A Frame Is an Angle

Contents

1. Why the phrase will not go away
2. A frame is a choice of axes
3. The minus sign that makes it spacetime
4. Why my time is different from yours
5. The tilt of “now”
6. The twins, done properly
7. What nobody disagrees about
8. What this does not license
9. On “math is why”
10. Exercises

01 Why the phrase will not go away

“It depends on your frame of reference” is the most overworked sentence in popular science, and it survives because it is pointing at something real that it never quite says. People reach for it when two accounts of the same situation disagree and neither party is lying. That is a genuine phenomenon with an exact mathematical description, and the description is far more disciplined than the phrase suggests.

The whole of it is this. Some quantities are properties of the thing. Some quantities are properties of the description of the thing. Confusing the two is the single most common error in physics, and once you can tell them apart, relativity stops being paradoxical and becomes almost boring — which is the highest compliment you can pay a physical theory.

We already have the machinery. In part four a boost turned out to be a rotation, and momentum turned out to inherit its axis from velocity. Now we ask the obvious follow-up: rotation of what, relative to what?

02 A frame is a choice of axes

Forget physics for two minutes. Lay a stick on a table and draw a pair of axes next to it. Read off the components: so far across, so far up. Now rub out the axes and draw a new pair, rotated thirty degrees. Read again. Both components changed.

Nothing happened to the stick. You did not move it, heat it, or observe it harder. You changed the basis you were expressing it in. And one number refused to change: its length.

Figure 01

One stick, two frames, two sets of components, one length

• Your frame
• My frame, rotated 30°
• The invariant

That is a frame of reference. Not a personality, not a point of view in any psychological sense — a set of axes. Two cameras bolted to the same table have different frames. Nothing conscious needs to be involved.

One precision, since this series cares about them. In relativity a frame is more than a basis: it is a basis plus an origin, so changing frames is an affine transformation rather than a purely linear one. The full set of moves — rotations, boosts, translations in space and time — is the Poincaré group. Set the origins to coincide and you are left with the linear part, the Lorentz group, which is where the interesting geometry lives.

03 The minus sign that makes it spacetime

Rotations preserve x² + y². That is what makes them rotations; it is the definition, not a consequence. Ask which transformations preserve some other quadratic quantity and you get a different geometry with a different family of “rotations.”

Spacetime’s invariant differs from Euclid’s by exactly one character:

Transformations preserving the second quantity are hyperbolic rotations — boosts. They are parameterised not by an angle that wraps around but by rapidity, written φ, with tanh φ = v/c. And then the Lorentz factor everyone has heard of is nothing but a hyperbolic cosine:

Rapidity is the honest variable, and here is the demonstration. Velocities famously refuse to add: run at half light speed, then run half light speed again from there, and you do not get c. Rapidities add perfectly.

Figure 02

The angle behaves; the speed does not

φ(0.5c) = 0.5493    φ + φ = 1.0986    tanh(1.0986) = 0.8c naïve 0.5c + 0.5c = 1.0c  —  wrong, and it would break causality

• v = 0.10c1.005
• v = 0.50c1.155
• v = 0.87c2.028
• v = 0.99c7.089
• v = 0.999c22.37

• γ, the Lorentz factor (value at right)
• φ, the rapidity

04 Why my time is different from yours

Now the question you actually asked.

Draw your worldline: the path you trace through spacetime just by existing. It has a length. In this geometry that length is called proper time — it is what your wristwatch reads, the arc length of your own path, and every observer in the universe computes the same value for it.

My “your time” is something else entirely. It is the projection of your worldline onto my time axis. A component. And components depend on the angle between our axes, exactly as in Figure 01.

So when the textbook says Δt = γ Δτ, that is a statement about a shadow, not about your watch malfunctioning. Your watch is fine. It is measuring the thing that does not change. I am measuring its shadow on my wall.

And here the geometry does something genuinely strange, which is worth stating precisely because it is where the Euclidean intuition betrays you. Tilt a stick and its shadow on the wall gets shorter. Tilt a worldline and its projection onto my time axis gets longer. The inequality flips. In Minkowski geometry, for timelike intervals, the triangle inequality runs backwards:

Read that as a fact about lives rather than vectors and it is startling. Between two events in flat spacetime, of all the possible ways to get from one to the other, the one who coasts — the one who never changes frames — ages the most. Every deviation costs you time. There is no path of maximum ageing other than the straight one, and there is no minimum at all: you can make your proper time arbitrarily close to zero by zig-zagging near the speed of light. The infimum is zero and nothing attains it.

Your watch measures the length of your path. My clock measures its shadow on my wall. Neither of us is wrong, and the conversion between us is exact.

A caution on the classic light-clock derivation, since every popular account uses it. Bouncing a photon between two mirrors and applying Pythagoras does give you γ, and it is a fine motivation. But it quietly assumes that lengths perpendicular to the motion are unaffected — which it does not prove — and it delivers nothing about simultaneity or length contraction. Used alone it is the direct cause of the objection every student raises next: if each of us sees the other’s clock running slow, who is right? The answer is in the next section, and the light clock cannot reach it.

05 The tilt of “now”

Here is the piece that is usually left out, and it is the one that actually answers the question.

When I say two distant things happened “at the same time,” I am naming a slice through spacetime — the set of events I count as simultaneous. That slice is a surface perpendicular to my time axis. Tilt the time axis and the slice tilts with it.

So we do not merely disagree about durations. We disagree about which events are in the same instant. Your “now” and mine are two different cuts through the same block, and for events far enough apart, we can even disagree about the order.

Figure 03

Two events, simultaneous for you, an hour apart for me

• Your axes
• My axes, boosted at β = 0.5
• The two events

Notice the discipline in it. A and B can swap order only because they are spacelike separated — too far apart for light to travel between them, so neither could have caused the other. Any two events that could be causally connected keep their order in every frame, permanently. Relativity reshuffles what it is harmless to reshuffle and nothing else. Causation is never up for negotiation.

06 The twins, done properly

One twin flies out and comes back younger. Students object: motion is relative, so why is it not symmetric?

The sloppy answer is “because one of them accelerates.” That identifies the asymmetry correctly — and it is a real, frame-independent asymmetry, since only one twin’s accelerometer ever reads anything — but it is not the explanation, and it teaches the wrong lesson.

The explanation is section 04. Proper time is the length of a path, and the two twins take different paths between the same two events. The stay-at-home takes the straight one, which by the reverse triangle inequality is the longest. That is the entire calculation.

To see that acceleration is a marker rather than a cause, use the clock hypothesis: an ideal clock’s rate depends only on its instantaneous speed, not on its acceleration. This is a separate assumption, not a theorem, and it has been tested to extraordinary precision — muons in storage rings survive accelerations around 10¹⁸ times gravity with their lifetimes tracking speed alone. Granted it, you can make the turnaround as brief and as brutal as you like without changing the age difference. Whatever is doing the work, it is not the acceleration.

07 What nobody disagrees about

If some quantities are description, the obvious question is which ones are not. Relativity answers precisely, and the answer is the most useful thing in the theory.

Figure 04

The audit: what survives a change of frame

Invariant — every frame agrees

• The spacetime interval s²
• Proper time along a given worldline what a particular watch reads
• Rest mass
• The speed of light defined exactly since 1983: 299,792,458 m/s
• Total electric charge charge density is not
• The order of causally connectable events
• Whether a separation is timelike, spacelike, or null

Frame-dependent — a component, not a fact about the thing

• Simultaneity of separated events
• Duration between separated events
• Length, as measured rest length is invariant
• Energy and momentum conserved within a frame ≠ invariant across frames
• Velocity
• The order of spacelike-separated events
• The electric field E and B mix; the invariants are B² − E²/c² and E·B

08 What this does not license

Sooner or later someone in the room will say that relativity shows truth is relative — your truth, my truth, who is to say. It shows the exact opposite, and the confusion is a pun on the word.

The transformation law is public and exact

I can compute your description from mine with no residue and no negotiation. Disagreeing about components while agreeing completely about facts is the relationship between metres and feet, not between competing realities.

The invariants are shared absolutely

Every observer in the universe agrees on every interval, every proper time, every rest mass, every charge, and on what caused what. There is no frame in which the shooting came after the funeral.

The meta-fact is not itself relative

Relativity states which quantities are frame-dependent and which are not, and that statement holds in every frame. A theory that told you nothing absolute could not be tested at all.

And the frames are not points of view

A frame is a state of motion, specified numerically, related to every other by a known group. It is not a standpoint, a culture, or an opinion. Two thermometers have different frames. Neither has a perspective.

Where I would push back on myself

“Only invariants are real” is a slogan, not a result, and taken literally it is false — energy and momentum are frame-dependent and entirely real. The defensible version is narrower: invariants are the frame-independent facts, and looking for them is the most reliable method in physics. That is how Weyl and Wigner put it, and they were careful for a reason. “Invariant” is also meaningless until you say invariant under what group.

The best line on all of this belongs to Einstein, who thought the theory was misnamed. In a 1921 letter he remarked that invariant theory would have described the method better — then declined to change it, on the grounds that renaming an established term causes more confusion than it cures. He was right twice.

09 On “math is why”

This series runs on Max Tegmark’s claim that the physical world does not obey a mathematical structure but is one, and that what we call physical properties are positions inside it. Frames of reference are the cleanest case the idea has.

Because look at what the physics is actually telling you. There is one spacetime. Frames are charts drawn on it. “Your time” and “my time” are coordinate labels; the interval and the proper time are the geometry. On the Tegmark reading this is not an analogy about maps and territories — the structure is the territory, and the coordinates are the only thing that was ever a map.

Minkowski said it first, three years after Einstein’s paper and one year before his own death, opening his Cologne lecture in September 1908:

Henceforth space by itself, and time by itself, are doomed to fade away into mere shadows, and only a kind of union of the two will preserve an independent reality.

Hermann Minkowski, “Raum und Zeit,” 21 September 1908

Two honest caveats, because this is philosophy and it should be labelled as such. First, the position that structure is what is fundamentally real has a name — ontic structural realism — and a literature. James Ladyman named it in 1998 and developed it with Steven French; John Worrall’s earlier and more cautious 1989 proposal was epistemic structural realism, the claim that structure is what we can know, which is a different and weaker thesis. The two are routinely conflated and should not be.

Second, and more important: relativity is compatible with this reading and suggestive of it, but does not prove it. Someone who thinks objects are fundamental can accept every invariant in Figure 04 and hold that structure supervenes on the objects rather than replacing them. The physics underdetermines the metaphysics, and structural realism’s strongest evidence comes from quantum mechanics rather than from relativity anyway. Teach the geometry as established and the philosophy as an argument you find persuasive. Students can tell the difference, and they trust you more when you mark it.

10 Exercises

1. Do Figure 01 by hand Rotate (110, 170) by 30° and confirm the components come out (180.3, 92.2). Then verify the sum of squares is unchanged. Which line of your working is the reason the length survives?
2. Add three rapidities Three collinear boosts of 0.5c each. Compute the final speed two ways — by applying the velocity-addition formula twice, and by adding rapidities and taking a tanh. Then explain why the second method makes it obvious you can never reach c.
3. Break the diagram In Figure 03, measure the interval from the origin to event B with a ruler, using the unprimed scale, and then compute it properly. Explain the discrepancy, and state the general rule for when a Minkowski diagram may be read metrically.
4. Find the causal floor Construct two events whose order does reverse under some boost, and prove no signal could connect them. Then show that if a signal could connect them, no boost reverses them. What have you just proved about relativity and causality?
5. Argue the other side Take the position that coordinates are as real as invariants — that a frame-dependent energy is not a lesser kind of fact. Make the strongest case in one paragraph, then say what section 09 has to give up if you are right.