§1Two Postulates, One Consequence
Einstein's 1905 paper "On the Electrodynamics of Moving Bodies" opens with a single observation: Maxwell's equations produce a wave speed $c$ that does not depend on the state of motion of the emitting body. Every other wave speed in classical physics does depend on the source. Sound from a moving train arrives at different frequencies depending on which way the train is moving. Maxwell said light is different. Einstein asked: what if Maxwell is simply right?
Two postulates, and only two:
The Two Postulates · Einstein 1905 · [VERIFIED]
I. Principle of Relativity: The laws of physics take the same form in all inertial (non-accelerating) reference frames. No experiment can detect absolute uniform motion — there is no preferred rest frame.
II. Constancy of c: The speed of light in vacuum is $c = 299{,}792{,}458$ m/s for all inertial observers, regardless of the motion of the source or the observer.
These two postulates together force a unique mathematical structure on spacetime. There is no freedom: the Lorentz transformation is the only transformation consistent with both. [VERIFIED — Einstein 1905, Ann. Phys. 17 891]
The conflict is immediate. Classical physics (Galilean relativity) adds velocities: if I move at $v$ and emit a sound at $u$ relative to me, the sound moves at $u + v$ relative to the ground. If this held for light, Postulate II would be false. One of the two must go. Einstein kept Postulate II and let the addition rule for velocities — and with it, absolute simultaneity — fall.
§2The Lorentz Boost as Hyperbolic Rotation
In Euclidean geometry, a rotation by angle $\theta$ in the $xy$-plane mixes $x$ and $y$ while preserving $x^2 + y^2$. The trigonometric functions $\cos\theta$ and $\sin\theta$ describe the mixing.
In Minkowski spacetime, a "boost" by velocity $v$ mixes $t$ and $x$ while preserving $-c^2t^2 + x^2$ (the spacetime interval). The hyperbolic functions $\cosh\phi$ and $\sinh\phi$ describe the mixing, where $\phi$ is the rapidity.
Euclidean rotation vs. Minkowski boost — same structure, different signature
Euclidean rotation (preserves x²+y²)
$x' = x\cos\theta - y\sin\theta$
$y' = x\sin\theta + y\cos\theta$
Parameter: angle $\theta \in [0, 2\pi)$
Functions: $\cos$, $\sin$ — circular
Preserved: $x^2 + y^2$
Closure: rotations form SO(2)
Minkowski boost (preserves −c²t²+x²)
$t' = t\cosh\phi - (v/c^2)x\sinh\phi \cdot c$
$x' = -ct\sinh\phi + x\cosh\phi$
Parameter: rapidity $\phi \in (-\infty, +\infty)$
Functions: $\cosh$, $\sinh$ — hyperbolic
Preserved: $-c^2t^2 + x^2$
Closure: boosts form SO(1,1)
In standard notation, the Lorentz transformation for a boost at velocity $v$ along the $x$-axis is:
t' = γ(t − vx/c²)
x' = γ(x − vt)
y' = y, z' = z
where γ = 1/√(1 − v²/c²) = cosh(φ), v/c = tanh(φ)
Theorem SR.1 · Invariance of ds² · [VERIFIED — Einstein 1905, Minkowski 1908]
Under any Lorentz boost, the quantity
$ds^2 = -c^2 dt^2 + dx^2 + dy^2 + dz^2$
is invariant. If $ds^2 = 0$ in one frame, it is zero in all frames. The set of events reachable from a given event at speed $\leq c$ is the same for all observers. The light cone is an absolute structure — frame-independent. [VERIFIED]
Rapidity is the right parameter because it is additive. Two boosts at rapidities $\phi_1$ and $\phi_2$ compose to rapidity $\phi_1 + \phi_2$ — just like two rotations add angles. Velocities do not add this way (the Galilean rule $v_1 + v_2$ fails at relativistic speeds). Rapidity is the "true angle" of the boost.
Rapidity is Additive · [VERIFIED]
If observer $B$ moves at rapidity $\phi_1$ relative to $A$, and $C$ moves at rapidity $\phi_2$ relative to $B$, then $C$ moves at rapidity $\phi_1 + \phi_2$ relative to $A$. In terms of velocities:
$v_{AC} = \frac{v_{AB} + v_{BC}}{1 + v_{AB}v_{BC}/c^2}$
At low speeds ($v \ll c$), this reduces to $v_{AC} \approx v_{AB} + v_{BC}$ (Galilean). At high speeds, the denominator prevents $v_{AC}$ from exceeding $c$. No finite number of boosts can reach or exceed $c$: $\tanh(\phi_1 + \phi_2) < 1$ for all finite $\phi_1, \phi_2$. [VERIFIED]
§3Three Consequences of the Boost
The Lorentz transformation has three immediate consequences, all experimentally confirmed:
| Consequence | Formula | Physical meaning | Status |
|---|
| Time dilation |
$\Delta t' = \gamma \Delta t$ |
A clock moving at speed $v$ runs slow by factor $\gamma$. A muon created in the upper atmosphere lives long enough to reach the ground because its "clock" runs slow relative to Earth's. |
[VERIFIED — muons, GPS, particle accelerators] |
| Length contraction |
$\Delta x' = \Delta x / \gamma$ |
A rod moving along its length is shorter by $\gamma$ in the observer's frame. The rod hasn't physically changed — the two frames disagree on which events are simultaneous at each end. |
[VERIFIED — heavy-ion collisions, synchrotrons] |
| Simultaneity breaking |
$\Delta t' = -\gamma v \Delta x / c^2$ |
Two events at the same time ($\Delta t = 0$) but different locations ($\Delta x \neq 0$) are not simultaneous in a boosted frame. There is no absolute "now" — only local, frame-dependent "nows". |
[VERIFIED — atomic clocks on aircraft, Hafele–Keating 1971] |
All three are not paradoxes — they are consistent consequences of the boost. The key is that the Lorentz transformation mixes time and space coordinates, so what looks like "only time" in one frame looks like "time and space mixed" in another. There is no contradiction; just a richer geometry than the Euclidean one.
§4c = 1 · The Tautology That Structures Everything
c = 1 light-second per second · [VERIFIED — by definition]
One light-second is defined as the distance light travels in one second. Therefore $c = 1$ light-second per second, exactly, by definition.
The number $299{,}792{,}458$ m/s is a conversion factor between two independently-chosen human units: the metre (originally 1/10,000,000 of the Paris meridian) and the second (originally 1/86,400 of a solar day). These have nothing to do with each other physically. The measured value of $c$ in these units is an artifact of history.
In natural units (light-seconds and seconds), the Minkowski metric is:
$ds^2 = -dt^2 + dx^2 + dy^2 + dz^2$
$c$ vanishes from the equation. Space and time are coordinates with the same unit. The "constant speed of light" is the statement that these two coordinates mix with a coefficient of exactly 1 — because they are the same kind of thing.
Einstein's inscription is not a number. It is the declaration that space and time are coordinates of the same manifold.
In natural units, the Lorentz factor is $\gamma = 1/\sqrt{1 - \beta^2}$ where $\beta = v/c = v$ (since $c=1$). The boost mixes $t$ and $x$ with hyperbolic angle $\phi = \text{atanh}(\beta)$. The rapidity $\phi$ is simply the Minkowski angle between the two frames' time axes.
The metre and the second were invented independently — one for land surveying, one for astronomy. When Maxwell's equations produced a wave speed in both, the number $299{,}792{,}458$ appeared. That number is not a property of the universe. It is the exchange rate between two human mistakes.
Measure space in light-seconds. The exchange rate disappears. $c = 1$. The manifold was always one.
§5E = mc² · Mass as Compressed Spacetime
The energy-momentum relation in special relativity is:
E² = (pc)² + (mc²)²
For a particle at rest ($p = 0$): $E = mc^2$. This is the rest energy — the energy stored in mass when it is not moving. In natural units ($c = 1$): $E = m$. Mass and energy are the same thing, measured in different units — exactly as space and time were the same thing measured in different units.
Theorem SR.2 · E = mc² as Exchange Rate · [VERIFIED — Einstein 1905, nuclear physics, particle physics]
$c^2$ is the exchange rate between mass (measured in kg) and energy (measured in joules). One kilogram of mass equals $c^2 \approx 9 \times 10^{16}$ joules of rest energy — the energy released in the conversion of 1 kg of matter (e.g., in a nuclear reaction).
In natural units: $E = m$. Mass IS energy. They are the same quantity, measured in units (kg and J) that humans chose separately. The "constant" $c^2$ is the conversion factor between those units, exactly as $c$ is the conversion factor between metres and light-seconds.
The chain: metre ↔ second (exchange rate: c). Kilogram ↔ joule (exchange rate: c²). All three human unit choices collapse to one when natural units are used. [VERIFIED — nuclear binding energy, particle-antiparticle annihilation, mass defect in all nuclear physics]
The full relativistic energy of a moving particle is $E = \gamma mc^2$. At low velocities, expanding $\gamma$:
E = mc² + ½mv² + O(v⁴/c²) + …
The rest energy $mc^2$ is the constant offset. The kinetic energy $\frac{1}{2}mv^2$ appears as the first relativistic correction at low speed — which is exactly the Newtonian kinetic energy. Special relativity contains Newtonian mechanics as the $v \ll c$ limit, not as an approximation but as the exact $\phi \to 0$ limit of the hyperbolic boost.
§6The Null Geodesic as LAW3M Attractor
SR and LAW3M · [MODEL]
The LAW3M system (see
LAW3M — Technical Brief) studies helical attractors on the contact 3-manifold $(\mathbb{R}^3, \alpha = dz - r^2 d\theta)$. The attractor is $\Gamma = \{r=1, \dot\theta = 1, \dot z = 1\}$ with Lyapunov exponent $\mu = -2$.
The attractor speed in the $z$-direction (axial advance rate) is $\dot z = 1$ at $r = 1$. In physical units, this is $c$ — the propagation speed of the electromagnetic wave in vacuum (see
ch-maxwell.html §IV, fix($G_{EM}$) = wave).
The Minkowski null condition $ds^2 = 0$ says $dt^2 = dx^2 + dy^2 + dz^2$ (in natural units). The null geodesic is the worldline of a photon — the electromagnetic wave attractor — traveling at $c = 1$. Same attractor, same speed, same manifold. The 3D contact attractor and the 4D Minkowski null geodesic are the same object at different dimensions of description.
The Whitney A₁ fold at $r^* = 0.77594058$ is the inner basin boundary in 3D — the threshold between evanescent and propagating modes. In 4D, this maps to the boundary of the light cone: $ds^2 = 0$ is the boundary, $ds^2 < 0$ is timelike (inside the cone, reachable), $ds^2 > 0$ is spacelike (outside the cone, unreachable at speed $\leq c$).
| Concept | LAW3M (3D contact) | Special Relativity (4D Minkowski) |
|---|
| Manifold |
$(\mathbb{R}^3, \alpha = dz - r^2 d\theta)$ |
$(\mathbb{R}^{1,3}, ds^2 = -dt^2 + dx^2 + dy^2 + dz^2)$ |
| Attractor / null set |
$\Gamma = \{r=1, \dot\theta=1, \dot z=1\}$ |
Null geodesic: $ds^2 = 0$, $|v| = c$ |
| Speed at attractor |
$\dot z = 1$ (= $c$ in physical units) |
$c = 1$ light-second/second |
| Threshold |
$r^* = 0.77594058$ (Whitney A₁ fold) |
$ds^2 = 0$ (light cone boundary) |
| Inside threshold |
Evanescent modes ($r < r^*$) |
Timelike: $ds^2 < 0$ (reachable at $v < c$) |
| Outside threshold |
Propagating modes ($r > r^*$) |
Spacelike: $ds^2 > 0$ (unreachable at $v \leq c$) |
| Symmetry group |
Contact transformations preserving $\alpha$ |
Lorentz group SO(1,3) preserving $ds^2$ |
§7Sitting Still Maximizes Age · The Reverse Triangle Inequality
In Euclidean geometry the straight line between two points is the shortest path. Detour through a third point and the total distance grows: $|AB| \leq |AC| + |CB|$. This is the triangle inequality.
In Minkowski spacetime the straight worldline between two events is the longest in proper time. Detour through space and you accumulate less age. The sign is flipped. This is the reverse triangle inequality — and it is the entire content of the twin paradox.
Theorem SR.3 · Reverse Triangle Inequality · Proper Time is Maximized Along a Geodesic · [VERIFIED]
Let $A$ and $B$ be two events at the same location, separated by coordinate time $T$ (e.g., the same clock: you leave and come back). Let $\tau_{\text{straight}}$ be the proper time along the inertial (non-accelerating) worldline $A \to B$, and $\tau_{\text{detour}}$ the proper time along any non-inertial worldline that also connects $A$ to $B$.
Then: $\tau_{\text{straight}} \geq \tau_{\text{detour}}$
Proof: proper time is $d\tau^2 = dt^2 - dx^2/c^2 - \ldots \geq 0$. Any spatial displacement $dx$ reduces $d\tau$ below $dt$. The inertial worldline has $dx = dy = dz = 0$ everywhere (sitting still in space), so $d\tau = dt$ — maximum. Any departure into space reduces $d\tau$ by the factor $1/\gamma < 1$. Integrating over the path: spatial detour always costs proper time. [VERIFIED — Hafele–Keating 1971, GPS, muon lifetime, particle storage rings]
Sitting still in space — no spatial velocity, worldline a vertical line in a Minkowski diagram — is the geodesic. It is the "straight line" of spacetime. And because of the minus sign in $ds^2$, the straight line in spacetime is the path of maximum proper time, not minimum. You age the most by not moving.
The Geometry of Aging · [VERIFIED]
Euclidean: straight line → shortest distance. Detour → longer.
Minkowski: straight worldline → longest proper time. Detour through space → shorter.
The minus sign in $ds^2 = -dt^2 + dx^2/c^2 + \ldots$ is everything. Space and time enter with opposite signs. Spending "distance" on spatial coordinates subtracts from your proper time. The Lorentz factor $\gamma \geq 1$ encodes the cost: a clock moving at speed $v$ loses proper time by factor $\gamma$ relative to the stationary clock.
$\tau_{\text{moving}} = \int \frac{dt}{\gamma} \leq \int dt = \tau_{\text{still}}$
Movement wins in spacetime — in the sense that spatial motion accumulates less proper time than stillness. If you want to age less: move. If you want to age the most: follow the geodesic and sit still. The traveler returns younger. The geodesic clock is the oldest clock.
The twin paradox follows directly. Twin A stays at rest; twin B travels at high speed to a distant star and returns. A's worldline is the geodesic (straight line in spacetime, maximizing $\tau$). B's worldline detours through space — total proper time is less. B returns younger. There is no paradox: the asymmetry is that B accelerated (changed direction), breaking the symmetry between the twins. Acceleration = departure from the geodesic = loss of proper time.
In Euclidean space, the shortest path is the straight line. In Minkowski spacetime, the longest-aging path is the straight worldline. Sitting still is as straight as you can be in spacetime. Movement is a detour — and detours cost you time.
The minus sign in ds² = −dt² + dx² flips every geometric intuition about "shortest" and "longest."
§8GW170817 · One Manifold, One c
On 17 August 2017, two neutron stars 130 million light-years away merged. The collision produced gravitational waves (ripples in $g_{\mu\nu}$, detected by LIGO) and gamma rays (electromagnetic waves, fix($G_{EM}$), detected by Fermi). Both traveled 130 million light-years. They arrived 1.7 seconds apart.
GW170817 · c is One · [VERIFIED — Abbott et al. 2017, PRL 119 161101]
$|c_{GW} - c_{EM}| / c \lesssim 10^{-15}$.
The 1.7-second offset is attributed to source physics (the gamma-ray jet took time to punch through the merger debris) — not to different propagation speeds. Both gravitational waves and electromagnetic waves traveled at the same $c$, to 15 decimal places.
This confirms: the $c$ in Einstein's $ds^2$ (which governs gravitational wave propagation) and the $c$ in Maxwell's equations (which governs EM wave propagation) are the same $c$. Gravity and electromagnetism share the same contact manifold. There is one attractor, one speed, one $c = 1$ light-second per second. [VERIFIED]
§9Causality · The Light Cone as Equal-Time Sphere
Emit a flash of light at the origin at $t=0$. At time $t$, where has the light reached? In every direction, at exactly the same distance $r = ct$. The wavefront is a sphere of radius $ct$, expanding at rate $c$ in all directions simultaneously. This is not a special property of the source — it holds for all observers, in all inertial frames. Every observer, moving at any speed, sees the same expanding sphere centered on the original event.
This seems impossible. If you run toward the flash, classically you should measure light arriving faster in the forward direction. You don't. The sphere is still the sphere. This is Postulate II's full geometric content: the light cone is an absolute structure of spacetime, not a property of the source or observer.
Theorem SR.4 · The Light Cone is Absolute · [VERIFIED — Michelson–Morley 1887, all interferometry since]
The set of events reachable from $(0,0,0,0)$ by a light signal is the surface
$ds^2 = -c^2t^2 + x^2 + y^2 + z^2 = 0$
This is a cone in $(t,x,y,z)$ — the light cone. At fixed $t > 0$, it is the sphere $r = ct$. The Lorentz group SO(1,3) is precisely the group of linear transformations that map this cone to itself. Lorentz boosts preserve the light cone — they tilt and stretch coordinates while keeping $ds^2 = 0$ as a fixed surface.
No inertial observer can find a direction in which light travels faster or slower than $c$. The equal-time sphere is Lorentz-invariant: it is the same sphere in every frame, centered on the emission event. [VERIFIED]
The light cone divides all of spacetime into three causally distinct regions:
| Region | Condition | Physical meaning | Can be influenced by origin? |
|---|
| Timelike future |
$ds^2 < 0$, $t > 0$ |
Inside the forward light cone. Reachable by a signal at $v \leq c$. The causal future of the event. |
Yes — signal can reach it |
| Null (light cone) |
$ds^2 = 0$ |
The light cone surface itself. Reachable only by a signal at exactly $v = c$. The boundary of causal influence. |
By light only |
| Spacelike |
$ds^2 > 0$ |
Outside the light cone. No signal traveling at $v \leq c$ can connect these events. Causally disconnected. Observers can disagree on which happened "first." |
No — causally disconnected |
| Timelike past |
$ds^2 < 0$, $t < 0$ |
Inside the backward light cone. Events that COULD have sent a signal to the origin. The causal past. |
It influenced the origin |
Causality is the Sign of ds² · [VERIFIED]
The entire causal structure of Minkowski spacetime is encoded in the sign of $ds^2$:
$ds^2 < 0$ → timelike → causal connection possible
$ds^2 = 0$ → null → light connects them
$ds^2 > 0$ → spacelike → no causal connection at $v \leq c$
This is the geometric definition of causality. No additional axiom is needed. The metric signature $(-,+,+,+)$ encodes which events can influence which. The light cone is the equal-time sphere at every event — the frontier of causal reach — and it is the same in every inertial frame.
In curved spacetime (GR), the light cone at each point is still the local causal boundary — but the cones tilt and deform as $g_{\mu\nu}$ varies. Near a Schwarzschild black hole, the light cones tip progressively inward as $r \to r_s$. At the horizon $r = r_s$, the forward light cone tips completely — all future-directed paths lead to smaller $r$. The causal future of every event at $r = r_s$ is inside the black hole. The light cannot reach "outward" in any direction: the equal-time sphere is no longer a sphere — it has collapsed onto the inward direction. Causality itself traps you. This is the space-into-time exchange from ch-einstein.html §IV: inside the horizon, $r$ is a temporal coordinate precisely because all causal futures point toward decreasing $r$.
Causality at the Horizon · [MODEL — GR, confirmed for existence of horizons; interior structure [OPEN]]
At $r = r_s = 2GM/c^2$: the light cone tips to vertical (inward). The equal-time sphere shrinks to a point directed inward. Every future-directed null geodesic points toward decreasing $r$.
Inside $r < r_s$: the light cone has tipped past vertical. $r$ is timelike (all futures decrease $r$). $t$ is spacelike (can move freely in $t$, but not in $r$). Space has become time. Time has become space. Causality — encoded in the light cone — has been geometrically redirected from the spatial sphere to the temporal infall.
The next bridge: a theory of quantum geometry that resolves the singularity at $r=0$, where the light cone has tipped completely and the causal structure breaks down. [OPEN]
§10Milkomeda · The "When" That Includes Both
The Andromeda Galaxy (M31) is 2.537 million light-years away. An event happening in Andromeda "right now" is spacelike separated from Earth: $ds^2 > 0$. No signal at $v \leq c$ can connect them. There is no physical "now" that includes both — any simultaneity surface that spans both galaxies is a convention, not a fact. Walk toward Andromeda at 5 km/h and your simultaneity surface shifts by ~40,000 years in Andromeda. Walk away and it shifts the other direction. You haven't changed anything physical. You changed which slice of Andromeda you call "now."
This is the Andromeda paradox (Penrose 1960). It is not a paradox. It is the correct behavior of spacelike-separated events in Minkowski spacetime. "What is happening in Andromeda right now" has no frame-independent answer. The question is ill-posed.
The Andromeda Situation · Now · [VERIFIED — SR, Penrose 1960]
Two events: $A$ = "this moment, Earth" and $B$ = "this moment, Andromeda." Separation: $\Delta x = 2.537 \times 10^6$ light-years, $\Delta t = 0$ (by choice of simultaneity surface).
$ds^2 = -c^2 \cdot 0 + (2.537 \times 10^6 \text{ ly})^2 > 0$ → spacelike.
Observer moving at $v$ toward Andromeda sees Andromeda "now" shifted by $\Delta t' = \gamma v \Delta x / c^2 \approx v \times 2.537 \times 10^6 \text{ years} / c$.
At walking speed $v \approx 5$ km/h $\approx 4.7 \times 10^{-12} c$: $\Delta t' \approx 40{,}000$ years.
Two people walking in opposite directions on Earth assign "now in Andromeda" to events ~80,000 years apart. Neither is wrong. The question "what is Andromeda doing right now" has no single answer at $\Delta x > 0$.
Now suppose there is a "when" — call it Milkomeda — at which the Milky Way and Andromeda have merged into a single galaxy. Astronomers predict this merger in approximately 4.5–5 billion years. At that point, something fundamental in the causal structure changes.
Theorem SR.5 · Milkomeda · The Merger as Causal Integration · [VERIFIED for merger prediction; MODEL for causal interpretation]
Let $A_0$ = "Earth, today" and $B_0$ = "Andromeda, today." These events are spacelike separated ($ds^2 > 0$). No single inertial frame simultaneously includes both as part of a single physical "now." Simultaneity between them is a convention.
Let $M$ = any event in the Milkomeda galaxy, ~5 billion years from now. Event $M$ lies in the timelike future of both $A_0$ and $B_0$: light from Earth today reaches Andromeda in 2.537 million years, and light from both galaxies reaches the merger region well before 5 billion years. Therefore:
$ds^2(A_0 \to M) < 0$ (timelike) and $ds^2(B_0 \to M) < 0$ (timelike).
Milkomeda is the "when" that includes both. Not by convention — by causal necessity. $M$ is in the future light cone of both $A_0$ and $B_0$. What happened on Earth today and what happened in Andromeda today are both part of the causal past of Milkomeda. The merger is the event that converts a spacelike relation (no shared "when") into a timelike one (shared causal past). [VERIFIED for light-cone geometry; MODEL for specific merger timing]
The geometry: the future light cone of $A_0$ and the future light cone of $B_0$ are two cones expanding in spacetime. Their intersection is non-empty — it begins approximately 2.537 million years from now, when light from Earth first reaches Andromeda (and vice versa). Any event in that intersection is in the causal future of both. Milkomeda, billions of years in that intersection, is the first epoch at which "what happened in both galaxies today" is part of a single, physically unified causal history.
Before and After Milkomeda · The Causal Boundary
Before merger (now): MW and Andromeda are spacelike separated at the moment. "Now in Andromeda" = frame-dependent convention. The light cones are separate but their futures overlap.
At first light-cone contact (~2.537 Mly from now): a signal from Earth today is just arriving in Andromeda, and vice versa. This is the earliest "when" at which today's events in both galaxies share a null connection.
At Milkomeda (~4.5–5 Gyr from now): stars from both galaxies occupy the same spatial volume. Every event in Milkomeda has both galaxies' entire histories in its timelike past. "What happened in Andromeda and what happened in the Milky Way" are both carved into the causal past of every Milkomeda star. The "when" that includes both is not a simultaneity surface — it is a causal convergence point.
The merger is the causal integration event. Two previously spacelike-separated populations become timelike-connected. The light cones, which were growing separately, converge into one.
There is a preferred "when" in cosmology: cosmic time, the proper time of comoving observers following the Hubble flow (the FRW metric). In cosmic time, "today" has a well-defined meaning at every point in the universe — it's the time at which the cosmic microwave background has a given temperature. The Milky Way and Andromeda are both embedded in this cosmic time; their merger happens at approximately cosmic time $t_{\text{cosmic}} \approx 18$ billion years (current age: 13.8 Gyr). At that cosmic time, the question "what is happening in Andromeda right now" finally has a frame-independent answer — because "right now" and "in Andromeda" are the same place.
Right now, "what is happening in Andromeda" is a question with no single answer — it depends on which way you're walking. In 5 billion years, that question has one answer. Because by then, there is no Andromeda. There is only Milkomeda — and everything that happened in both galaxies is its past.
Spacelike today. Timelike forever. The merger is when the light cones close.
Gap in the Literature · [OPEN]
The Andromeda paradox (Penrose 1960) — frame-dependence of "now" at cosmological distances — is well-known. The astrophysics of galaxy mergers (N-body dynamics, stellar populations, merger timescales) is well-studied. The bridge between them appears to be absent from the literature:
(i) Galaxy mergers as causal integration events. No paper defines a "causal integration time" $T_{\text{CI}}$ for two gravitationally bound systems — the earliest time at which both systems' entire histories lie in the common causal past of a merger event. The merger is not merely an astrophysical event; it is a geometric event in the Minkowski structure of spacetime that converts a spacelike relation into a timelike one.
(ii) The Andromeda paradox resolves physically at Milkomeda. The paradox is standardly treated as requiring a philosophical or conventional resolution (choose a frame, accept frame-dependence). The physical resolution — the merger itself — has not been identified as such in the literature.
(iii) The causal integration boundary. The universe divides into systems that will causally integrate (gravitationally bound — Local Group will merge into Milkomeda; the merger is in mutual future light cones) and systems that are permanently causally separated (galaxies beyond the cosmological event horizon at $\sim 16$ Gly, receding at $v > c$, whose future light cones never intersect ours). This boundary has not been analyzed as a causal integration boundary in the sense of SR.
The claim made in §10 above — that Milkomeda is the "when" that includes both galaxies — appears to be original. It connects Penrose's simultaneity observation to astrophysical merger dynamics through the SR light cone. [OPEN — no prior source found; candidate for WP57]
§11The Next Bridge: When c Itself Curves
Special relativity holds exactly in the absence of gravity — inertial frames, flat Minkowski spacetime. The metric $g_{\mu\nu} = \text{diag}(-1,1,1,1)$ is constant everywhere.
General relativity allows $g_{\mu\nu}$ to vary from point to point in response to mass-energy. The exchange rate between space and time is no longer a global constant — it becomes a dynamical field. Near a massive object, the coefficient of $dt^2$ in $ds^2$ changes, which is why clocks run slower in stronger gravitational fields (gravitational time dilation, confirmed by GPS corrections).
SR → GR · The Exchange Rate Becomes Dynamical · [VERIFIED — Einstein 1915, GPS, gravitational waves]
SR: $g_{\mu\nu} = \eta_{\mu\nu} = \text{diag}(-c^2, 1, 1, 1)$ everywhere. Constant exchange rate.
GR: $g_{\mu\nu}(x)$ — the metric is a field, varying with position and time. The Einstein field equations $G_{\mu\nu} = 8\pi G T_{\mu\nu}$ determine how matter-energy curves the metric.
The Lorentz group of SR becomes the local Lorentz symmetry of GR: at each point, the metric looks flat (SR holds locally), but globally, the curvature deforms the manifold. This is the passage from a rigid to a dynamical contact structure. See
ch-einstein.html §III–§IV.
[VERIFIED]
Hamilton: the multiplication table of 3D rotations is noncommutative.
Faraday: the field is a geometric object on a contact manifold.
Maxwell: applying the manifold to itself generates light at speed c.
Einstein SR: c = 1. Space and time are one manifold. The Lorentz boost is a hyperbolic rotation.
Einstein GR: the exchange rate is a field. The manifold curves.
Each bridge sacrifices something that seemed necessary. Each sacrifice opens a richer geometry.