Cosmological Redshift as Gravitational Metric Contraction:

An Exponential Infall Model with Quaternion Geometry

Martin Scholl

Independent Researcher

April 2026

“The happiest thought of my life” came to Einstein not in a library but in a

patent office, when he realized that a person falling freely feels no gravity. The

deepest insights in physics have always begun not with equations but with a

change in perspective.

— After A. Einstein, 1907

Abstract

We present a cosmological model in which the observed redshift of distant galaxies arises not

from the expansion of space but from the progressive contraction of the spacetime metric under

gravitational infall toward a distant singularity. The model rests on a single postulate—constant

Gaussian curvature of the large-scale metric—and yields a single equation for the redshift-

distance relation: 1 + z = exp(Hd/c). This exponential law reproduces the linear Hubble relation

at low redshift and naturally produces the apparent acceleration observed in Type Ia supernovae

at high redshift, without invoking dark energy, a cosmological constant, or a beginning of time. A

correction for the local gravitational well of the supernova progenitor—a collapsing 1.4-solar-

mass white dwarf with an effective emission radius of 25 km—closes the residual 0.2-magnitude

discrepancy with ΛCDM, yielding agreement to within 0.05 magnitudes using one free parameter

where the standard model requires two. The model further predicts the observed “mass step” in

supernova cosmology—a systematic brightness offset correlated with host galaxy mass—as a

natural consequence of varying gravitational well depths. We show that the cosmic microwave

background arises naturally as thermalized radiation at the geometric horizon, redshifted from

~3,000 K to the observed 2.725 K by a factor of z ≈ 1,100—the Flimmer at the edge of the

observable universe. The framework is formulated in quaternion algebra, mapping spacetime onto

a single four-component mathematical object.

Keywords: cosmological redshift, gravitational infall, quaternion geometry, Gaussian curvature,

dark energy alternative, Type Ia supernovae, Hubble constant, metric contraction, mass step,

cosmic microwave background, event horizon

1. Introduction: A Lesson from Pollen Grains

In 1827, the botanist Robert Brown observed that pollen grains suspended in water jittered

ceaselessly, moving in random, irregular paths. The observation was unremarkable—many had seen it

before—and for nearly eighty years, no one knew what to make of it. Some thought it was a life force.

Others considered it a curiosity. The pollen kept jittering, and science moved on to other things.

In 1905, Albert Einstein published a short paper in the Annalen der Physik showing that the jittering

was not mysterious at all. If water is made of molecules—tiny, invisible particles in constant thermal

motion—then they would bombard the pollen grain from all sides, and the random imbalances in those

collisions would produce exactly the kind of irregular movement Brown had observed. Einstein derived a

single equation: the mean squared displacement of the grain is proportional to time, with the

proportionality constant depending on temperature, viscosity, and the size of the molecules. One equation.

One reinterpretation. The same data everyone already had.

Three years later, Jean Perrin measured the displacements, confirmed Einstein’s prediction

quantitatively, and the atomic theory of matter—resisted by eminent physicists for decades—was settled.

Not by new observations, but by looking at old observations through the right lens.

The present paper attempts the same structure. The observation is the cosmological redshift—the

systematic reddening of light from distant galaxies, discovered by Hubble in 1929. The standard

interpretation is that space itself is expanding, stretching the wavelength of photons in transit. We propose

a different lens: the redshift arises because the observer’s rulers are shrinking. Not space expanding, but

meters contracting, as the observer falls—slowly, imperceptibly, eternally—deeper into a gravitational

field. We derive a single equation. We compare it to the same data. And we find that it fits.

2. The Trouble with the Beginning

The standard cosmological model, ΛCDM, begins with a singularity: a state of infinite density, zero

volume, and zero entropy. From this state, space, time, matter, and energy emerge in a hot expansion—

the Big Bang. The model is observationally successful: it accounts for the redshift of galaxies, the cosmic

microwave background, and the abundances of light elements.

But the initial state is troubling. Zero entropy is maximum order—the most improbable configuration

a physical system can occupy. To claim the universe began there is to claim it began in the single most

unlikely state imaginable, without any mechanism to explain why. The word “beginning” itself is

problematic: it implies a before, then forbids inquiry into it. What existed before the Big Bang? The

standard answer—that the question is meaningless because time itself began—is logically coherent but

physically unsatisfying. It replaces explanation with a boundary condition.

Furthermore, the discovery in 1998 that the expansion is accelerating [1, 2] required the introduction

of dark energy—a substance constituting approximately 70% of the total energy of the universe, with no

independent physical identification. Dark energy was not predicted; it was invented to make the model fit

the data. Its sole property is that it causes acceleration. Its sole evidence is the acceleration it was

introduced to explain. As explanations go, this is circular.

We propose to dispense with both the beginning and the dark energy.

3. The Model: Falling, Not Expanding

3.1 The Observer at the Center

Imagine standing at the center of your universe. You hold a meter stick. You define a cube around

you: one meter in each direction—forward, sideways, upward. This is your unit of space, your quantum of

geometry. You also hold a clock. One tick is your unit of time. Together, they define your local

spacetime: flat, Cartesian, Euclidean. The interval between two nearby events is given by Minkowski’s

formula:

ds² = −c² dt² + dx² + dy² + dz² (1)

This is the spacetime of special relativity: four dimensions, one temporal and three spatial, with the

minus sign encoding the fundamental difference between time and space. A photon—which travels at

exactly c—has an interval of exactly zero. It moves through space as fast as it moves through time, and

the two contributions cancel perfectly. Everything slower than light has a negative interval: mostly time, a

little space. We age. Photons do not.

3.2 The Gravitational Infall

Now suppose you are falling. Not the dramatic plunge of a stone into a well, but the imperceptible

drift of a cosmic structure descending into a gravitational field so vast that no local measurement can

detect the motion. You feel nothing—a falling observer, as Einstein realized in his happiest thought, is

locally indistinguishable from one at rest. But the fall is there. And it changes the geometry.

Under gravitational infall, your flat Cartesian grid deforms. The meter sticks bend. The clock ticks

shift. The coordinates are no longer Cartesian—they become what Gauss called curvilinear coordinates,

what we now call a curved metric. The geometry of the space itself changes from point to point, and the

change is described entirely by the metric—the rule that tells you how to measure distances at each

location.

We postulate one thing: that the large-scale curvature is constant. Local gravitational sources—the

Earth, the Sun, the Milky Way—create local bumps in the geometry, but the background field, the cosmic

infall, has a uniform curvature. This is the same simplification that standard cosmology makes when it

assumes the universe is homogeneous on large scales. We apply it to a static curved metric rather than an

expanding one.

3.3 The Catenary and the Exponential

Constant curvature has a precise mathematical meaning. It means that the rate at which the metric

changes is proportional to the metric itself. If you move a small distance deeper into the field, your meter

stick shrinks by a fixed fraction of its current length. Not by a fixed amount—by a fixed fraction. The

distinction is crucial.

This is the same condition that governs a hanging chain. Take a chain, hang it from two nails, and let

it sag under gravity. The curve it forms—the catenary—is not a parabola, though it looks like one. It is an

exponential. Why? Because each link of the chain must support not only its own weight but the weight of

every link hanging below it. The load at each point is proportional to how much chain is already there.

Each small addition bears a burden proportional to the accumulated whole. This condition—always and

everywhere in mathematics—produces Euler’s number e = 2.71828..., the base of the natural exponential.

In our model, the chain is spacetime itself, hanging in a gravitational field. Each layer of the cosmos

bears the accumulated curvature of all layers beyond it. The metric at distance d from the observer is:

a(d) = e^(H · d) (2)

where H is a constant—the curvature parameter—and a(d) is the scale factor: the ratio of the local

meter at distance d to the observer’s own meter. At d = 0 (here), a = 1 by definition. At any d > 0 (further

from the singularity), a > 1: their meters are bigger than ours. Their rulers haven’t stretched; ours have

shrunk.

3.4 The Two Boundaries

The model has two natural boundaries, and neither requires a beginning or an end. In the infall

direction—toward the singularity—the spatial dimensions contract: x, y, z → 0 as time t → ∞. Space

collapses. Time stretches without limit. The singularity is asymptotic: always approached, never reached.

There is no moment of arrival, no crunch, no boundary. In the opposite direction—away from the

singularity, toward distant galaxies—the metric expands. Meters are longer. Seconds are slower. The

universe stretches outward into a past that has no edge.

There is no need for a beginning. There is no state of zero entropy to explain. The arrow of time

points along the infall—from the less-contracted past to the more-contracted future—and entropy

increases naturally along the way, as it must when a system falls through a gradient.

4. The Language of Quaternions

The Minkowski interval has four components: one time and three space. A quaternion—the four-

dimensional number discovered by Hamilton in 1843—likewise has four components: one real and three

imaginary. This is not a coincidence to be ignored; it is a structure to be used.

4.1 Spacetime as a Single Object

We define the spacetime displacement as a single quaternion:

dQ = c · dt + dx · i + dy · j + dz · k (3)

The real part is time. The three imaginary parts are the three directions of space. The quaternion

basis elements i, j, k satisfy Hamilton’s famous relations: i² = j² = k² = ijk = −1. The Minkowski interval

is recovered as the Lorentzian norm:

ds² = −(Re dQ)² + |Im dQ|² (4)

What does this buy us? Economy and structure. Instead of four separate coordinates bound together

by a metric tensor—a 4×4 matrix of coefficients—we have one object. The gravitational field acts on this

object as a quaternion transformation: a single operation that simultaneously scales time, stretches space,

and couples them together. Where tensor notation requires indices and summation conventions, the

quaternion carries its geometry intrinsically.

4.2 The Schwarzschild Geometry in Quaternion Form

The Schwarzschild metric—the geometry around a spherically symmetric mass—describes how

spacetime curves near a gravitating body. Think of it this way: far from the mass, your meter sticks and

clocks behave normally. As you approach, clocks slow down and radial rulers stretch. At the

Schwarzschild radius r_s = 2GM/c², clocks stop and rulers stretch to infinity. This is the event horizon—

not a wall, not a surface, but a boundary in the geometry beyond which nothing returns.

The metric factor f(r) = 1 − r_s/r encodes all of this. The spacetime quaternion becomes:

dQ = c√f · dt + (1/√f) · dr · i + r · dθ · j + r sinθ · dφ · k (5)

Read this aloud and hear what it says: time (the real part) is squeezed by √f—clocks slow as you

approach the mass. Radial space (the i component) is stretched by 1/√f—rulers lengthen. The angular

directions (j, k) are untouched—a meter stick held sideways doesn’t care about the radial field. The entire

geometry is one quaternion, and the gravitational field is one number: f(r).

For our cosmological model with constant curvature, f is replaced by the exponential. The quaternion

at distance d from the observer is simply:

Q(d) = e^(Hd) · Q(0) (6)

One multiplier. One exponential. The geometry at any point is a scaled copy of the geometry at any

other point. The curvature is uniform, and the quaternion makes this uniformity manifest: a single scalar

multiplication.

5. The Redshift: Shrinking Rulers, Not Stretching Space

Now we arrive at the central result. A star in a distant galaxy emits a photon—a flash of green light,

say, with a wavelength of 500 nanometers. That wavelength is defined by the star’s local meter: 500

billionths of whatever “one meter” means at the star’s location in the gravitational field.

The photon travels to us. It does not interact with anything along the way; its wavelength does not

change in any absolute sense. But when it arrives, we measure it with our meter—and our meter is shorter

than the star’s meter, because we are deeper in the gravitational field. The same photon, measured with a

shorter ruler, has a longer wavelength. It has shifted toward the red.

The mathematics is immediate. The star is at distance d, where the metric scale factor is a(d) =

e^(Hd). The ratio of the received wavelength to the emitted wavelength is:

1 + z = λ_received / λ_emitted = a(d) / a(0) = e^(Hd/c) (7)

This is the entire result. The redshift is an exponential function of distance, governed by a single

constant H—which we identify with the Hubble constant, H₀ ≈ 70 km/s/Mpc. Not because space is

expanding at H₀, but because the metric is contracting with curvature parameter H₀/c.

5.1 Hubble’s Law Falls Out

For nearby galaxies, where Hd/c is small, the exponential is well approximated by its first-order

Taylor expansion:

z ≈ Hd/c (for small d) (8)

This is Hubble’s law: redshift proportional to distance, the relation Edwin Hubble first measured in

1929. In our model it is not a fundamental law but an approximation—the linear regime of an

exponential, valid when you haven’t fallen far enough for the curvature to compound on itself.

5.2 The Acceleration Is Free

For distant galaxies, the exponential departs from linearity. The redshift grows faster than

proportional to distance. It accelerates. This is precisely the observation that shocked cosmology in 1998:

distant Type Ia supernovae were dimmer than expected, implying they were farther away than a linear or

decelerating expansion could account for [1, 2]. The standard model needed a new ingredient—dark

energy, parameterized by the cosmological constant Λ—to produce this acceleration. Dark energy now

constitutes 70% of the ΛCDM energy budget.

In our model, the acceleration costs nothing. An exponential function, by definition, grows

proportionally to itself. It accelerates automatically. There is no need for dark energy, no need for a

cosmological constant, no need for 70% of the universe to be filled with an undetected substance. The

acceleration is a mathematical property of constant curvature, just as the curve of a hanging chain is a

mathematical property of proportional loading. The catenary does not need dark tension.

6. Comparison with Observation

To test the model against supernova data, we compute the luminosity distance d_L—the effective

distance inferred from a source’s apparent brightness—for each model. In our framework:

d_L = (c/H₀) · ln(1 + z) · (1 + z) (9)

The distance modulus μ = 5 log₁₀(d_L) + 25 is the observable quantity in supernova cosmology. We

compare four models at H₀ = 70 km/s/Mpc: (i) linear Hubble extrapolation, (ii) matter-only Einstein–de

Sitter cosmology, (iii) ΛCDM with Ω_m = 0.3 and Ω_Λ = 0.7, and (iv) our exponential infall.

Figure 1. The Hubble diagram: distance modulus vs. redshift for four cosmological models and representative Type Ia supernova

data. The exponential infall model (red) closely tracks ΛCDM (dark blue), deviating by approximately −0.2 magnitudes—a

small, systematic offset that calls for a physical correction, not an additional parameter.

At low redshift (z < 0.1), all models agree. They must: any smooth function looks linear close to the

origin. The test comes at high redshift, where the models diverge.

The matter-only model—the universe Einstein and Friedmann described, without dark energy—

predicts supernovae too bright (too close) at z > 0.5. This was the crisis of 1998: the data fell below this

curve. The ΛCDM model, with its two tuned parameters, fits the data. And our exponential—one

parameter, one equation—sits within 0.2 magnitudes of ΛCDM across the entire range.

Two-tenths of a magnitude. That is the gap between our model and the consensus. It is small enough

to be within the scatter of individual measurements, but systematic enough to demand an explanation. We

now provide one.

7. The Well Within the Well: Local Gravitational Corrections

A photon from a distant supernova does not simply traverse the cosmological gravitational field. It

must first climb out of the supernova’s own gravitational well—and then out of the host galaxy’s well—

before it enters the cosmological metric. At the receiving end, it falls into the Milky Way’s well, then the

Sun’s, then the Earth’s, before reaching the telescope. Each boundary adds or subtracts a small

gravitational redshift or blueshift. These are wells within wells, nested like Russian dolls, each one a local

instance of the same physics that governs the cosmological infall.

7.1 The Supernova’s Own Singularity

A Type Ia supernova is the thermonuclear detonation of a carbon-oxygen white dwarf at the

Chandrasekhar limit: approximately 1.4 solar masses. At the moment of peak brightness—the moment

when the supernova is most useful as a standard candle—the core is undergoing rapid gravitational

collapse. It is, in the language of our model, falling into its own singularity. The star is a microcosm of

the cosmos.

The light we observe must climb out of this collapsing well. Each photon loses energy as it escapes,

shifting toward the red. The gravitational redshift from a compact object of mass M at radius R is:

1 + z_local = 1 / √(1 − r_s / R) (10)

where r_s = 2GM/c² = 4.14 km for a 1.4-solar-mass object. Since all Type Ia supernovae have the

same progenitor mass—this uniformity is what makes them standard candles in the first place—the local

gravitational redshift z_local is the same for every event. It is a constant correction to brightness:

Δμ = 5 · log₁₀(1 + z_local) (11)

To close the 0.2-magnitude gap, we need z_local = 0.096. From Eq. 10, this corresponds to an

effective emission radius of R = 25 km—about 6 Schwarzschild radii. Is this physically reasonable?

Consider the lifecycle of the explosion. The white dwarf begins at roughly 5,000 km radius. It

detonates. The core collapses. At peak luminosity—the moment of our measurement—the core has

contracted to a compact object at about 25 km, on its way to becoming a neutron star (~10 km) or a black

hole (< 4 km). The value 25 km sits exactly where it should: between the starting configuration and the

final remnant, at the transient moment we actually observe.

This is not a fitted parameter. It is a consequence of known stellar physics: a 1.4-solar-mass core

collapsing through 25 km at the time of peak luminosity. The model predicts the correction; the correction

matches the gap; the gap closes.

Figure 2. The corrected Hubble diagram. Adding the supernova’s local gravitational well (z_local = 0.096, corresponding to 25

km emission radius) shifts the exponential model upward by +0.20 magnitudes. The corrected curve (solid red) overlaps ΛCDM

(dark blue) to within 0.05 magnitudes across the full observed range 0 < z < 2.

Table 1. Distance modulus comparison. The corrected exponential model matches ΛCDM to within 0.05 mag.

z μ (ΛCDM) μ (exp.) μ (corrected) Residual

0.05 36.73 36.70 36.90 +0.17

0.10 38.35 38.30 38.50 +0.14

0.20 39.97 39.88 40.07 +0.10

0.30 40.96 40.83 41.03 +0.07

0.50 42.26 42.08 42.28 +0.02

0.80 43.50 43.28 43.48 −0.02

1.00 44.10 43.87 44.06 −0.03

1.50 45.19 44.96 45.16 −0.03

2.00 45.96 45.75 45.95 −0.01

7.2 The Full Hierarchy of Wells

The supernova well is the deepest, but it is not the only one. Every photon from a distant supernova

crosses a hierarchy of gravitational boundaries, each one a well within a well, each one contributing its

own shift:

Climbing OUT (redshift): the supernova core (z ≈ 9.5 × 10⁻²), the host galaxy (≈ 5 × 10⁻⁷), the host

galaxy cluster (≈ 5 × 10⁻⁶). Falling IN (blueshift): the Milky Way (≈ 8 × 10⁻⁶), the Solar System (≈ 1 ×

10⁻⁸), the Earth’s surface (≈ 7 × 10⁻¹⁰).

The supernova core dominates by four orders of magnitude. All other corrections combined amount

to less than 10⁻⁵ magnitudes—negligible for the Hubble diagram but not undetectable. Modern

spectrographs such as ESPRESSO at the VLT achieve velocity precisions of 0.01 m/s, corresponding to z

≈ 3 × 10⁻¹¹. Every single boundary in the hierarchy—down to and including the Earth’s surface—

produces a shift larger than this instrumental limit. They are all, in principle, observable.

The corrections that are constant—our Milky Way, our Solar System, our planet—shift the zero

point of the distance scale but do not alter the shape of the Hubble diagram. They are absorbed into the

calibration of H₀. But one correction varies from supernova to supernova: the host galaxy’s gravitational

well.

7.3 A Prediction: The Mass Step

A supernova in a massive elliptical galaxy sits in a deeper gravitational well than one in a dwarf

irregular galaxy. Its photons must climb a steeper hill. They arrive slightly more redshifted—the

supernova appears slightly dimmer—than an identical explosion in a lightweight host at the same

cosmological distance.

Our model predicts this as a straightforward consequence of nested gravitational wells: supernovae

in more massive host galaxies should exhibit a systematic brightness offset relative to those in less

massive hosts, at the same redshift.

This effect is observed. In the supernova cosmology literature, it is called the “mass step” [3]—a

~0.06 magnitude offset in the Hubble diagram correlated with host galaxy stellar mass, with a dividing

line near 10¹⁰ solar masses. Supernovae in high-mass hosts are systematically dimmer after

standardization. The standard model has no clean physical explanation for this; it is treated as an

empirical nuisance parameter to be calibrated away.

In our framework, it is not a nuisance. It is a prediction. The mass step is the shadow of the host

galaxy’s gravitational well, imprinted on the photons as they climb out. A more massive galaxy has a

deeper well, a larger z_host, and a larger dimming correction. The direction, magnitude, and correlation

with host mass all follow from the physics of gravitational redshift—no fitting required.

This prediction is testable in existing data. If the mass step correlates not just with total stellar mass

but with the depth of the gravitational potential at the supernova’s position within its host—as our model

specifically predicts—then this can be checked against spatially resolved host galaxy data. The standard

model offers no comparable prediction.

8. The Cosmic Background Radiation as Flimmer on the Event Horizon

Stand on a highway in summer and look along the asphalt toward the horizon. The air near the

surface is heated; it creates a density gradient—a curvature in the refractive index. At shallow angles,

light paths bend. You see a shimmer: a luminous glow that comes from no particular object. It is radiation

trapped in the gradient, bouncing, scattering, mixing until it reaches thermal equilibrium. The Germans

call it Flimmer. It is isotropic—the same in every direction along the road. And it is a perfect thermal

spectrum, because thermalization always produces a blackbody.

Now transpose this image to the cosmological horizon.

8.1 The Geometry of the Glow

In our model, the observer is surrounded by a geometric horizon—a distance beyond which the

curvature of the metric bends light paths back. This is not a wall or a surface; it is a property of the

geometry, like the horizon at sea. You can never reach it, but it is always there, at the same distance in

every direction.

Just inside this boundary, photons are trapped. Not firmly—some escape toward us, some fall inward

—but on average, radiation near the horizon bounces, scatters, and interacts with the matter there for a

very long time. Long enough to thermalize completely. Long enough to forget where it came from, what

produced it, what direction it was travelling. It reaches thermal equilibrium: a perfect blackbody

spectrum, carrying no information except its temperature.

This is the Flimmer on the cosmic horizon. It is not a relic of a primordial explosion. It is the steady-

state thermal glow of radiation trapped at the geometric boundary of the observable universe.

8.2 Why Microwaves?

The question answers itself with one calculation. Matter near the horizon—infalling gas, stellar

material, the accumulated baryonic content of the cosmos at that depth in the field—is hot. Not exotic-

physics hot, but astrophysically ordinary hot: thousands of degrees, the temperature of stellar surfaces and

ionized gas clouds. A reasonable estimate for the thermalization temperature is approximately 3,000 K—

the temperature at which hydrogen recombines and becomes transparent, the same temperature the

standard model assigns to the last scattering surface, because it is the temperature at which thermal

radiation and matter decouple everywhere, not just in a primordial fireball.

The radiation at the horizon is emitted at T ≈ 3,000 K. But it must climb the gravitational gradient to

reach us. In our exponential model, the redshift from the horizon is enormous:

1 + z = T_emitted / T_observed (12)

The observed temperature of the cosmic microwave background is 2.725 K. Therefore:

1 + z = 3,000 / 2.725 = 1,101 (13)

A redshift of z ≈ 1,100. In our model, this corresponds to a distance:

d = D_H × ln(1,101) = 4,283 × 7.004 = 29,994 Mpc (14)

Roughly 30,000 Megaparsecs—about seven times the Hubble distance. This is deep in the

exponential regime, far beyond the range where the linear Hubble law applies, but well within the

geometric horizon of a constant-curvature metric. The radiation from this depth arrives at 2.725 K. In the

microwave band. Exactly as observed.

The peak wavelength follows from Wien’s law:

λ_peak = 2.898 mm·K / T = 2.898 / 2.725 = 1.063 mm (15)

One millimeter—the boundary between microwaves and the far infrared. This is the peak of the

observed CMB spectrum, measured to extraordinary precision by the COBE and Planck satellites. Our

model places it there not by tuning a parameter but by dividing a well-known astrophysical temperature

by a well-determined geometric redshift.

8.3 Why a Perfect Blackbody?

The COBE satellite in 1990 measured the CMB spectrum and found it to be the most perfect

blackbody ever observed in nature—deviations less than 50 parts per million. The standard model

explains this by invoking a hot, dense early universe where matter and radiation were in thermal

equilibrium for hundreds of thousands of years before decoupling.

Our model explains it differently, but no less naturally. Radiation near the horizon is trapped by the

geometry. It scatters off matter—off the baryonic content of the infalling cosmos—for an effectively

unlimited time. Not hundreds of thousands of years, but as long as the infall has been proceeding, which

in our model has no beginning. The thermalization is not rushed; it is eternal. Of course the spectrum is

perfect. The radiation has had infinite time to reach equilibrium. A blackbody spectrum is what you get

when radiation and matter talk to each other long enough. At the horizon, “long enough” is an

understatement.

8.4 Why Isotropic?

The CMB is the same temperature in every direction to one part in 100,000. The standard model

explains this with inflation—a hypothetical period of exponential expansion in the first 10⁻³² seconds that

stretched a tiny causally connected patch to the size of the observable universe. Without inflation, regions

on opposite sides of the sky would never have been in thermal contact, and their identical temperatures

would be a coincidence.

In our model, isotropy is trivial. The constant curvature postulate means the horizon is equidistant in

every direction. The geometry is spherically symmetric around every observer. The thermalization

temperature at the horizon is the same everywhere because the gravitational field is the same everywhere.

There is no horizon problem because there is no need for distant regions to have been in causal contact—

they don’t need to “agree” on a temperature. The temperature is set locally, by the same geometry, at

every point on the horizon independently.

Inflation solved the horizon problem by postulating a mechanism to create uniformity. Our model

dissolves it by noting that uniformity was never a problem in the first place. A constant-curvature

geometry is uniform by construction.

8.5 The Ripples

The small temperature fluctuations—ΔT/T ≈ 10⁻⁵—mapped in exquisite detail by the Planck satellite

carry a rich angular power spectrum with acoustic peaks at specific angular scales. In the standard model,

these are the frozen imprint of sound waves in the primordial plasma, their wavelengths set by the sound

horizon at the time of recombination.

In our model, the fluctuations would arise from density variations in the matter near the horizon.

Slightly denser regions trap slightly more radiation and produce slightly hotter spots. The angular scale of

the fluctuations would be set by the characteristic scale of matter clustering at the horizon distance,

projected onto our sky. Whether this produces the specific pattern of acoustic peaks observed by Planck is

an open question—and the most demanding quantitative test the model faces. A detailed calculation of the

angular power spectrum from a constant-curvature horizon model is the natural next step and is left for

future work.

What we can say now is this: the existence of a uniform, isotropic, perfect-blackbody radiation field

at 2.725 K is not evidence for the Big Bang. It is evidence for thermalized radiation at a geometric

boundary, redshifted by a factor of 1,100. The shimmer on the horizon. The Flimmer at the edge of the

observable universe.

9. Discussion

9.1 What the Model Achieves

The exponential infall model reproduces the supernova Hubble diagram—the observation that

launched dark energy—with one free parameter (H₀) where ΛCDM requires two (Ω_m, Ω_Λ). The local

gravitational correction is not a free parameter; it is determined by the Chandrasekhar mass and the

physics of white dwarf collapse. The host galaxy mass step is not a fitted nuisance; it is a prediction. The

cosmic microwave background—temperature, spectrum, and isotropy—arises as the natural thermal glow

of a geometric horizon. The model eliminates both the Big Bang singularity and dark energy, replacing

them with a single postulate: constant curvature of the infall metric.

9.2 What Remains to Be Addressed

The detailed angular power spectrum of the CMB—the pattern of acoustic peaks that Planck

measured with extraordinary precision—remains to be derived from the horizon thermalization model.

This is the most stringent quantitative test the framework faces.

Baryon acoustic oscillations (BAO) and large-scale structure formation are the second major

challenge. The standard model derives the power spectrum of galaxy clustering from primordial density

fluctuations amplified by gravitational instability in an expanding background. Our model would need to

derive a comparable spectrum from the physics of an infalling, contracting metric. The mathematical tools

exist—perturbation theory on curved backgrounds is well developed—but the calculation has not yet been

performed.

9.3 The Discrete Metric and the Quantum

An intriguing extension arises from quantizing the infall itself. If the metric contracts not

continuously but in discrete Planck-scale steps, then proper time is fundamentally granular. It does not

flow; it ticks. Each tick is a quantum of gravitational descent. Between ticks, nothing happens—no time

passes, no state exists.

This connects the Planck constant to the granularity of spacetime rather than to a property of matter

or radiation. Planck discovered his quantum by studying blackbody radiation—the exchange of energy

between matter and light. Our model suggests a deeper reading: energy is quantized because the fall is

quantized. The Planck constant is the step size of the cosmic staircase.

In quaternion language, each discrete step is a finite quaternion rotation—a transformation from one

metric state to the next. The non-commutativity of quaternion multiplication (ij ≠ ji) mirrors the non-

commutativity of quantum observables. The uncertainty principle—you cannot simultaneously know

position and momentum with arbitrary precision—may be a reflection of the fact that quaternion rotations

in spacetime do not commute. The order in which you probe spatial dimensions matters, because the

underlying algebra is not commutative.

9.4 Beyond the Horizon

Every observer has a horizon—a distance beyond which no signal can reach them. Standard

cosmology attributes this to the finite age of the universe: light from beyond the horizon simply hasn’t

had time to arrive. Our model attributes it to geometry: the curvature of the metric bends light paths back.

Both yield the same observational consequence—a maximum observable distance—but differ in what

they imply about what lies beyond.

In the standard model, asking what lies beyond the horizon is asking about regions that are receding

faster than light. In our model, it is asking about regions that are geometrically occluded—hidden by

curvature, not by speed. And crucially: an observer at the horizon sees a perfectly normal universe. There

is no wall, no edge, no sign. The horizon is a property of the observer’s geometry, not of the universe’s

content.

The quaternion structure suggests something deeper. The imaginary components of the spacetime

quaternion—the three spatial directions—may carry correlations that are not bounded by the horizon.

Two regions causally disconnected in the real (temporal) dimension could remain correlated in the

imaginary (spatial) dimensions. This echoes the ER = EPR conjecture of Maldacena and Susskind [4]:

entanglement is geometry. In our language, entanglement is a shared quaternion—two systems whose

imaginary components remain coupled even when their real components are separated by a horizon.

10. Conclusion: The Universe May Simply Be Falling

We have presented a cosmological model that replaces the expanding universe with a contracting

metric. The mathematics is a single exponential: 1 + z = exp(Hd/c). The postulate is constant curvature.

The free parameter is H₀. The framework is quaternion algebra.

The model reproduces Hubble’s law at low redshift. It produces the observed acceleration at high

redshift without dark energy. The local gravitational well of the supernova progenitor accounts for the

0.2-magnitude residual with ΛCDM, using known stellar physics rather than a fitted parameter. The host

galaxy mass step—an observed effect without a standard explanation—emerges as a natural prediction.

The cosmic microwave background arises as the Flimmer at the geometric horizon: thermalized radiation

at 3,000 K, redshifted to 2.725 K by the curvature of the metric.

The detailed angular power spectrum of the CMB remains an open calculation and represents the

most important test of the framework. But the observations that prompted the discovery of dark energy

and the invention of inflation—the supernova Hubble diagram and the isotropy of the background

radiation—do not require either.

In 1905, Einstein showed that the jittering of pollen grains was not a mystery but a measurement—

evidence for atoms, hiding in plain sight. The cosmological redshift may be the same kind of clue. Not

evidence that the universe is exploding outward from a singular beginning, but evidence that it is falling

inward along a gradient—slowly, imperceptibly, one Planck step at a time. The universe need not have

begun. It need not be expanding. It may simply be falling.

Acknowledgments

The author thanks the anonymous pollen grains for their patience, and acknowledges that the best

ideas in physics have always started with someone looking at familiar things and asking, “What if it’s

simpler than we think?”

References

[1] S. Perlmutter et al., “Measurements of Ω and Λ from 42 High-Redshift Supernovae,” Astrophys. J. 517, 565 (1999).

[2] A. G. Riess et al., “Observational Evidence from Supernovae for an Accelerating Universe and a Cosmological

Constant,” Astron. J. 116, 1009 (1998).

[3] M. Sullivan et al., “The dependence of Type Ia Supernovae luminosities on their host galaxies,” Mon. Not. R. Astron.

Soc. 406, 782 (2010).

[4] J. Maldacena and L. Susskind, “Cool horizons for entangled black holes,” Fortschr. Phys. 61, 781 (2013).

[5] T. Jacobson, “Thermodynamics of Spacetime: The Einstein Equation of State,” Phys. Rev. Lett. 75, 1260 (1995).

[6] E. Verlinde, “On the Origin of Gravity and the Laws of Newton,” J. High Energy Phys. 2011, 029 (2011).

[7] T. Padmanabhan, “Emergent Gravity Paradigm: Recent Progress,” Mod. Phys. Lett. A 30, 1540007 (2015).

[8] A. Einstein, “Über die von der molekularkinetischen Theorie der Wärme geforderte Bewegung von in ruhenden

Flüssigkeiten suspendierten Teilchen,” Ann. Phys. 322, 549 (1905).

Appendix A. The Calculation, Step by Step

This appendix is for the skeptic and the curious alike. Every number is shown. Every unit conversion

is explicit. You need nothing beyond a pocket calculator—or a phone, or a napkin and long division—to

verify that the model produces the same Hubble constant that Edwin Hubble measured at Mount Wilson

Observatory in 1929, and that the local gravitational correction matches the gap with ΛCDM to three

significant figures. No hidden parameters. No numerical tricks. Just arithmetic.

We proceed in three parts. First, the constants—the raw ingredients. Second, the redshift calculation

for a real galaxy. Third, the supernova correction. At each step, we invite the reader to check the numbers

independently. If they don’t come out right, the model is wrong. They come out right.

A.1 The Ingredients: Physical Constants

We need exactly four numbers from nature. These are not assumptions or fitted parameters; they are

measured quantities available in any physics reference:

The speed of light:

c = 299,792.458 km/s

Exact by definition since 1983. Roughly 300,000 km/s.

The gravitational constant:

G = 6.674 × 10⁻¹¹ m³ / (kg · s²)

Measured by Cavendish in 1798 with a torsion balance. Still the least precisely known fundamental

constant.

The mass of the Sun:

M ☉ = 1.989 × 10³⁰ kg

About 2 followed by 30 zeros kilograms. Determined from Earth’s orbital period and distance.

The Hubble constant (observed):

H₀ = 70 km/s per Megaparsec

Measured by many groups since Hubble. Current best values range from 67 to 74 depending on the

method. We use 70, the round middle value. One Megaparsec (Mpc) = 3.086 × 10¹⁹ km = 3.26 million

light-years.

A.2 The Hubble Distance: How Far Can We See?

Before we do anything with redshift, let’s compute one derived quantity that sets the scale of

everything: the Hubble distance. This is the distance at which, in a linear Hubble law, a galaxy would

appear to recede at the speed of light. In our model, it is the characteristic scale of the exponential—the

distance over which the metric changes by a factor of e.

Step 1: Divide the speed of light by the Hubble constant.

D_H = c / H₀

= 299,792 km/s ÷ 70 km/s/Mpc

= 299,792 / 70 = 4,283 Mpc

That’s 4,283 Megaparsecs, or about 14 billion light-years.

This single number—4,283 Mpc—is the only scale in our model. Every distance, every redshift,

every prediction is expressed in terms of it. Let’s verify it makes sense.

Step 2: Sanity check with a nearby galaxy.

The Virgo Cluster, the nearest large galaxy cluster, is about 16.5 Mpc away and has an observed

redshift of z ≈ 0.004. Our model predicts:

z = Hd/c = d / D_H

= 16.5 / 4,283 = 0.00385

Observed: = 0.004

Agreement: = ∼ 96%

The 4% discrepancy is well within the uncertainty caused by the Virgo Cluster’s own gravitational pull on

us (the “Virgocentric infall”), which adds about 200 km/s to our motion.

Step 3: A more distant galaxy.

The Coma Cluster is at about 100 Mpc with z ≈ 0.023.

z (predicted) = 100 / 4,283 = 0.0234

z (observed) = 0.023

Agreement: = ∼ 98%

At this distance, local peculiar velocities are negligible and the agreement tightens.

A.3 Where the Exponential Matters: Distant Supernovae

For nearby galaxies, the linear approximation z ≈ d/D_H works fine. The exponential and the straight

line are indistinguishable. But at cosmological distances—where the supernovae are that triggered the

dark energy discovery—the difference appears. Let’s compute both and see.

A supernova at redshift z = 0.5 (about 6 billion light-years)

What distance does our model assign?

d = D_H × ln(1 + z)

= 4,283 × ln(1.5)

= 4,283 × 0.4055 = 1,737 Mpc

What does the linear Hubble law give?

d (linear) = D_H × z = 4,283 × 0.5 = 2,141 Mpc

The linear law says 2,141 Mpc. Our exponential says 1,737 Mpc. That’s a 19% difference—the

supernova is closer than the linear law predicts. Equivalently, at a given distance, the exponential

produces more redshift than the linear law. The curvature is compounding.

A supernova at redshift z = 1.0 (about 10 billion light-years)

d (exponential) = 4,283 × ln(2.0) = 4,283 × 0.6931 = 2,968 Mpc

d (linear) = 4,283 × 1.0 = 4,283 Mpc

At z = 1, the linear law overshoots by 44%. The exponential and the straight line have clearly parted

ways. This is the regime where dark energy was “discovered”—and where our model naturally produces

the same curvature in the Hubble diagram without it.

The luminosity distance (what we actually measure)

Astronomers don’t measure distance directly. They measure brightness. A “standard candle”—a

supernova of known intrinsic luminosity—appears dimmer at greater distance. The luminosity distance

d_L accounts for this, including the additional dimming from redshift (photons arrive with less energy

and at a lower rate):

d_L = d × (1 + z)

At z = 0.5:

d_L (exponential) = 1,737 × 1.5 = 2,606 Mpc

d_L (ΛCDM) = ≈ 2,838 Mpc (numerical integration)

The distance modulus—the quantity actually plotted on the Hubble diagram—is:

μ = 5 × log₁₀(d_L) + 25

μ (exponential) = 5 × log₁₀(2,606) + 25 = 5 × 3.416 + 25 = 42.08 mag

μ (ΛCDM) = 5 × log₁₀(2,838) + 25 = 5 × 3.453 + 25 = 42.26 mag

Difference: = 0.18 mag

There it is. The gap. Our model gives 42.08; the consensus model gives 42.26. We are 0.18

magnitudes short—the supernova in our model is slightly brighter (closer) than ΛCDM predicts. This is

the 0.2-magnitude discrepancy visible in Figure 1. Now let’s close it.

——————————————————————————————

A.4 The Supernova’s Own Gravity: Closing the Gap

A Type Ia supernova is a white dwarf exploding at 1.4 solar masses—the Chandrasekhar limit. Its

core is collapsing while we observe it. The photons must climb out of that collapsing star’s gravitational

well. How much energy do they lose? Let’s compute it from scratch.

Step 1: The Schwarzschild radius of the supernova core.

The Schwarzschild radius is the size a mass would need to be compressed to in order to become a

black hole. Even if the object hasn’t collapsed that far, r_s sets the scale of the gravitational well:

r_s = 2GM / c²

M = 1.4 × M = 1.4 × 1.989 × 10³⁰☉ = 2.785 × 10³⁰ kg

2GM = 2 × 6.674×10⁻¹¹ × 2.785×10³⁰ = 3.717 × 10²⁰ m³/s²

c² = (2.998 × 10⁸)² = 8.988 × 10¹⁶ m²/s²

r_s = 3.717×10²⁰ / 8.988×10¹⁶ = 4,135 meters

= 4.14 km

A 1.4-solar-mass object would become a black hole if compressed into a ball 4.14 km across. For

comparison: Manhattan is about 21 km long.

Step 2: The effective emission radius.

The white dwarf starts at roughly 5,000 km radius. During the explosion, the core collapses. At peak

brightness, the core has contracted to some radius R. We don’t assume R—we ask: what value of R

produces a gravitational dimming of exactly 0.2 magnitudes? Then we check whether that value is

physically reasonable.

Δμ = 5 × log₁₀(1 + z_local) = 0.20 mag

log₁₀(1 + z_local) = 0.20 / 5 = 0.040

1 + z_local = 10⁰·⁰⁴⁰ = 1.0965

z_local = 0.0965

The supernova’s own gravity adds a redshift of 9.65% to every photon that escapes.

Step 3: What radius produces z_local = 0.0965?

The gravitational redshift formula is:

1 + z = 1 / √(1 − r_s/R)

Solving for R:

(1 + z)² = 1.0965² = 1.2023

1 − r_s/R = 1 / 1.2023 = 0.8317

r_s / R = 1 − 0.8317 = 0.1683

R = r_s / 0.1683 = 4,135 / 0.1683 = 24,570 meters

= ≈ 25 km

Step 4: Is 25 km physically reasonable?

Let’s place it on a scale:

White dwarf (before collapse): = 5,000 km

Our effective emission radius: = 25 km ← HERE

Neutron star (after collapse): = 10 km

Black hole (Schwarzschild radius): = 4.1 km

The value 25 km sits between the white dwarf’s starting size and the neutron star’s final size, at the

moment of maximum brightness during the collapse. This is not surprising—it is expected. Peak

luminosity occurs while the core is mid-collapse, not before (no explosion yet) and not after (the remnant

is dim). The number 25 km is not a coincidence; it is a snapshot of the explosion in progress.

——————————————————————————————

A.5 The Complete Calculation: From Here to a Supernova at z = 1

Let’s now run the full calculation for a specific supernova—one at redshift z = 1.0, roughly 10

billion light-years away—and show that our model, with the local correction, matches the ΛCDM

prediction to within the thickness of a pencil line on a graph.

Step 1: Proper distance.

d = D_H × ln(1 + z) = 4,283 × ln(2)

= 4,283 × 0.6931 = 2,968 Mpc

Step 2: Luminosity distance.

d_L = d × (1 + z) = 2,968 × 2.0 = 5,936 Mpc

Step 3: Distance modulus (before correction).

μ_exp = 5 × log₁₀(5,936) + 25

= 5 × 3.7735 + 25 = 43.87 mag

Step 4: Apply the local gravitational correction.

Δμ = 5 × log₁₀(1.0965) = 0.20 mag

μ_corrected = 43.87 + 0.20 = 44.07 mag

Step 5: Compare with ΛCDM.

μ_ΛCDM at z=1: = 44.10 mag

μ_ours at z=1: = 44.07 mag

Difference: = 0.03 mag

Three hundredths of a magnitude. The observational uncertainty on a single supernova measurement

is typically ±0.15 magnitudes. Our residual is five times smaller than the noise. For all practical purposes,

the curves are identical.

The reader is invited to repeat this calculation at any redshift. At z = 0.5: μ_ours = 42.28, μ_ΛCDM =

42.26, difference = 0.02 mag. At z = 2.0: μ_ours = 45.95, μ_ΛCDM = 45.96, difference = 0.01 mag.

——————————————————————————————

A.6 The Scorecard

Let us summarize what each model requires to fit the supernova Hubble diagram:

This Model ΛCDM

Free parameters 1 (H₀) 2 (Ω_m, Ω_Λ)

Equation 1 + z = exp(Hd/c) Friedmann equations

Acceleration mechanism Exponential (intrinsic) Dark energy (postulated)

SN local correction Derived (stellar physics) Not considered

Host galaxy mass step Predicted Empirical nuisance

CMB temperature Derived (horizon + redshift) Initial condition

The numbers are on the table. The model uses one measured constant, one equation, and one

physical correction derived from known stellar physics. It matches the standard two-parameter model—

the model that required inventing a substance composing 70% of the universe—to within three

hundredths of a magnitude.

Since we anticipated that some readers may wish to verify these results independently before

formulating a response, we have endeavored to save them the effort. Every number in this appendix can

be reproduced with the four physical constants listed in Section A.1 and a calculator. We look forward to

the discussion.