Cosmology That ClicksBonus 4 (main-episode level) · making the series' foundation rigorous

The "space stretches = light slows" used by feel all along, now proved with conditions attached

When Are Spatial Curvature
and Slowing Light Equivalent? The watchword since Episode 1, "space stretches = light slows" — when does it hold, and when does it break?
Bonus 3's "what to protect is α" becomes, directly, half of the equivalence condition.

Tools you'll need: division, proportionality, that light travels distance÷time c_B · a = constant / α invariant

Since Episode 1, this series has repeated that "space stretches" and "light slows" are rephrasings of the same thing. In Bonus 3 we also saw why VSL failed by getting this rephrasing wrong. Now, head-on — under exactly what conditions are these two equivalent? And where does it break? Not by feel, but by laying out the conditions. To state the conclusion first: equivalence holds only when two conditions are met, and one of them is that very $\alpha$ invariance that VSL failed to protect in Bonus 3.

01First, write "space stretches" from the light side

In an expanding universe, the distance to a far galaxy increases with time. What represents that stretch is the scale factor $a(t)$. Set the present to $a=1$; the past had smaller $a$, the future larger. Picture the ruler of space stretching and shrinking wholesale by a factor $a(t)$.

Here we switch views. Fix space as a "non-stretching lattice (comoving coordinates)" and instead consider the speed at which light travels over that lattice. Because space has stretched by a factor $a$, seen from the fixed lattice, light appears to cover fewer "lattice cells" in the same time. In an equation —

Translating "space stretch" into the language of the speed of light

$$c_B(t) = \frac{c_0}{a(t)}\qquad(\text{the effective speed of light measured on the fixed lattice})$$

The larger $a$ (the more space has stretched), the smaller $c_B$ (the slower light). This is the equation-form of Episode 1's "space stretches = light slows." Multiply both sides by $a$ and you get the form familiar from the series.

The identity of the watchword

$$c_B(t)\cdot a(t) = c_0 = \text{constant}$$

"Speed of light × space stretch = constant." Space stretching by a factor $a$, and light dropping by a factor $1/a$, are exact flip sides in this one equation. This is the first equivalence condition.

Figure: c_B (slowing light, amber) and a (space stretch, blue) always have a constant product. Move one and the other necessarily moves as its inverse.

Space stretch a (universe grows to the right) 1.0

a = 1.00 → c_B = c₀/a = 1.00 → product c_B·a = 1.00 (constant)

space stretch a speed of light c_B = c₀/a product c_B·a (doesn't move)

02But this alone is only "apparent equivalence"

$c_B\cdot a=\text{constant}$ certainly rewrites "space stretch" as "slowing light." But — this alone might be just swapping symbols. As Bonus 2 showed, a quantity with dimensions ($c$ or length) can change its appearance freely by swapping the standard. To truly say "the same physics," you must guarantee that the rewriting changes not one observation.

What determines observations — the colors of light atoms emit, the rhythms of atomic clocks, the behavior of matter — is, as we saw in Episodes 2 and 6, the unit-free ratio, the fine-structure constant $\alpha$. So for the equivalence to be real, $\alpha$ must not move across the rewriting.

The second equivalence condition

$$\alpha = \frac{e^2}{4\pi\varepsilon_0\hbar c}\ \text{is invariant across the rewriting}$$

If you move $c_B$, you must move $e,\hbar$ in step to protect $\alpha$ — that is, "slowing light" must be an operation that rewrites all of physics as a set, units and all. Then atoms, clocks, and rulers all scale together, the changes cancel in the ratio, and observation doesn't change one iota. This is the second condition — the very one VSL failed to protect in Bonus 3.

Collecting on Bonus 3 VSL did the first condition (move $c$) but broke the second ($\alpha$ invariant) — because it moved only $c$ and held $e,\hbar$ fixed. So it became a rewriting that is "only apparently equivalent, but changes observation (conflicts with atomic clocks)." Of the two equivalence conditions, it satisfied only one — this is the rigorous statement of VSL's near-miss.

◇ ◇ ◇

03Laying out where equivalence holds and where it breaks

With both conditions met, "space stretch" and "slowing light" are perfectly equivalent — compute either way, and no observation can tell them apart. Conversely, if any condition breaks, equivalence breaks. Let's organize.

✓

Condition 1: $c_B\cdot a=\text{constant}$ (slowing light and space stretch are flip sides)Close to a definition. Write the space stretch in the language of the speed of light and it's automatically satisfied.

✓

Condition 2: $\alpha$ invariant (= rewrite all of physics, units and all)The guarantee that the rewriting is "merely a relabeling of symbols." This is what secures observational invariance.

✗

Break case A: move only $c$, moving $\alpha$ (VSL-type)Condition 2 breaks. Observation changes, conflicting with atomic clocks. Not equivalence but "different physics."

✗

Break case B: very short distances / extreme gravity (Planck scale)A region where the premise "the space stretch is uniform in all directions and smooth" becomes doubtful. Corresponds to Episode 5's "edge of the band." Equivalence is not guaranteed here.

In other words, the "space stretch = slowing light" the series has used all along is rigorously correct only in the range where Conditions 1 and 2 both hold. That range covers nearly all the physics we can touch — which is why it was safe to use as a watchword. But it was not "always, unconditionally." We can also point clearly to where it breaks (VSL-type rewriting, the Planck scale).

Why the two conditions imply "observational invariance" (the reasoning)

① The way light travels doesn't change

By Condition 1, the distance light reaches, $\int c_B\,dt$, is the same value whether written as space stretch or as slowing light (confirmed numerically in Episode 4). Causal structure, redshift, and the horizon are invariant.

② The matter side doesn't change either

By Condition 2, the $\alpha$ that sets atomic sizes, energy levels, and clock rhythms is invariant. So the behavior of matter is invariant too. By ① and ②, neither light nor matter changes — if everything observable is the same, the two views are the same physics. This is what "equivalent" means.

04So you may tell it either way — but with conditions

The conclusion. In the range where Conditions 1 and 2 both hold, "space stretches" and "light slows" are two ways of telling one and the same physics. Like the record and CD of Episode 5, there are just two media; the music playing is one. So when Episode 1 said "tell it in the version that clicks (slowing light)," that was not a whim but a legitimate choice supported by these two conditions.

And this episode, while making the series' foundation rigorous, also lit up VSL's lesson from the other side — the key to preserving equivalence is not how you move the dimensionful $c$, but whether you can protect the dimensionless $\alpha$. Bonus 3's diagnosis, "what to protect is $\alpha$," was directly the second equivalence condition. The two episodes shake hands at the same single point.

The honest line

"Equivalent" can be said only within the observable range (the region where Conditions 1 and 2 hold). Whether, at extremes like the Planck scale, space really stretches smoothly, and whether equivalence holds there, is unsolved — like Episode 5's "edge of the band," no one has yet checked it by experiment.

Also, this equivalence is not a claim that "the specific model $c\cdot t=\text{constant}$ is correct." What's equivalent is the ways of telling; which way of expanding (the form of $a(t)$) is real is a separate matter, decided by observation. Equivalence guarantees only "freedom of viewpoint," not the correctness of a model.

Practice problems (solvable with just this episode's formulas)

When space has stretched twofold ($a=2$), by $c_B\cdot a=\text{constant}$, what fraction of its original value is the speed of light $c_B$?
Show answer
$c_B=c_0/a=c_0/2$. Half the original. Space stretching twofold = light dropping to half is the flip-side relation.
Is a rewriting that "satisfies $c_B\cdot a=\text{constant}$ yet moves $\alpha$" equivalent or not? Give the reason.
Show answer
Not equivalent. It satisfies Condition 1 but breaks Condition 2 ($\alpha$ invariant), so observation (atomic clocks, etc.) changes. This is the VSL-type failure — only "apparent equivalence."
Name one place where "space stretch = slowing light" can break, from this episode and Episode 5.
Show answer
The Planck scale (very short distances / extreme gravity). A region where the premise "space stretches smoothly and uniformly" isn't guaranteed — Episode 5's "edge of the band." Equivalence isn't secured here.

Wrap-upEquivalence stood on "two conditions"

"Space stretches = light slows" is rigorously equivalent only when two conditions are met — ① $c_B\cdot a=\text{constant}$ (slowing light and space stretch are flip sides), ② $\alpha$ invariant (rewrite all of physics, units and all). ① guarantees the invariance of how light travels, ② the invariance of how matter behaves, and together they secure "all observations the same = the same physics."

It breaks for VSL-type rewritings that can't protect ②, and at the Planck scale where the premise becomes doubtful. The watchword since Episode 1 was correct only in this range — and that range covers nearly everything we touch. The key to preserving equivalence is not how you move $c$, but whether you protect $\alpha$ — Bonus 3's lesson became, directly, the equivalence condition, and closed the series' foundation.

The series continues The foundation of "spatial curvature = slowing light" is now rigorous, with conditions. The big doors that remain are the "do the three forces meet at a point (grand unification)" teased in Episode 6 Part 2, and "can gravity be put into this picture?" Condition 2 ($\alpha$ invariant), which bounds the range of equivalence, will keep being a guidepost even in those unsolved questions.

This is Bonus 4 of "Cosmology That Clicks," a reading piece for curious high-schoolers and undergraduates. $c_B=c_0/a$ and $c_B\cdot a=\text{constant}$ rewrite spatial expansion in an FLRW universe as an effective speed of light in the comoving frame, corresponding to the conformal-time representation. The equivalence (invariance of causal structure, redshift, and the horizon) holds rigorously under reparametrization of dimensionful quantities plus invariance of the dimensionless $\alpha$. The locally measured speed of light is always $c_0$. Discreteness at the Planck scale and whether equivalence holds there are unsolved problems in physics. The "equivalence" here is between ways of telling (choices of coordinates and units), not a claim about the validity of any particular expansion model. — To print, use your browser's Print → Save as PDF (in the printed version, the slider and answers are static/hidden).

Print / save as PDF: ⌘+P (Ctrl+P on Windows). On screen, the slider shows c_B·a staying constant. "Show answer" opens the solutions.

When Are Spatial Curvatureand Slowing Light Equivalent? The watchword since Episode 1, "space stretches = light slows" — when does it hold, and when does it break?Bonus 3's "what to protect is α" becomes, directly, half of the equivalence condition.

01First, write "space stretches" from the light side

02But this alone is only "apparent equivalence"

03Laying out where equivalence holds and where it breaks

04So you may tell it either way — but with conditions

Wrap-upEquivalence stood on "two conditions"

When Are Spatial Curvature
and Slowing Light Equivalent? The watchword since Episode 1, "space stretches = light slows" — when does it hold, and when does it break?
Bonus 3's "what to protect is α" becomes, directly, half of the equivalence condition.