength contraction from the Lorentz transformation

The goal is to earn the formula $L = L_0/\gamma$ honestly from the boost equations, and in doing so to expose exactly where the physics enters — not in the algebra, which is a two-line subtraction, but in the measurement condition that both ends be recorded at one instant of the measuring frame. We place the object at rest in $S'$ , write down its endpoint worldlines, transform them, and impose simultaneity in $S$ . The one subtle move is that "measure the length in $S$ " means " $\Delta t = 0$ in $S$ ", and getting that condition right is the entire difference between the correct contraction and the classic wrong answer.

Step 1. Let the object lie at rest in frame $S'$ with its back at $x' = 0$ and its front at $x' = L_0$ ; these values hold for all $t'$ because the object does not move in its own frame. In frame $S$ the object glides at speed $v$ along $x$ . Why: fixing the endpoints in the rest frame is what makes $L_0$ the proper length — a length read off where nothing is moving, so no simultaneity subtlety can contaminate it.

Step 2. Use the boost $x' = \gamma(x - vt)$ for each endpoint, evaluating both at one common lab time $t$ : the back obeys $0 = \gamma(x_\text{back} - vt)$ and the front obeys $L_0 = \gamma(x_\text{front} - vt)$ . Why: recording both ends at the same instant $t$ in $S$ is the physical definition of measuring a moving object's length — this shared $t$ is the measurement condition, and everything hinges on imposing it in the frame that does the measuring.

Step 3. Subtract the two equations. The $-\gamma vt$ terms are identical and cancel, leaving $L_0 = \gamma(x_\text{front} - x_\text{back}) = \gamma L$ , where $L \equiv x_\text{front} - x_\text{back}$ is precisely the length measured in $S$ . Why: the cancellation of the $vt$ terms is only legitimate because the two ends share the same $t$ ; it is the algebraic footprint of the simultaneity condition.

Step 4. Solve for the measured length, $L = L_0/\gamma = L_0\sqrt{1 - v^2/c^2}$ . Since $\gamma \ge 1$ , the moving object is shorter than its proper length, with equality only at $v = 0$ . Why: this is the destination — a purely kinematic shrinkage with no force acting on the body, set entirely by the relative speed through $\gamma$ . The boxed result is $L = L_0/\gamma = L_0\sqrt{1 - v^2/c^2}$ , reproducing the standard treatment in Schutz Ch. 1, Carroll §1.2, and Hartle Ch. 4.

Go deeper · Derivation B · 2 minReciprocity without paradox

Derivation BReciprocity without paradox

The goal is to resolve the objection that most rattles a first-time student: if $A$ says $B$ 's ruler is short and $B$ says $A$ 's ruler is short, surely one is wrong. The claim to establish is that both are right simultaneously, and that the apparent contradiction dissolves once one notices the two observers are not measuring the same pair of spacetime events. We take two identical rulers of proper length $L_0$ sliding past one another, let each observer measure the other, and trace why the two "contractions" never collide into a paradox.

Step 1. Observer $A$ (frame $S$ ) marks the two ends of $B$ 's ruler at one instant of $A$ 's clock, $\Delta t = 0$ in $S$ . By the previous derivation $A$ finds $L_{A\to B} = L_0/\gamma$ . Why: $A$ is entitled to use $A$ 's own simultaneity; there is no privileged frame that could overrule it.

Step 2. Observer $B$ (frame $S'$ , moving at $+v$ ) marks the two ends of $A$ 's ruler at one instant of $B$ 's clock, $\Delta t' = 0$ in $S'$ . By the identical algebra with the inverse boost, $B$ finds $L_{B\to A} = L_0/\gamma$ . Why: the principle of relativity guarantees the physics is the same in both frames, so the contraction factor must come out identical — neither ruler is objectively the short one.

Step 3. The resolution is that $\Delta t = 0$ in $S$ does not imply $\Delta t' = 0$ in $S'$ : $A$ 's pair of "simultaneous" marking events is not simultaneous for $B$ , and vice versa. The two observers literally measure different pairs of events on the rulers' worldtubes. Why: with simultaneity relative, "the length now" picks out a different spacetime cross-section for each observer, so the two shortenings refer to different measurements and cannot contradict.

Step 4. Hence both statements are true, each in its own frame, and the symmetry is mandatory rather than paradoxical — swapping the labels $A \leftrightarrow B$ maps one correct measurement onto the other. Why: a paradox would require a frame-independent fact that came out two ways; there is none, only two frame-dependent measurements that each frame reports consistently. The boxed result is $L_{A\to B} = L_{B\to A} = L_0/\gamma$ , matching the resolution given in MTW §2.1 and Wald Ch. 1.

Go deeper · Derivation C · 2 minWhy only the parallel direction contracts

Derivation CWhy only the parallel direction contracts

The goal is to prove that lengths perpendicular to the motion are exactly unchanged — not approximately but identically — using a symmetry argument that needs no algebra, only the principle of relativity and the reversibility of the setup. Students often assume a moving object shrinks in every dimension; showing the transverse direction is protected sharpens what contraction really is (a projection along the boost) and justifies the " $L_\perp = L_{0,\perp}$ " the formula leaves implicit.

Step 1. Take two identical rings, $A$ and $B$ , of the same rest radius, and set them moving toward each other along their common axis, each carrying a fine scriber at its rim pointed at the other's rim. Ask whether one ring passes inside the other. Why: threading the question through a physical mark (a scratch) makes the outcome frame-independent — either $A$ scratches inside $B$ or it does not, and every observer must agree on which scratch exists.

Step 2. Suppose transverse lengths contracted. In $A$ 's frame $B$ is moving, so $B$ 's ring would be the smaller one and $A$ would scratch $B$ on the outside. But in $B$ 's frame $A$ is moving, so $A$ 's ring is smaller and $B$ scratches $A$ on the outside — an opposite, incompatible physical fact. Why: the two frames must agree on a permanent scratch mark, so any rule that makes "the moving ring smaller" produces a genuine contradiction, not a mere disagreement of perspective.

Step 3. The only assignment consistent for both frames is that the transverse radii are equal in every frame: neither ring is inside the other, and the scribers just graze. Hence $L_\perp = L_{0,\perp}$ exactly. Why: symmetry under $A \leftrightarrow B$ forbids any transverse change, because a change would have to pick a preferred ring, which relativity forbids.

Step 4. The argument fails for the parallel direction precisely because there simultaneity enters: a "length along the motion" requires marking two ends at one time, and the two frames disagree about that time — so no single frame-independent scratch pins the parallel length, and contraction is allowed. Why: this contrast is the crux — transverse lengths are absolute because they need no simultaneity convention, while parallel lengths are relative because they do. The boxed conclusion is $L_\perp = L_{0,\perp}$ with contraction confined to the boost direction, the standard result underlying the "pancaking" of fields and nuclei used later.

Limiting cases

The contraction factor $L/L_0 = 1/\gamma = \sqrt{1-\beta^2}$ governs every regime; reading its behaviour at the extremes is the fastest sanity check on any relativistic length calculation.

Case	Result	Meaning
Low speed (Newtonian)	$\beta \ll 1 \Rightarrow \dfrac{L}{L_0} \approx 1 - \tfrac{1}{2}\beta^2$	Contraction is second-order in $v/c$ and utterly negligible in daily life — at highway speed $\beta \sim 10^{-7}$ , so $L/L_0$ differs from $1$ by $\sim 10^{-15}$ ; lengths are effectively absolute.
Ultra-relativistic	$\beta \to 1 \Rightarrow \dfrac{L}{L_0} = \sqrt{1-\beta^2} \to 0$	The length collapses toward zero as $v \to c$ , but never reaches it, since a massive object always has $v < c$ ; the drop is arbitrarily steep near the light barrier.
Transverse direction	$L_\perp = L_{0,\perp}$	Dimensions perpendicular to the boost are exactly unchanged — the symmetry argument protects them; only the parallel dimension shrinks.
Rest frame	$v = 0 \Rightarrow L = L_0$	The proper length is recovered; $L_0$ is the maximum length any inertial observer measures, achieved only by the rest frame.

Common trapLength contraction squeezes the atoms of the moving body, like a compression.

It does nothing of the kind, and the confusion hides two separate errors. First, no force acts: contraction is a statement about how two frames slice the same worldtube, not about stresses inside the object — the atoms are undisturbed in the rest frame, where nothing is happening to them at all. Second, and more operationally deadly, is forgetting that a length demands both endpoints be marked at the same time in the measuring frame. If you naively transform two events that are simultaneous in the rest frame, you land on events at different times in the moving frame, the object drifts between them, and you recover the wrong (often uncontracted, or even dilated) answer. Impose $\Delta t = 0$ in the frame doing the measuring, remember that $L_0$ is the largest length, and that only the parallel dimension changes — get those three right and contraction never bites you.

Worked examples

1 · Why cosmic-ray muons reach the ground (fully numeric)

Muons are created about $15\ \text{km}$ up when cosmic rays strike the atmosphere. A muon's proper mean lifetime is $\tau_0 = 2.2\ \mu\text{s}$ , and a typical fast muon travels at $v = 0.995c$ . Newtonian physics says almost none should reach the ground; relativity says about a tenth do. Show both, once via length contraction in the muon's frame and once via time dilation in the ground frame, and confirm they agree.

Step 1. Lorentz factor. With $\beta = 0.995$ , $\gamma = 1/\sqrt{1-0.995^2} = 1/\sqrt{1-0.990025} = 1/\sqrt{0.009975} = 1/0.09987 = \mathbf{10.0}$ . This is the red point on the contraction curve.

Step 2. Newtonian estimate (the puzzle). In one lifetime a muon moving at $0.995c$ covers $v\tau_0 = (0.995)(2.998\times10^8)(2.2\times10^{-6}) = \mathbf{657\ \text{m}}$ , about $0.66\ \text{km}$ . To survive from $15\ \text{km}$ non-relativistically it would need to last $\sim 15/0.66 \approx 23$ lifetimes, a survival fraction $e^{-23} \approx 10^{-10}$ — essentially none. That contradicts the observed flux, and relativity is what rescues it.

Step 3. Muon frame — length contraction. In the muon's rest frame the atmosphere rushes past at $0.995c$ , so its $15\ \text{km}$ thickness is contracted to $L = L_0/\gamma = 15\ \text{km}/10 = \mathbf{1.5\ \text{km}}$ . The muon must survive only the time to let $1.5\ \text{km}$ sweep by: $t' = L/v = 1500/(0.995\times2.998\times10^8) = 5.03\times10^{-6}\ \text{s} = 5.03\ \mu\text{s}$ , i.e. $2.29$ lifetimes. Survival fraction $e^{-5.03/2.2} = e^{-2.29} = \mathbf{0.10}$ .

Step 4. Ground frame — time dilation, as a cross-check. Here the full $15\ \text{km}$ must be crossed: $t = 15000/(0.995\times2.998\times10^8) = 50.3\ \mu\text{s}$ . But the muon's own clock, dilated, advances only $t/\gamma = 50.3/10 = 5.03\ \mu\text{s}$ during the trip — the same proper time as before. Survival fraction $e^{-5.03/2.2} = \mathbf{0.10}$ . Both frames predict about $10\%$ of the muons reach the ground, with contraction and dilation as the two frame-dependent names for one invariant fact — the proper time $5.03\ \mu\text{s}$ elapsed on the muon's clock.

2 · The pole-and-barn paradox, resolved with numbers

A runner carries a pole of proper length $10\ \text{m}$ toward a barn of proper length $5\ \text{m}$ with a door at each end. How fast must the runner go for the pole to fit momentarily inside the barn in the barn's frame — and how is this consistent with the runner, who sees the barn as even shorter than the pole?

Step 1. Barn frame requirement. The barn measures the moving pole as $L = 10/\gamma$ . For the pole to fit between the two doors we need $L \le 5\ \text{m}$ , i.e. $10/\gamma \le 5$ , so $\gamma \ge 2$ . At exactly $\gamma = 2$ the pole is contracted to $10/2 = \mathbf{5\ \text{m}}$ and just fits, both doors able to be shut for an instant with the pole entirely inside.

Step 2. Required speed. $\gamma = 2$ gives $\sqrt{1-\beta^2} = 1/2$ , so $1-\beta^2 = 1/4$ , $\beta^2 = 3/4$ , and $v = \sqrt{3}/2\,c \approx \mathbf{0.866c}$ — the amber point on the curve.

Step 3. Runner's frame — the apparent contradiction. To the runner the pole is a full $10\ \text{m}$ , while the barn is now the moving object and is contracted to $5/\gamma = 5/2 = \mathbf{2.5\ \text{m}}$ . A $10\ \text{m}$ pole cannot possibly fit inside a $2.5\ \text{m}$ barn, so the runner denies it was ever enclosed. Both observers seem right and the "paradox" is that they disagree about a physical event.

Step 4. Resolution by simultaneity. "The pole fits" means both doors are shut at the same time with the pole inside — and simultaneity is frame-dependent. In the barn frame the two door-closings are simultaneous; in the runner's frame they are separated by $\Delta t' = \gamma v\,\Delta x/c^2 = (2)(0.866c)(5\ \text{m})/c^2 = 8.66/(2.998\times10^8) = \mathbf{28.9\ \text{ns}}$ . The far door opens to let the pole's front out before the near door shuts behind its back, so at no single runner-frame instant is the pole trapped. There is no contradiction: "enclosed" is not an invariant, and the $28.9\ \text{ns}$ gap in door-closings is exactly what reconciles the two accounts.

Physics of the portrait

Length contraction is unusual among physical laws in that it was written down, as a bold ad hoc rescue, more than a decade before anyone understood what it meant. The name attached to it belongs to the physicist who made it part of a systematic transformation of space and time.

Origins · 1 minHendrik Lorentz and how this was discovered

Hendrik Lorentz

1853–1928 · the electron theory and the transformations of space and time

Hendrik Antoon Lorentz was the towering figure of theoretical physics between Maxwell and Einstein, and the contraction is one thread of a life's work on how matter and the electromagnetic field share the same space. The puzzle that provoked it was the Michelson–Morley experiment of 1887, which stubbornly found no motion of the Earth through the supposed luminiferous ether. In 1892, and independently of George FitzGerald who had suggested the same idea in 1889, Lorentz proposed that a body moving through the ether physically shrinks along its direction of motion by just the factor $\sqrt{1 - v^2/c^2}$ — enough to null the experiment's fringe shift. For Lorentz this was at first a dynamical hypothesis about how intermolecular forces, themselves electromagnetic, would deform under motion. Over the next decade he built it into his "theory of corresponding states," introducing a fictitious "local time" for moving frames and, by 1904, the full set of coordinate transformations that Poincaré would name after him — the same equations from which contraction is derived in this article. Lorentz shared the 1902 Nobel Prize with Zeeman for the electron theory of the Zeeman effect. What he never fully embraced was the last step: Einstein's 1905 recognition that the contraction and the "local time" are not dynamical accidents of the ether but kinematic features of space and time themselves, symmetric between all inertial frames and requiring no ether at all. The mathematics is Lorentz's; the meaning is Einstein's; and Minkowski's 1908 geometry showed the two were one.

History

1887

— Michelson and Morley find no motion of the Earth through the ether

1889

— FitzGerald proposes that bodies contract along their direction of motion

1892

— Lorentz independently advances the contraction to explain the null result

1904

— Lorentz publishes the full transformation equations that carry his name

1905

— Einstein derives contraction kinematically from two postulates, discarding the ether

1908

— Minkowski recasts contraction as the geometry of spacetime

Going further

References. Misner, Thorne & Wheeler, Gravitation, §1.3, §2.1 · Wald, General Relativity, Ch. 1 · Carroll, Spacetime and Geometry, §1.2 · Hartle, Gravity, Ch. 4 · Schutz, A First Course in General Relativity, Ch. 1. All formulae are standard special-relativity results. Key numerical anchors: $\gamma = 1.25$ at $v = 0.6c$ , $\gamma = 2$ at $v = \sqrt{3}/2\,c \approx 0.866c$ , and $\gamma = 10.0$ at $v = 0.995c$ ; the muon survival fraction $e^{-2.29} = 0.10$ follows from both the contracted $1.5\ \text{km}$ path and the dilated $5.03\ \mu\text{s}$ proper time; and the pole-and-barn door-closing gap is $\Delta t' = \gamma v\,\Delta x/c^2 = 28.9\ \text{ns}$ . The collider figures ( $\gamma \sim 100$ at RHIC, $\gamma \sim 2500$ at the LHC) are order-of-magnitude machine values.

Builds on: Postulates of Special Relativity · Lorentz Transformations · Relativity of Simultaneity · Proper Length. Leads to: Time Dilation · Spacetime Interval · Lorentz Boosts and Rapidity.

Ask the tutor: Ask me to walk you through the pole-and-barn or twin scenario on a spacetime diagram, and watch the contraction fall out of the tilt of "now."

Is the moving object "really" shorter, or does it just look shorter?

It is really shorter as measured in that frame — but "measured" and "looks" are different operations, and the distinction matters. Length contraction is what you get by marking both ends simultaneously and subtracting; it is a genuine, reproducible coordinate fact of the measuring frame, not an optical illusion. What a single camera or eye sees, however, is different again: because light from the far end left earlier than light from the near end, a fast object can appear rotated or even skewed rather than simply squashed (the Penrose–Terrell effect). So contraction is real as a measurement, but the naive "photograph" of a relativistic object is a separate, more subtle question.

Which length is the "true" one, L or L₀?

Both are true; neither is privileged. $L_0$ , the proper length, is special only in being the length in the object's own rest frame and the largest value any inertial observer assigns — but it is not more "real" than the contracted $L$ measured elsewhere. Relativity's whole point is that there is no preferred frame, so there is no frame-independent "the length." What is invariant is the spacetime interval between the endpoint events; the proper length is the particular slice of it taken along the rest-frame simultaneity surface. Different observers, slicing differently, legitimately report different lengths, all consistent with the same underlying geometry.

Why does the perpendicular direction not contract at all?

Because measuring a transverse length needs no simultaneity convention, and that removes the loophole contraction exploits. The ring argument in the third derivation makes it airtight: if a moving ring were transversely smaller, then by symmetry the other ring (moving in the first one's frame) would also have to be smaller, and the two would leave contradictory scratch marks — a physical impossibility, since a permanent scratch is frame-independent. The only assignment consistent for both frames is that transverse sizes are unchanged. Parallel lengths escape this argument precisely because they require marking two ends "at the same time," and the two frames disagree about what that means.

How is length contraction related to time dilation?

They are the same relativistic effect projected onto different axes. Boosting to a moving frame is a hyperbolic rotation of spacetime that mixes the space and time coordinates; the projection onto the space axis is length contraction and the projection onto the time axis is time dilation, both scaled by the identical factor $\gamma$ . The muon worked example shows the equivalence concretely: the ground observer explains the muon's survival by a dilated decay clock, while the muon explains it by a contracted atmosphere, and the two accounts agree exactly because they compute the same invariant proper time. You can never have one without the other.

Does an object need to accelerate or feel a force to become contracted?

No. Contraction depends only on the relative velocity between the object and the observer, not on any history of acceleration and not on any force. A rod that has always drifted inertially past you is already contracted in your frame, and a rod at rest beside you is not, even though neither ever accelerated in your presence. The contraction is a relation between the object's worldtube and your simultaneity surface, established the instant you specify a relative speed. Acceleration matters for other relativistic stories (the twin paradox turns on it), but plain length contraction is a pure function of relative velocity through $\gamma$ .

At everyday speeds, is there any contraction at all?

Yes, but it is fantastically small because the effect is second-order in $v/c$ . Expanding for $\beta \ll 1$ gives $L/L_0 \approx 1 - \tfrac{1}{2}\beta^2$ , so a jetliner at $250\ \text{m/s}$ has $\beta \approx 8\times10^{-7}$ and a fractional contraction of only $\sim 3\times10^{-13}$ — a $70\ \text{m}$ aircraft shortened by about the width of an atomic nucleus. This quadratic smallness is exactly why length was believed absolute for all of pre-relativistic physics: you need speeds a substantial fraction of $c$ , as in accelerators or cosmic rays, before the $\sqrt{1-\beta^2}$ factor departs measurably from one.