Relativity of Simultaneity

Three of the most famous puzzles in relativity look like three different problems. A pole is sealed inside a barn too short for it. A traveller comes home younger than her twin, and cannot account for six of her sister's years. A thread strung between two identical rockets snaps for no reason anyone can point at. They are the same problem. All three are what happens when you ask a question whose answer turns on the word now — and forget that now is not a feature of the world. It is a slice through the world, and every observer cuts a different one. The offset formula itself was derived in `aq-sr-relativity-of-simultaneity-s1` and is not re-derived here. What is done here is to give it its geometric name — a frame's "now" is the spacelike hyperplane Minkowski-orthogonal to its four-velocity, and boosting tilts it — and then to use it as a key. It opens all three doors. It also settles the single most persistent numerical error in this chapter: why two of the three offsets carry a factor of $\gamma$ and one does not. The rule is one sentence long, and it is exact.

Key takeaways

Signature $(+,-,-,-)$ : $\Delta s^2 = c^2\Delta t^2 - \Delta x^2$ , timelike $>0$ , spacelike $<0$ . Four-velocity $u^\mu = \gamma(c,\mathbf v)$ , with $u_\mu u^\mu = c^2$ .
The offset $\Delta t' = -\gamma v\Delta x/c^2$ is recalled from `aq-sr-relativity-of-simultaneity-s1`, not re-derived.
Owned result: $\;t' = u_\mu x^\mu/c^2$ . A frame's surfaces of simultaneity are the level sets of $u_\mu x^\mu$ — the hyperplanes Minkowski-orthogonal to its four-velocity. The orthogonal complement of a timelike vector is spacelike, so a "now" is a spacelike hypersurface. Boosting rotates $u^\mu$ ; the hyperplane tilts with it.
Owned result — the $\gamma$ rule, which fixes a live corpus-wide defect: $\gamma$ appears when the answer is an interval of PRIMED coordinate time. It does NOT appear when the answer is what an UNPRIMED clock READS — because an unprimed clock's reading is unprimed coordinate time. The two differ by exactly $\gamma$ .
One key, three doors: `aq-sr-pole-and-barn-paradox` ( $41.32$ ns, $\gamma$ ), `aq-sr-twin-paradox-asymmetry` ( $6.4$ yr, no $\gamma$ ), `aq-sr-bell-spaceship-paradox` ( $192.58$ ns, $\gamma$ ).
Key idea: simultaneity is relative; causal order is not. A boost can reorder two events precisely when no signal could have joined them — so relativity buys you the freedom to re-slice, and charges you nothing.

The core insight

Every inertial observer carries a four-velocity, and every four-velocity comes with a hyperplane at right angles to it — right angles in the Minkowski sense, $u_\mu \Delta x^\mu = 0$ . That hyperplane is her "now": a slice of spacetime on which every displacement is spacelike. Boost her and the four-velocity swings; the hyperplane swings with it, tilting toward the light cone. Nothing about spacetime changed. Only the slice did. Every "paradox" in this chapter is a question quietly asked about a slice and mistakenly answered as though it were about the world.

Where this lives today

The hyperplane picture is not decoration; it is how modern physics formulates dynamics. "Give me the state of the field everywhere at one time and I will evolve it forward" — and "at one time" means on a spacelike hypersurface, nothing else. A relativistic field theory's initial-value problem is posed on a Cauchy surface, a spacelike hypersurface each worldline crosses exactly once, and the demand that it be spacelike is exactly Step 4 above: no two points of the initial data may influence each other, or the data would not be free. In canonical general relativity the ADM formalism is the machinery of such slicings, and the freedom to pick one is the relativity of simultaneity promoted to a gauge symmetry. Quantum field theory inherits the same discipline as microcausality: field operators at spacelike separation must commute. Otherwise an observer could exploit a re-slicing to push influence across a spacelike interval — and, because another observer's slice reverses the order, into their own past. The commutator condition is exactly what makes the frame-dependence of simultaneity harmless. At the practical end, GPS is a simultaneity machine. A receiver solves for its position and its own clock offset, from signals that left twenty-odd satellites at twenty-odd different moments. There is no shared "now" — only emission events, light cones, and one intersection. GPS does not correct for the relativity of simultaneity; it is built out of it.

Origins · 2 minSpacelike hypersurface and how this was discovered

Spacelike hypersurface

1898 — present · Paris · Göttingen · Bern

The concept has no single discoverer, and that is fitting: it was assembled by three people asking three different questions, none of whom used the word. Henri Poincaré asked first, and philosophically. In "La mesure du temps" (Revue de métaphysique et de morale 6, 1–13, 1898) he observed that "we have no direct intuition of simultaneity", and that two distant clocks can only be synchronised by a rule — exchange light signals and assume the two directions take equal time. Simultaneity, he concluded, is not given by nature but defined by a convention. In his 1900 Lorentz jubilee lecture he applied that synchronisation to Lorentz's "local time" and showed it was what a moving observer would actually measure. Poincaré saw that "now" was a choice. He did not take the next step. Albert Einstein took it in 1905, by refusing to treat the convention as merely a convention: he made light-signal synchronisation the definition of simultaneity, derived the Lorentz transformation from it, and thereby found that different inertial frames make different choices — and that both are right. His 1917 popular exposition put the train and the embankment into every classroom on Earth. Hermann Minkowski gave it the geometry, in "Raum und Zeit" (Cologne, 1908). Once spacetime has an indefinite metric, a frame's "now" is the orthogonal complement of its four-velocity — and that complement is a spacelike hypersurface, a slice on which no point can influence another. Boost, and the four-velocity swings; the slice swings with it. What Poincaré found puzzling and Einstein found necessary became, in Minkowski's hands, a triviality of linear algebra — and every later use of the idea, from Cauchy surfaces to the ADM slicing of general relativity, is that one sentence repeated.

History

1898

— Henri Poincaré, in "La mesure du temps" (Revue de métaphysique et de morale 6, 1), argues that simultaneity at a distance is defined by a convention, not perceived.

1900

— Poincaré, in the Lorentz jubilee volume, interprets Lorentz's local time as what a moving observer measures using light-signal synchronisation.

1905

— Einstein makes light-signal synchronisation the definition of simultaneity in "Zur Elektrodynamik bewegter Körper", and derives the relativity of simultaneity from the two postulates.

1908

— Minkowski, in "Raum und Zeit", recasts a frame's now as the hyperplane orthogonal to its four-velocity: a spacelike hypersurface.

1917

— Einstein's popular exposition Über die spezielle und die allgemeine Relativitätstheorie puts the train-and-embankment argument into general circulation.

1959

— Dewan and Beran, in Am. J. Phys. 27, 517, turn the simultaneity offset into a physical stress: the seed of the Bell spaceship paradox.

1961

— Rindler, in Am. J. Phys. 29, 365, resolves a length-contraction paradox by the same offset, in the rod-and-grid form.

The physical picture

The result this article stands on is owned elsewhere and quoted here verbatim. From `aq-sr-relativity-of-simultaneity-s1`, which derives it from the Lorentz transformation: if two events are simultaneous in $S$ ( $\Delta t = 0$ ) and separated by $\Delta x$ along the boost axis, then in a frame $S'$ moving at $v$ ,

\Delta t' = \gamma\!\left(\Delta t - \frac{v\,\Delta x}{c^2}\right) = -\frac{\gamma\, v\,\Delta x}{c^2} .

That is the whole of the machinery. What is not owned elsewhere — and what this article supplies — is the answer to two questions the formula alone does not settle.

First: what kind of object is a "now"? The formula gives a number; the geometry gives a surface. Every inertial observer carries a four-velocity $u^\mu = \gamma(c, \mathbf v)$ , a timelike vector along her worldline. Contract it with a displacement:

u_\mu \Delta x^\mu = \gamma\left(c\cdot c\Delta t - \mathbf v\cdot\Delta\mathbf x\right) = c^2\,\gamma\!\left(\Delta t - \frac{\mathbf v\cdot\Delta\mathbf x}{c^2}\right) = c^2\,\Delta t' .

So $\Delta t' = u_\mu\Delta x^\mu / c^2$ : the surfaces of constant $t'$ are the level sets of $u_\mu x^\mu$ — the hyperplanes Minkowski-orthogonal to her four-velocity. Her "now" is not a formula; it is a plane, fixed the moment her velocity is. Change the velocity and the plane rotates. That is the entire relativity of simultaneity, said once, geometrically.

Second: which $\gamma$ ? This is where the corpus bleeds. Ask "what is the $S'$ time separation of two events simultaneous in $S$ and $\Delta x$ apart in $S$ ?" and the answer is $\gamma v\Delta x/c^2$ — carries $\gamma$ . Ask "two clocks at rest and synchronised in $S$ , a rest distance $L_0$ apart: at one instant of $S'$ , what do their faces READ?" and the answer is $vL_0/c^2$ — no $\gamma$ . Both are correct; they are different questions. The rule that tells them apart is one sentence, which we prove below: $\gamma$ appears when the answer is an interval of $S'$ coordinate time; it does not appear when the answer is the reading of an $S$ -clock — because an $S$ -clock's reading is $S$ coordinate time. Moving clocks tick at $1/\gamma$ of the $S'$ coordinate rate, and that factor is exactly what cancels the $\gamma$ . Get this wrong and you print $2Lv\gamma/c^2$ for the twin's turnaround jump — as two of the base corpus's own articles do — and your books fail to balance by four years.

Figure

PUNCHLINE — one key, three doors. The same tilted hyperplane opens all three paradoxes, and the γ rule sorts out which offset carries a Lorentz factor. LEFT: boosting rotates the four-velocity u and the hyperplane orthogonal to it scissors toward the light cone by the same angle, arctan β = 30.96° at β = 0.6. RIGHT: the three offsets, with real numbers. The barn (41.32 ns) and Bell (192.58 ns) ask for an interval in the PRIMED frame's coordinate time — those carry γ. The twin's turnaround jump (6.4 yr) asks what an EARTH clock READS — and an Earth clock's reading IS Earth coordinate time, so it carries none

The core equations

Recalled, not derived — the offset, from `aq-sr-relativity-of-simultaneity-s1`:

\Delta t' = \gamma\!\left(\Delta t - \frac{v\,\Delta x}{c^{2}}\right) \;\xrightarrow{\;\Delta t = 0\;}\; \Delta t' = -\frac{\gamma\,v\,\Delta x}{c^{2}} \qquad (1)

where: $\Delta x$ is measured in $S$ , the frame in which the two events are simultaneous. Physical reading: the leading event in $S$ (larger $x$ ) happens earlier in $S'$ . The effect is proportional to the separation, not to the elapsed time — which is why it is invisible for co-located events and enormous for distant ones. Now the article's owned geometry. With four-velocity $u^\mu = \gamma(c,\mathbf v)$ and displacement $\Delta x^\mu = (c\Delta t, \Delta\mathbf x)$ :

\boxed{\;\Delta t' = \frac{u_\mu \Delta x^\mu}{c^{2}}\;}\qquad\Longrightarrow\qquad \boxed{\;\text{“now” in } S' \;=\; \{\,\Delta x^\mu : u_\mu \Delta x^\mu = 0\,\}\;} \qquad (2)

where: $u_\mu u^\mu = \gamma^2(c^2 - v^2) = c^2 > 0$ , so $u$ is timelike (in our signature, timelike is positive). Physical reading: an observer's set of simultaneous events is the hyperplane Minkowski-orthogonal to her four-velocity. Because the orthogonal complement of a timelike vector contains only spacelike vectors, that hyperplane is a spacelike hypersurface — hence the name. Boost her, and $u^\mu$ swings through a hyperbolic rotation; the hyperplane swings with it. In a $(ct,x)$ diagram both the $ct'$ -axis and the $x'$ -axis tilt toward the light line by the same angle $\arctan\beta$ — they scissor, and the light cone, which is its own orthogonal complement, does not move at all. And the owned disambiguation — the two offsets, and the $\gamma$ that separates them:

\underbrace{\Delta t'_{\text{coord}} = -\frac{\gamma v \Delta x}{c^{2}}}_{\text{an } S' \text{ coordinate interval}} \qquad\text{versus}\qquad \boxed{\;\underbrace{\Delta \tau_{\text{read}} = -\frac{v L_0}{c^{2}}}_{\text{what } S\text{-clocks READ}}\;} \qquad (3)

where: $L_0$ is the rest separation of two clocks at rest in $S$ and synchronised there. Physical reading: "leading clocks lag" by $vL_0/c^2$ , with NO $\gamma$ . The ratio of the two quantities in (3) is exactly $\gamma$ , and the reason is that the $S$ -clocks, viewed from $S'$ , tick at $1/\gamma$ of the $S'$ coordinate rate — so an $S'$ coordinate gap of $\gamma vL_0/c^2$ registers on their faces as a reading gap of $vL_0/c^2$ . Do not add a $\gamma$ to the boxed formula. Do not remove one from the unboxed formula. Finally, the boundary that keeps relativity honest:

\text{order of two events is frame-dependent} \iff (\Delta s)^{2} < 0 \iff |c\,\Delta t| < |\Delta x| \qquad (4)

where: the reversing boost is $v_{\text{crit}} = c^2\Delta t/\Delta x$ , and $v_{\text{crit}} < c$ precisely when the interval is spacelike. Physical reading: simultaneity is relative; causal order is not. You may re-slice freely, and you can never re-slice a cause after its effect, because doing so would require $v > c$ . Relativity gives away the slice and keeps the causality.

The geometry of a now

Derivation AYour now is the hyperplane orthogonal to your four-velocity

The relativity of simultaneity is usually taught as an algebraic surprise — a term $-\gamma v\Delta x/c^2$ that appears in the Lorentz transformation and refuses to vanish. True, but the wrong altitude. At the right altitude it is a statement about orthogonality, and seen that way the tilting of the "now" is no more mysterious than the fact that turning your head turns your field of view. This derivation supplies that altitude, and it is what earns this article its portrait term.

Step 1. Write the observer's four-velocity: $u^\mu = \gamma(c, \mathbf v)$ , the tangent to her worldline, normalised by $u_\mu u^\mu = \gamma^2(c^2 - v^2) = c^2$ . Why: in $(+,-,-,-)$ a timelike vector has positive square. The four-velocity is the one geometric object an inertial observer carries with her, and everything about "her frame" must be expressible in terms of it.

Step 2. Contract with a displacement $\Delta x^\mu = (c\Delta t, \Delta\mathbf x)$ , lowering with $\eta_{\mu\nu} = \mathrm{diag}(+,-,-,-)$ :

u_\mu \Delta x^\mu = \gamma c \cdot c\Delta t \;-\; \gamma\,\mathbf v\cdot\Delta\mathbf x = c^2\,\gamma\!\left(\Delta t - \frac{\mathbf v\cdot\Delta\mathbf x}{c^2}\right) = c^2\,\Delta t' .

Why: the right-hand side is precisely the Lorentz transformation of the time coordinate. We have not derived anything new — we have recognised that the transformed time is a scalar product, and scalar products are geometry.

Step 3. Read off the surface. Two events are simultaneous in $S'$ iff $\Delta t' = 0$ iff $\boxed{u_\mu \Delta x^\mu = 0}$ — iff their separation is Minkowski-orthogonal to $u$ . The set of such $\Delta x^\mu$ is a three-dimensional hyperplane through the observer. Why: her "now" is not a formula and not a convention in disguise; it is the orthogonal complement of her four-velocity, determined by her velocity and nothing else.

Step 4. Prove the hyperplane is spacelike, which is what makes it a legitimate "now". Work in $u$ 's own rest frame, where $u^\mu = (c, \mathbf 0)$ . Then $u_\mu w^\mu = c\,w^0 = 0$ forces $w^0 = 0$ , so $w^\mu = (0, \mathbf w)$ and $w_\mu w^\mu = -|\mathbf w|^2 < 0$ for $\mathbf w \ne 0$ . Every non-zero vector in the hyperplane is spacelike. Since $u_\mu w^\mu$ is a scalar, this holds in every frame. Why: this is the theorem that licenses the whole idea of a "now". A slice on which every separation is spacelike is a slice on which no event can influence another — a genuine instant, containing no causation. That is what a spacelike hypersurface is, and it is why the concept has a name.

Step 5. Watch it tilt. In a $(ct, x)$ diagram $u^\mu \propto (1, \beta)$ , so the $ct'$ -axis makes an angle $\arctan\beta$ with the $ct$ -axis; and $\Delta t' = 0$ gives $c\Delta t = \beta\,\Delta x$ , so the $x'$ -axis makes the same angle with the $x$ -axis, on the other side. At $\beta = 0.6$ : $\arctan(0.6) = 30.96°$ — both axes swing $30.96°$ toward the $45°$ light line, closing on it symmetrically. Why: the famous scissoring of the Minkowski diagram, now a corollary rather than a picture. The light cone is fixed because a null vector is its own orthogonal complement — a boost slides null rays along themselves and never tilts them.

Step 6. Collect the payoff. A "now" is a spacelike hyperplane; there is one for every four-velocity; boost, and the hyperplane follows. No slice is privileged, and no slice is a fact about the world. Every paradox in this chapter presupposed a privileged slice. Why: this is the master key, and the next card cuts it. *Verified in verify_aq-sr-relativity-of-simultaneity-s3.py: $u_\mu u^\mu = c^2$ symbolically; $u_\mu\Delta x^\mu = c^2\Delta t'$ against the explicit Lorentz transform; the orthogonal complement of $u$ shown to contain only spacelike vectors; and $\arctan(0.6) = 30.96°$ .*

Go deeper · Derivation B · 3 minThe gamma rule: two offsets, one factor, and a four-year error

Derivation BThe gamma rule: two offsets, one factor, and a four-year error

Now the disambiguation — the most useful thing in this article, and the one that repairs a live defect in the corpus. Two quantities share the same units, the same $vL/c^2$ shape, and the same physical origin, and one carries a $\gamma$ while the other does not. Confusing them is how the twin paradox's turnaround jump acquires a spurious Lorentz factor in print, and how the books stop balancing. We derive both, prove their ratio is exactly $\gamma$ , and state the rule that separates them.

Step 1. Set up two clocks. Clocks $A$ and $B$ are at rest in $S$ , at $x = 0$ and $x = L_0$ , synchronised in $S$ — so at any $S$ -time $t$ , both faces read $t$ . An observer $S'$ moves at $+v$ . In her frame the clocks sweep past her at $-v$ , so $A$ is the leading clock (it is the one further along the direction the pair is travelling). Why: "leading" is defined by the clocks' motion in $S'$ , and getting this backwards flips the sign of the whole result.

Step 2. Ask the first question — the coordinate one. Two events simultaneous in $S$ ( $\Delta t = 0$ ) separated by $\Delta x = L_0$ have, by (1), an $S'$ coordinate-time separation

\Delta t'_{\text{coord}} = -\frac{\gamma v L_0}{c^{2}} .

Why: this is a difference of $t'$ values. It carries $\gamma$ , and it must, because it is an interval of the primed frame's coordinate time.

Step 3. Ask the second question — the reading one. What do the two clock faces show at a single instant $t'$ of $S'$ ? Invert the transform on each worldline. Clock $A$ ( $x = 0$ ): $t' = \gamma t \Rightarrow t = t'/\gamma$ . Clock $B$ ( $x = L_0$ ): $t' = \gamma(t - vL_0/c^2) \Rightarrow t = t'/\gamma + vL_0/c^2$ . So at a single $S'$ -instant,

\text{face of }B \;-\; \text{face of }A \;=\; \frac{vL_0}{c^{2}}, \qquad\text{i.e.}\qquad \boxed{\;\text{the LEADING clock } A \text{ lags by } \frac{vL_0}{c^{2}}\;}

Why: no $\gamma$ . And this is not an approximation or a special case — it is exact, for every $\beta$ .

Step 4. Prove the ratio is exactly $\gamma$ , and say why. Divide: $\Delta t'_{\text{coord}} / \Delta\tau_{\text{read}} = \gamma$ . The reason is time dilation, applied once. The two clocks are moving in $S'$ , so their faces advance at $1/\gamma$ of the $S'$ coordinate rate. A coordinate gap of $\gamma v L_0/c^2$ therefore registers on their faces as a reading gap of $\tfrac{1}{\gamma}\cdot\gamma vL_0/c^2 = vL_0/c^2$ . Why: the $\gamma$ of the transformation and the $1/\gamma$ of time dilation cancel. That cancellation is the entire content of the discrepancy, and once you have seen it you cannot un-see it.

Step 5. State the rule. A factor of $\gamma$ appears when the answer is an interval of the PRIMED frame's coordinate time. It does not appear when the answer is what an UNPRIMED-frame clock READS — because an unprimed clock's reading is unprimed coordinate time. Why: one sentence, and it decides every case in this chapter without further thought. Ask yourself which frame's clock you are reading, and the $\gamma$ is determined.

Step 6. Apply it to the twin, and price the error. At the turnaround the traveller asks: what does the Earth clock read now? That is a reading of an $S$ -clock. No $\gamma$ . Before the turn her slice puts Earth at $L/v - vL/c^2$ ; after, at $L/v + vL/c^2$ ; the jump is $2vL/c^2$ . At $\beta = 0.8$ , $L = 4$ ly: $5 - 3.2 = 1.8$ yr, then $5 + 3.2 = 8.2$ yr, a jump of $6.4$ yr. Her two legs credit Earth with $2L/(v\gamma^2) = 3.6$ yr, and $3.6 + 6.4 = 10.0$ — Earth's actual elapsed time, exactly. Insert the spurious $\gamma$ and the jump becomes $10.67$ yr, the audit totals $14.27$ yr, and $4.27$ years of Earth history have been invented out of nothing. Why: this is not a stylistic quibble. It is a wrong number, it is in print, and the rule above is what catches it in one line. *Verified in verify_aq-sr-relativity-of-simultaneity-s3.py: both offsets derived symbolically from the Lorentz transform; their ratio shown to be exactly $\gamma$ for all $\beta$ ; the leading-clock sign; and the twin audit $3.6 + 6.4 = 10.0$ closing, against the $\gamma$ -contaminated $3.6 + 10.67 = 14.27$ .*

Go deeper · Derivation C · 2 minOne key, three doors

Derivation COne key, three doors

The claim advertised in the overview is that the barn, the twin and the thread are one problem. This card cashes it: each paradox is equation (1) evaluated on a different pair of events, and the $\gamma$ rule of the previous card decides, in each case, whether a Lorentz factor rides along. Nothing new is computed — that is exactly the point. Three famous puzzles, one formula, three substitutions.

Step 1. Open the barn (`aq-sr-pole-and-barn-paradox`). Events: two door-closings, simultaneous in the barn frame, separated by the barn's depth $D$ . Question asked: their separation in the pole frame's coordinate time. That is a primed interval, so $\gamma$ rides:

\Delta t' = -\frac{\gamma v D}{c^{2}} = -41.32\ \text{ns} \qquad (D = 6\ \text{m},\ \beta = 0.9,\ \gamma = 2.2942).

Why: the pole "does not fit" only because the runner's slice is tilted. Untilt it and the paradox is gone — and so is the pole's excuse for being outside the barn.

Step 2. Open the twin (`aq-sr-twin-paradox-asymmetry`). Events: what an Earth clock reads, on the traveller's slice, immediately before and immediately after she changes frames. Question asked: a reading of an unprimed clock, so no $\gamma$ :

\Delta t_{\text{jump}} = \frac{2vL}{c^{2}} = 6.4\ \text{yr} \qquad (L = 4\ \text{ly},\ \beta = 0.8).

Why: the traveller's "missing" six years of her sister's life are the Earth-events her slice swept past when it swung. She did not lose them; she re-sliced across them.

Step 3. Open the thread (`aq-sr-bell-spaceship-paradox`). Events: two engine ignitions, simultaneous in the launch frame, separated by the ships' coordinate spacing $L$ . Question asked: their separation in the comoving frame's coordinate time. Primed interval, so $\gamma$ rides:

\Delta t' = -\frac{\gamma v L}{c^{2}} = -192.58\ \text{ns} \qquad (L = 100\ \text{m},\ \beta = 0.5,\ \gamma = 1.1547).

Why: in the ships' own frame the front ship lit its engine first and has been accelerating longer — so it is pulling away, the gap grows, and the thread parts. The thread breaks because two events that the launch frame calls simultaneous, the ships do not.

Step 4. Notice what the three share, and what they do not. All three are equation (1); all three depend on a separation $\Delta x$ and vanish as $\Delta x \to 0$ . Two carry $\gamma$ and one does not, and the difference is not in the physics — it is in the question. Why: this is the unification. The paradoxes are not three phenomena; they are three ways of forgetting that a slice is not a fact.

Step 5. State the boundary, so the key is not mistaken for a skeleton key. In none of the three is any causal order reversed. The barn's door events are spacelike ( $(\Delta s)^2 = -D^2 < 0$ ); the ignition events are spacelike ( $-L^2$ ); the Earth events the twin re-slices across are spacelike from her. Simultaneity is relative; causal order is not, and the criterion is (4): a boost reverses two events' order iff $|c\Delta t| < |\Delta x|$ — iff no signal could have joined them. Why: relativity gives away exactly what it can afford. It hands you the freedom to re-slice and charges nothing, because the only orderings it lets you scramble are the ones nothing depended on. *Verified in verify_aq-sr-relativity-of-simultaneity-s3.py: all three offsets recomputed from the same Lorentz transform ( $41.32$ ns, $6.4$ yr, $192.58$ ns); the presence/absence of $\gamma$ in each checked against the rule; and the spacelike character of the barn's and Bell's event pairs confirmed via $(\Delta s)^2 = -D^2$ , $-L^2$ .*

Figure

The two offsets that differ by exactly γ, for Einstein's train: two clocks 300 m apart at rest on the platform, synchronised there, viewed from a carriage at β = 0.6 (γ = 1.25). LEFT: what the two clock FACES read at a single instant of the carriage — they differ by vL₀/c² = 600.4 ns, with NO γ. RIGHT: the coordinate-time separation, in the carriage's frame, of the two platform-simultaneous events — γvL₀/c² = 750.5 ns, WITH γ. The ratio is 1.25, on the nose. Same setup, two questions, one Lorentz factor between them — and this is the exact distinction that the twin paradox's turnaround jump gets wrong in print

Limiting cases

Case	Result	Meaning
$\Delta x = 0$	$\Delta t' = 0$	co-located events are simultaneous in every frame. Simultaneity is only relative at a distance
$v \to 0$	$\Delta t' \to 0$	Newtonian absolute simultaneity. The hyperplane stops tilting
$v \to c$	$\Delta t' \to -\Delta x/c$	the "now" plane closes onto the light cone and stops being a slice
$\|c\Delta t\| > \|\Delta x\|$	order absolute	timelike: $v_{\text{crit}} > c$ is unattainable. Cause precedes effect in every frame
$\|c\Delta t\| < \|\Delta x\|$	order reversible	spacelike: some legal boost flips it, and nothing depended on the order anyway
$\|c\Delta t\| = \|\Delta x\|$	order absolute, zero margin	null: a light ray's order survives every boost, but only just

Common trapThe single most-copied error in this chapter

The twin paradox's turnaround jump is $\;2Lv/c^2$ , not $2Lv\gamma/c^2$ . The base corpus prints the $\gamma$ version in two separate articles — a propagating defect — and it is wrong because the quantity wanted is the reading of an Earth clock, and Earth-clock readings are Earth coordinate time. The audit identity is the cheapest possible check:

\frac{2L}{v\gamma^{2}} + \frac{2Lv}{c^{2}} = \frac{2L}{v} \qquad\text{(closes, because } \tfrac{1}{\gamma^2}+\beta^2 = 1\text{)}

At $\beta = 0.8$ , $L = 4$ ly: $3.6 + 6.4 = 10.0$ ✓. With the spurious $\gamma$ : $3.6 + 10.67 = 14.27$ ✗ — four extra years of Earth history from one misplaced Lorentz factor. A simultaneity offset that does not close its books is not "approximately right". It is wrong.

Common trapRelative does not mean arbitrary

"Simultaneity is a convention" is a true sentence that is almost always over-read. Two things are not conventional. (i) Causal order: if a signal could have gone from $E_1$ to $E_2$ , then $E_1$ precedes $E_2$ in every frame — the reversing boost would need $v_{\text{crit}} = c^2\Delta t/\Delta x > c$ . (ii) The hyperplane itself: fix a four-velocity and the "now" is determined, not chosen — it is the orthogonal complement, and there is exactly one. What is conventional is which four-velocity you privilege, the same freedom as choosing where to stand in a room. (The conventionality-of-simultaneity literature, Poincaré 1898 onward, concerns a further freedom in the one-way speed of light; nothing here depends on it.)

Worked examples

1 · Einstein's train: both offsets, and the $\gamma$ between them

Two clocks sit $300$ m apart on the platform, synchronised in the platform frame. A carriage passes at $\beta = 0.6$ . (a) At one instant of the carriage's frame, what do the two clock faces read? (b) Two events simultaneous on the platform, at the two clocks: what is their time separation in the carriage's frame?

Step 1 — identify. $\gamma = (1-0.36)^{-1/2} = 1/0.8 = 1.25$ . These are two different questions, and the whole point of the example is that they have different answers.

Step 2 — (a), the readings. This asks what platform clocks show. Platform clocks read platform coordinate time, so no $\gamma$ :

\Delta\tau_{\text{read}} = \frac{vL_0}{c^2} = \frac{0.6 \times 300}{2.99792458\times10^8} = 6.004\times10^{-7}\ \text{s} = \mathbf{600.4\ \text{ns}} .

The leading clock — the one at the front, in the direction the platform is moving as the carriage sees it — lags by $600.4$ ns.

Step 3 — (b), the coordinate separation. This asks for an interval of the carriage's coordinate time, so $\gamma$ rides:

\Delta t'_{\text{coord}} = \frac{\gamma v L_0}{c^2} = 1.25 \times 600.4\ \text{ns} = \mathbf{750.5\ \text{ns}} .

Step 4 — the boundary number. $750.5 / 600.4 = 1.25 = \gamma$ , exactly. Same platform, same clocks, same speed — two answers differing by precisely one Lorentz factor, because they are answers to two different questions. Reach for the wrong one and you have committed the corpus's most common error.

Step 5 — sanity. At walking pace ( $v = 1.5$ m/s) the same $300$ m gives $vL_0/c^2 = 5\times10^{-15}$ s — five femtoseconds. That is why nobody noticed for three hundred years; and it is why the effect is nonetheless enormous when $\Delta x$ is a light-year.

2 · Which orders can a boost reverse?

Event $E_1$ at $(t,x) = (0,0)$ ; event $E_2$ at $(1\ \mu\text{s},\ 600\ \text{m})$ . Can any inertial observer see $E_2$ happen first? What if $E_2$ is instead at $(3\ \mu\text{s},\ 600\ \text{m})$ ?

Step 1 — identify. Compute $c\Delta t$ and compare with $\Delta x$ . That single comparison decides everything.

Step 2 — the first pair. $c\Delta t = (2.99792458\times10^8)(10^{-6}) = 299.79$ m, and $\Delta x = 600$ m. Since $299.79 < 600$ , the interval is spacelike: $(\Delta s)^2 = (299.79)^2 - 600^2 = -2.7012\times10^{5}\ \mathrm{m}^2 < 0$ .

Step 3 — find the reversing boost. $v_{\text{crit}} = c^2\Delta t/\Delta x = c\,(299.79/600) = \mathbf{0.49965\,c}$ . Anyone moving faster than that along $+x$ sees $E_2$ before $E_1$ — legal, unremarkable, and nothing depends on it: influence from $E_1$ to $E_2$ would have needed $600/299.79 = 2.0014$ times the speed of light.

Step 4 — the second pair. Now $c\Delta t = 899.38$ m $> 600$ m: timelike, $(\Delta s)^2 = +4.4888\times10^{5}\ \mathrm{m}^2 > 0$ . The reversing boost would need $v_{\text{crit}} = c(899.38/600) = 1.499\,c$ — unattainable.

Step 5 — the boundary number. $E_1$ precedes $E_2$ in every inertial frame, and a light signal has $299.79$ m of slack to spare. Simultaneity is relative; causation is not. The dividing line is exactly the light cone, and (4) says only that you cannot outrun light to get behind an event.

Going further

Texts. Rindler, Introduction to Special Relativity, 2nd ed., §§2, 8 · Taylor & Wheeler, Spacetime Physics, 2nd ed., ch. 3 · Minkowski, "Raum und Zeit" (1908) · Poincaré, "La mesure du temps", Revue de métaphysique et de morale 6, 1–13 (1898) · Misner, Thorne & Wheeler, Gravitation, ch. 21 (ADM slicing).

Provenance. Signature $(+,-,-,-)$ . Every number recomputed in verify_aq-sr-relativity-of-simultaneity-s3.py. Base defect repaired: the propagating $\gamma$ error in the twin turnaround jump ( $2Lv\gamma/c^2$ in two separate base articles; the truth is $2Lv/c^2$ ) is diagnosed here at its root and given a one-line audit test. The base's own "leading clocks lag" card is correct ( $-vL_0/c^2$ , no $\gamma$ ) and is preserved; what was missing — supplied here — is the rule saying why the same setup yields $\gamma vL_0/c^2$ when a different question is asked.

Scope. ⚠ This article is a recaller, not the owner. The derivation of the offset from the Lorentz transformation is owned by `aq-sr-relativity-of-simultaneity-s1`; the qualitative statement by `aq-sr-relativity-of-simultaneity`. Owned here: the four-velocity/hyperplane formulation, the $\gamma$ rule, and the unification of the three paradoxes — which belong to `aq-sr-pole-and-barn-paradox`, `aq-sr-twin-paradox-asymmetry` and `aq-sr-bell-spaceship-paradox`. Interval: `aq-sr-invariant-spacetime-interval`. Four-velocity: `aq-sr-four-velocity`.

Builds on. Lorentz Transformation · Constancy of the Speed of Light · Length Contraction. Leads to. The Twin Paradox · Asymmetry in the Twin Paradox · Relativistic Velocity Addition · Thomas–Wigner Rotation.

If simultaneity is a convention, is the relativity of simultaneity just bookkeeping?

No — and the distinction matters. Which four-velocity you stand on is your choice, in the same harmless sense that where you stand in a room is. Once it is fixed, the "now" is determined: the orthogonal complement, and there is exactly one. What is no convention at all is the consequence: because different observers' hyperplanes genuinely differ, a thread between two rockets genuinely snaps, a pole genuinely passes through a barn shorter than itself, and a traveller genuinely returns younger. Those are measurable, and no re-labelling removes them.

How do I decide, in a new problem, whether my answer should carry a γ?

Ask which frame's clock is giving you the number. Want an interval of the primed frame's coordinate time — the reading difference of two clocks at rest in $S'$ ? Then $\gamma$ rides. Want what an unprimed clock's face shows? No $\gamma$ , because an unprimed face is unprimed coordinate time. Then sanity-check against a conservation identity: the twin's audit ( $2L/(v\gamma^2) + 2Lv/c^2 = 2L/v$ ) is the model. A formula that fails to close its books is wrong, whoever printed it.

Can I always find a frame in which two given events are simultaneous?

Only if they are spacelike separated. The required boost is $v = c^2\Delta t/\Delta x$ , which is under $c$ exactly when $|c\Delta t| < |\Delta x|$ . For a timelike pair no such frame exists — but there is one in which they happen at the same place (the rest frame of an observer who attends both). Spacelike pairs can be made simultaneous; timelike pairs can be made co-located. The light cone is the boundary, and belongs to neither.

Does the tilting of the "now" plane mean the future already exists?

That is a metaphysical question in a physics costume, and this article does not answer it — but it does constrain it. The geometry says only that different observers slice the same four-dimensional structure differently, so an event in your "future" may lie in a passing observer's "now". Whether that makes the future real (the block-universe reading) or merely makes "now" a poor concept to do physics with is a matter on which honest philosophers disagree. What is not in dispute: no re-slicing lets you signal into your own past, because the reordering is confined to spacelike pairs. The physics is safe whichever way the metaphysics falls.

The effect is O(v/c²). Is it ever big?

Enormously, because it scales with the separation, not the speed. At walking pace across a $300$ m platform it is five femtoseconds. At $\beta = 0.8$ across four light-years it is $3.2$ years per slice — and $6.4$ years when the traveller swings from one slice to the other. It is the one relativistic effect that is unbounded: fix any $\beta > 0$ , take $\Delta x$ large enough, and the offset is as large as you please. That is why it, and not time dilation, is what the paradoxes are actually made of.

Why is the light cone unmoved by a boost, when everything else tilts?

Because a null vector is Minkowski-orthogonal to itself: $k_\mu k^\mu = 0$ . The $ct'$ - and $x'$ -axes both rotate toward the light line by the same angle $\arctan\beta$ , closing on it from either side and never crossing. What a boost does to a null ray is rescale it — by $e^{\pm\varphi}$ , the Doppler factor — never tilt it. The light cone is the fixed structure all the tilting happens around, which is another way of saying that the constancy of $c$ is what makes the geometry hang together.

Relativity of Simultaneity

Where this lives today

The physical picture

The core equations

The geometry of a now

Limiting cases

Worked examples

1 · Einstein's train: both offsets, and the $\gamma$ between them

2 · Which orders can a boost reverse?

Going further

Before this

Where this leads

Relativity of Simultaneity

Where this lives today

The physical picture

The core equations

The geometry of a now

Limiting cases

Worked examples

1 · Einstein's train: both offsets, and the γ\gammaγ between them

2 · Which orders can a boost reverse?

Going further

Before this

Where this leads

1 · Einstein's train: both offsets, and the $\gamma$ between them