The Winding Down
Years until the next pop — how a bubble runs out of afterglow, empties its sky, and hands its islands to the hourglass
The boil-timeline plate at the head of A Universe That Boils reads left to right from the pop: the fog, the lock, recombination, the dark ages, the first galaxies, the moraine wash, and at the right-hand edge the Rubin deep field we live in. Every chapter of this section so far has been a chapter about that plate — about the wind-up. Why Matter Won and The Two Ledgers are about the fog; The Quiet Majority and The Forge about what came out of it; Early Structure and the Erratics about the structure and the debris. The expansion has been the whole story, and in this framework the expansion is a bubble growing into the metastable phase that hosted it.
A bubble cannot grow forever against nothing. This chapter turns the plate around and asks what the expansion does to \mathcal{B}^0 once it has nothing left to push against — and how, out of that, the next pop is prepared. It is not a heat-death chapter. The framework does not have a heat death; it has a cycle, and the second half of the cycle has so far been asserted in a sentence (“the kettle is always simmering”) rather than walked through. Walking through it turns out to be cheap, because every rung is already on the paper’s books. What is new is the assembly, one identity, one correction to an open-problem entry, and a range.
The picture that emerges has one clock and one thermometer. The clock is the de Sitter e-fold time the residual \Lambda imposes once the moraine wash has passed. The thermometer is the cosmic microwave background, which does not drive any of it and reads all of it.
All numbers below are machine-checked in scripts/winding_down_timeline.py.
One clock
The gravity chapter’s tracker integrates the vacuum’s relaxation against Friedmann and finds the residual \rho_\Lambda “freezing toward de Sitter.” The DESI crust fit puts the last feature of the moraine wash — the harmonic edge — at z_\text{harm}=-0.25, in our future. After that there is nothing left in the expansion history but the frozen residual, and the expansion rate settles to a constant:
H_\Lambda \;=\; H_0\sqrt{\Omega_\Lambda} \;=\; 55.8\ \text{km/s/Mpc}, \qquad \tau_\Lambda \;=\; \frac{1}{H_\Lambda} \;=\; 17.5\ \text{Gyr}.
Under the crust reading the density we measure today is a crest, f(0)=1.25, and the base the wash relaxes onto is a factor 1.25 lower, which slows the clock to H_\Lambda=49.9 km/s/Mpc and \tau_\Lambda=19.6 Gyr. The twelve percent between the two is the difference between Planck’s \Lambda and the framework’s base; nothing below depends on which is used, because everything below is stated in e-folds,
N \;\equiv\; H_\Lambda\,t \;=\; \ln\frac{a(t)}{a_0},
with the year count following at 17.5–19.6 Gyr per e-fold. The wind-up was measured in 1+z; the wind-down is measured in N. They are the same logarithm read in opposite directions, and there is a symmetry in the counts worth stating at the outset: the CMB has been stretched by 7.0 e-folds since recombination and by 3.0 since it crossed the modon floor at z=18.3. What is ahead of it is about sixty more.
One thermometer
The CMB is the right instrument for reading the wind-down for the same reason it was the only instrument that ever read the dc1 lattice: it is everywhere, it interacts with nothing, and its energy tracks the scale factor exactly, E\propto1/a. Every marker in this chapter can therefore be quoted as a temperature of the afterglow:
T(N) \;=\; 2.725\ \text{K}\;e^{-N},\qquad \lambda_\text{peak}(N) \;=\; 1.87\ \text{mm}\;e^{N}.
| N | Years from now | T_\text{CMB} | \lambda_\text{peak} | in cells \xi |
|---|---|---|---|---|
| 0 | now | 2.73 K | 1.9 mm | 19 |
| 1 | 16 Gyr | 1.0 K | 5 mm | 52 |
| 3 | 51 Gyr | 0.14 K | 3.8 cm | 390 |
| 6 | 104 Gyr | 6.8 mK | 0.75 m | 7.8\times10^3 |
| 20 | 350 Gyr | 5.6 nK | 910 km | 9\times10^9 |
| 40 | 700 Gyr | 10^{-17} K | 4\times10^{14} m | 5\times10^{18} |
| 67 | 1.2 Tyr | 2\times10^{-29} K | \sim c/H_\Lambda | 10^{30} |
One correction to an intuition the plate invites, because it matters for what follows. By count, the CMB is the second-largest population the boil produced — 1.6\times10^9 photons per baryon, 1/1509 of the dc1. By energy it is nothing: \Omega_\gamma=5.4\times10^{-5} today, 1.7\times10^{-4} of the matter density, and falling as e^{-N}. The energy ledger already found that the light’s 99.9\% went into P\,dV work on the expansion, reversibly; what is left cannot drive anything. The stretch is not a factor in the transition. It is the readout of the transition. That is a stronger role, not a weaker one: the thermometer is the one thing in the bubble that never stops reporting.
The near future: the surf ends
The first marker is the one the boil plate already draws leaving the frame. The cyan trace of the moraine wash — the dispersive shock from \mathcal{B}^0’s wall meeting the remnant boundary of \mathcal{B}^{-1} — has its harmonic edge at z_\text{harm}=-0.25, which is a=4/3, N=0.29, and on the Planck background
t_\text{harm} \;\approx\; 4.4\ \text{Gyr from now}.
That is when the last ripple of the previous cycle passes us. Under the crust reading the dark-energy density we measure drops from f=1.25 to the base over that interval — a 20\% fall in \rho_\Lambda that no observer will be in a position to check, but which is the reason the crust chapter predicts w\neq-1 now: a transient cannot have w=-1. After the edge, \mathcal{B}^0 has finished registering its first contact with the outside, and its history is its own.
Two standard clocks happen to sit on the same tick, and the plate should draw them because they are what “the stars turning red” first means at home: the Sun leaves the main sequence at \sim5 Gyr, and the Milky Way and Andromeda merge at \sim4.5 Gyr. Neither is a framework result. Both are the wind-down beginning at the only place it will ever be watched from.
At N=0.4 — about 6 Gyr — comes the first framework-specific tick. The CMB cools through 1.88 K, the Hawking temperature of the lunar-mass floor hole M_\text{min}=\xi c^2/2G. From then on the lightest permitted black hole sheds more light than it banks. It is the first of the Hawking flips, and the section below says why none of them matter.
The sky empties: islands and doldrums
By N\approx6 — roughly 10^{11} years — every galaxy not gravitationally bound to our own has been redshifted past detectability.1 The sky is then one object: the merged Local Group, and beyond it the CMB at 7 mK. The universe has separated into islands, and the framework says precisely where the shoreline is.
A bound system holds its substrate where the ebbing current beats the Hubble flow, \sqrt{2GM/r}>H_\Lambda r:
r_\text{zv} \;=\; \left(\frac{2GM}{H_\Lambda^{2}}\right)^{1/3} \;=\; 1.4\ \text{Mpc}\ \Big(\frac{M}{10^{12}M_\odot}\Big)^{1/3},
the zero-velocity surface — 1.4 Mpc for a Milkomeda-class island, 14 Mpc for a Phoenix-class cluster. Inside it the lattice is held; outside it the lattice is carried off. And the dc1 held inside has a mass that is, by a one-line identity, a fixed fraction of the island:
\boxed{\;M_\text{res} \;=\; \tfrac{4}{3}\pi r_\text{zv}^{3}\,\rho_\text{DM} \;=\; \frac{8\pi G\rho_\text{DM}}{3H_\Lambda^{2}}\,M \;=\; \frac{\Omega_\text{DM}}{\Omega_\Lambda}\,M \;=\; 0.38\,M\;}
— the dark-sector ratio again, this time as the size of the reservoir a bound island keeps. A 10^{12}M_\odot island retains 4\times10^{11}M_\odot of lattice; a 10^{15}M_\odot cluster retains 4\times10^{14}M_\odot. This number is the fuel gauge for everything in the hourglass section below.
Outside the shoreline the story is the marginal point read forward. The framework’s present-day mean density sits at n_\text{tr} — that is why light has a single sharp c — and the mean has got there by clumping: above n_\text{tr} a uniform substrate is tachyonic and must break into self-bound structures at the close-packing density, and the mean falls as those structures separate. That engine switches off at n_\text{tr}. Below it the uniform medium is stable, and stable means it simply dilutes: the Hubble-flow remainder between the islands falls as e^{-3N} — a twentieth of marginal after one e-fold, 10^{-8} after six, 10^{-86} by the peak stretch. In that regime the log equation of state gives the photon mode a gap, \Delta=\sqrt{\epsilon_1\epsilon_2}, and the boil chapter’s dilute-bulk law c\propto\rho^{1/3} applies. Slow signals, heavy light, weak everything.
The boil chapter has a name for that medium. It is the doldrums — “low density, slow internal clock, weak gravity, weak everything” — the substrate between bubbles, the reason one cluster of pops cannot see the next. This chapter’s contribution is only to notice that we can watch it being made. The space between the islands is not heading toward emptiness in the abstract; it is heading toward the specific state the framework already needs the inter-bubble medium to be in. The wind-down manufactures the doldrums, and the islands are what is left standing in them.
The marginal-point section says the medium “parks at the edge of stability”; the Friedmann derivation says the dc1 mean dilutes as a^{-3}. Both are true today only because clumping has been doing the diluting while each clump holds n_\text{tr}. Below n_\text{tr} there is no instability left to clump with, so the two statements separate: the islands park, the Hubble-flow remainder dilutes. That is the reading this chapter uses. The alternative — that some self-tuning holds even the void substrate at n_\text{tr} against an exponential expansion — would need a mechanism the framework does not have, and would leave the boil chapter’s doldrums with nowhere to come from.
The Hawking flips, and why they do not matter
The hourglass section noted that every permitted black hole is presently a net absorber of background light, because even the floor hole’s Hawking temperature sits below 2.725 K. That is an epoch statement, and the epoch ends:
| Hole | T_H | Flip at N | Years from now |
|---|---|---|---|
| M_\text{min} (0.9 lunar) | 1.88 K | 0.4 | 6\times10^{9} |
| 10\,M_\odot | 6\times10^{-9} K | 20 | 3.5\times10^{11} |
| Sgr A^* | 1.5\times10^{-14} K | 33 | 5.7\times10^{11} |
| M87^* | 9\times10^{-18} K | 40 | 7.0\times10^{11} |
| Phoenix A | 6\times10^{-19} K | 43 | 7.5\times10^{11} |
By three quarters of a trillion years every black hole in the universe has flipped to a net emitter of light. In standard cosmology this is the beginning of the black-hole era: the holes evaporate over 10^{67}–10^{100} years and the universe ends in a thin gas of photons. In the framework it is bookkeeping. A hole is a compactor that drinks the lattice through the valveless ebbing current, and against that drain the Hawking channel is not a competitor:
| 10\,M_\odot hole | Power |
|---|---|
| Hawking emission | 9\times10^{-31} W |
| dc1 drain, fast clock (v_\text{reg}=v_L) | 1\times10^{16} W |
| dc1 drain, slow clock (v_\text{reg}=c) | 2\times10^{8} W |
Forty to forty-six decades. Nothing the framework permits ever evaporates while there is lattice to drink, and inside an island there always is. The Hawking flips mark the moment each hole stops banking the afterglow; they were never banking anything that mattered, and the black-holes chapter’s prediction 2 — no complete evaporation, ever — is unchanged.
The peak stretch
The intuition that the CMB “must reach a peak stretch” is right, and the framework can say where, because a winding cannot be stretched over a span longer than the causal patch that carries it. The de Sitter horizon is c/H_\Lambda=17.5 Gly, and the CMB’s peak wavelength reaches it at
N_\text{peak} \;=\; \ln\frac{c/H_\Lambda}{\lambda_\text{peak,0}} \;=\; 66.7, \qquad t \;\approx\; 1.2\times10^{12}\ \text{yr}
(1.3\times10^{12} on the crust clock). Two other statements land on the same e-fold, and the coincidence is not one:
| Reading | Condition | N |
|---|---|---|
| Span | \lambda_\text{peak}=c/H_\Lambda — the winding spans the horizon | 66.7 |
| Temperature | kT_\text{CMB}=\hbar H_\Lambda/\pi — the afterglow falls to the vacuum’s own de Sitter temperature | 68.6 |
| Mass | h\nu_\text{peak}=m_\gamma c^2\sim\hbar H_\Lambda — the photon’s energy equals the frozen photon mass | 68.5 |
The first is geometry. The second uses Volovik’s local de Sitter temperature, T=\hbar H/\pi, which the boil chapter already adopts as the wall’s thermalization bath at the pop (breadcrumb 3); here it is the same formula at the other end of the cycle — the temperature the emptied bubble has, felt by anything in the flow, and the CMB cooling down to meet it means the afterglow has become indistinguishable from the bath. The third is the framework’s own: the marginal point reads the vacuum’s frozen super-criticality as a photon mass m_\gamma c^2\sim\hbar H, and a quantum whose energy has been stretched down to its rest mass is a quantum that has stopped propagating. The afterglow and the residual \Lambda — the “fog” and the “un-relaxed remainder” of the genealogy — become one thing at the same moment three different ways.
Two head-counts sit on the same tick. A horizon volume today holds 8\times10^{87} CMB photons and 1.2\times10^{91} dc1; both dilute as e^{-3N}, so the last CMB photon leaves a horizon volume at N=67.5 and the last Hubble-flow dc1 quantum at N=69.9. The horizon-spanning photon is the last photon. After N\approx70 the space between the islands carries less than one quantum of anything per causal patch. The quiet-majority chapter’s ledger closes here: the 99.9\% the light had given up by today becomes, to any precision one likes, all of it; the P\,dV work is done; the windings have “relaxed back into the texture of the medium that made them,” as the boil chapter put it, and the medium itself has thinned to nothing where it is not held.
It is worth setting the count against its own history. The CMB spent 4 e-folds as a gas of compact modons (release at z=1090 to the floor at z=18.3), has spent 3 as a lattice winding, and has 60 ahead of it in the same state. The wind-up took 13.8 Gyr for its seven; the wind-down spends 17.5 Gyr on each of the sixty. The sentence “the CMB is not made of light any more” was true at z=18; at N=67 it stops being a distinction, because there is no light for it to be distinguished from.
The stars go out
Inside the islands the standard story of a dying universe runs its course, and the framework adopts its clocks without changing them.2 Star formation, already a tenth of its z\approx2 peak, declines exponentially; the merged island is a single elliptical by \sim10^{12} yr; the last stars — red dwarfs of a twelfth of a solar mass, burning for 10^{13} years — go out by \sim10^{14} yr. “The stars turning red” is literal twice over: the population reddens as its massive members die first, and every spectrum is redder through the thermometer’s own descent. By the time the last red dwarf dims, the CMB has cooled through more than five thousand e-folds and is not a background at all.
Then the island condenses. Over 10^{19}–10^{20} years stellar encounters relax the galaxy dynamically: most remnants are flung out past the shoreline into the doldrums, and the rest spiral into the central compactor. What is left is degenerate matter — white dwarfs cooled to black, neutron stars, planets, the dead — around a black hole.
Here the framework departs from the standard script in two places, and both departures are already on its books. The standard degenerate era ends in proton decay at 10^{32}–10^{40} yr, dissolving the remnants into leptons and light. The framework’s proton is a Borromean junction — three vortex channels that cannot be unlinked without cutting — and does not decay; a confirmed proton decay is already listed there as a falsifier. So the framework’s degenerate remnants never dissolve. They wait. And the standard black-hole era ends in evaporation; the previous section retired that. The island’s endpoint is therefore not a thin photon gas but a compactor with a retinue of cold, indestructible remnants — which is, recognizably, the population the erratics chapter needs \mathcal{B}^{-1} to have left behind: “cores dense enough to survive the wall cold.” The wind-down manufactures the erratics as surely as it manufactures the doldrums.
The hourglass on bound fuel
The hourglass established what fills a compactor: not the CMB (retired by twenty-one decades), not baryons (the Eddington valve, a finite reservoir), but the dc1 lattice itself through the ebbing current, at the Bondi rate \dot M=4\pi G^2M^2\rho_\text{dc1}/v_\text{reg}^3, a finite-time runaway with t_*\propto1/(M_0\rho). The wind-down sharpens it in two ways, one of which corrects an open-problem entry.
The Hubble-flow fuel is already spent. A hole drinking from the unbound remainder sees a density falling as e^{-3H_\Lambda t}, and the integral is finite: \int\rho\,dt=\rho_0/(3H_\Lambda) — the equivalent of 5.8 Gyr at today’s rate, total, forever. The Bondi solution on that supply runs away only if t_*<1/(3H_\Lambda); no permitted hole is within ten decades of that. For Phoenix A the unbound fuel changes its mass by a factor 1.00001. Whatever pops, pops on bound fuel — on the 0.38\,M of lattice the island keeps inside its zero-velocity surface.
The fuel density is pinned. WIP-34 lists, as its item 2, “galactic-center halo densities 10^2–10^5\times” the cosmic mean, shortening every t_* accordingly. In the framework’s own terms that shortening is not available. The lattice cannot be denser than n_\text{tr} without turning tachyonic (the marginal point); a “dark-matter halo” in this framework is the phonon-enhanced response of matter embedded in a uniform superfluid, not piled-up dc1; and the deeper engine holds each self-bound structure “near the marginal close-packing value,” not above it. The density a compactor drinks from is therefore n_\text{tr} — today’s cosmic mean to within the \mathcal{O}(1) the marginal-point placement allows — wherever it sits. The queue is ordered by M_0 alone, and the fast clock is as fast as it gets:
| Compactor | M_0 | t_*, fast clock (v_L) | t_*, slow clock (c) |
|---|---|---|---|
| Phoenix A | 10^{11}M_\odot | 5\times10^{14} yr (4\times10^4\,t_H) | 3\times10^{22} yr |
| M87^* | 6.5\times10^{9}M_\odot | 8\times10^{15} yr | 5\times10^{23} yr |
| Milkomeda’s hole | \sim1.5\times10^{8}M_\odot | 4\times10^{17} yr | 2\times10^{25} yr |
| Sgr A^* today | 4\times10^{6}M_\odot | 10^{19} yr | 10^{27} yr |
Read against the fuel gauge: t_* is when the mass formally diverges, and with a finite reservoir the compactor instead drains its island in a time \approx t_*(1-M_0/M_\text{res}) \approx t_*, ending at M_0+M_\text{res}. Whether the nucleation barrier is crossed before the reservoir is spent is the one thing this chapter cannot say — it is the same open coefficient as the barrier itself. What it can say is the floor: no island in \mathcal{B}^0 pops sooner than \sim5\times10^{14} years from now, forty thousand Hubble times, and that is the fast clock applied to the best-fed compactor in the sky. Our own island’s clock runs a thousand times slower. The seven decades between the fast and slow readings remain WIP-34’s item 1; the two decades of halo enhancement come off both ends.
At the earliest pop, N=H_\Lambda t_*\approx3\times10^4. The wind-down is not long the way the wind-up was long — seven e-folds spread over fourteen billion years. It is long the way inflation was long, sixty e-folds and more, and then five hundred times that again. By the time the first island pops, the space between the islands has been stretched by e^{30000}. Each island pops alone, into a parent that has emptied to nothing around it. What a wall expands into when the parent phase has been diluted below one quantum per horizon is a question the framework has not asked before and does not answer here; it is logged below.
The loop
Put the rungs in order and the second half of the cycle reads as one feedback loop, running on one clock:
- The residual sets the clock. After the surf ends the expansion is de Sitter at H_\Lambda; every subsequent rung is a count of e-folds.
- Expansion isolates. The islands separate past each other’s horizons by N\approx6; each holds its lattice at n_\text{tr} inside r_\text{zv} and keeps 0.38\,M of it as fuel.
- Expansion empties. The remainder between the islands dilutes below marginal into the doldrums, and the afterglow stretches to the horizon and goes to ground at N\approx67. The bubble forgets its own pop.
- Gravity condenses. Inside each island the stars burn down to indestructible remnants and the remnants fall inward; the compactor drains the bound lattice on the hourglass clock.
- The compactor saturates and pops. Its core, pressed to the density ceiling, is the pre-boil state; the barrier is crossed by the vortex-instanton tunneling the boil chapter’s breadcrumb 1 names; latent heat inflates \mathcal{B}^{+1}.
- The old island is the new moraine. \mathcal{B}^{+1}’s wall runs out through the island’s relaxed lattice and its retinue of remnants — and that is what the inhabitants of \mathcal{B}^{+1} will one day fit to their DESI and search for as erratics: the moraine is a bound island of the previous cycle, and the erratics are its black dwarfs. Then the loop returns to step 1 inside the child.
One reading follows that the boil chapter did not have. It describes the kettle boiling “in clusters” — a pop loading neighbouring compactors over their own thresholds — and this chapter has just isolated every island by e^{30000}. The two are consistent if the cluster of pops is the bound cluster: a Phoenix-class island keeps dozens of galaxies and their supermassive holes inside a 14 Mpc shoreline, all on their own runaway curves, and the first pop among them loads the rest. The cascade is intra-island. That would make “we are deep inside one cluster of bangs” a statement about the galaxy cluster \mathcal{B}^0 nucleated in, and would predict that a bubble’s siblings \mathcal{B}^0_j are few — as many as its parent island had compactors near the barrier — rather than a landscape.
Predictions and falsification
The chapter’s markers are, by construction, out of reach; what it exposes to test is the machinery that produces them.
- The proton must not decay. The degenerate remnants must survive to become the next cycle’s erratics, and the loop’s step 6 needs them. This is the same falsifier Why Matter Won already carries; this chapter raises its stakes from “the ledger” to “the cycle.” A confirmed proton decay at any lifetime breaks step 6 outright.
- The lattice density in halos must not exceed marginal. The 5\times10^{14}-year floor rests on the fuel density being pinned at n_\text{tr}. If the framework’s own halo profiles (galactic dynamics) ever require clumped dc1 at densities 10^2–10^5 above the cosmic mean — rather than the phonon-enhanced response of a uniform superfluid — the floor drops by the same factor, and the marginal-point stability table falls with it. The pinning is a load-bearing use of that table, and a discriminator between the two readings of a halo.
- Dark energy is a transient, and its transience is in the data now. The surf ends at z=-0.25; a density that will fall by 20\% cannot have w=-1 today. This is the crust chapter’s prediction, not a new one, but the wind-down depends on it in a specific way: the clock of everything above is the base H_\Lambda, not the crest. If the crust is an artifact of the candle (WIP-35), the clock speeds up by twelve percent and nothing else changes.
- No permitted black hole evaporates, ever. The drain beats Hawking emission by forty decades while any lattice is present. Black-holes prediction 2 states the remnant endpoint; this chapter adds that inside an island the endpoint is never reached because the hole never stops growing. An observed evaporation signature from a hole above the floor mass would break the valveless-drain premise.
Honest assessment
What is solid is the arithmetic, and the arithmetic is mostly standard physics in the framework’s dress. The de Sitter clock is Planck’s \Omega_\Lambda. The thermometer is E\propto1/a. The Hawking flips are Hawking’s formula against an exponentially cooling background. The peak stretch is the statement that a wavelength cannot exceed a horizon, made three ways; the two framework readings (Volovik’s local temperature, the frozen photon mass) are formulas the paper already uses elsewhere, evaluated at H_\Lambda. The bound-reservoir identity M_\text{res}=(\Omega_\text{DM}/\Omega_\Lambda)M is one line of Newtonian mechanics plus the framework’s identification of the ebbing current with the substrate’s inflow. The starved-hourglass integral is a convergent exponential. None of these can be wrong except in coefficient.
What is a reading is the assembly and three claims inside it. The fuel density is pinned at n_\text{tr}: this follows from the marginal-point table and the Khoury reading of halos, both already on the books, but it contradicts the working assumption in WIP-34’s item 2 and is here promoted from an inconsistency to a result — the framework should decide which it believes, and this chapter has said which. The island is the next cycle’s moraine and its remnants are the erratics: a natural closure of the boil chapter’s inheritance ledger, not a computation. The cascade is intra-island: a reconciliation of two statements that would otherwise conflict, offered as the simplest one.
What is inherited is every astrophysical timescale in The stars go out — the stelliferous era, the merger, the dynamical relaxation — taken from Adams & Laughlin and Krauss & Scherrer without modification.
What is open, and the chapter logs it rather than papering over it:
- The barrier (WIP-33) and the regulator (WIP-34). Between them they hold seven decades of the pop date and the yes-or-no of whether a given island pops at all before its fuel is spent. The floor (5\times10^{14} yr) survives both; the date does not exist until they are computed.
- What a wall expands into. The boil chapter’s bubble grows “into the metastable parent phase.” By the earliest pop the parent phase around the island has diluted to less than one quantum per horizon. Whether the child’s wall carries its own substrate outward, whether the doldrums at c\propto\rho^{1/3}\to0 act as a rigid boundary, or whether the island’s own bound lattice is the whole of the parent phase that matters, is not calculated. It is the same class of question as the wall’s ultimate fate, which the boil chapter already lists as open.
- The fork in the marginal-point reading (callout above): islands park, voids dilute. Stated, not derived.
Putting the section in context
The Big Bubble section has been a set of chapters about one plate — years after the pop — and this is the first about the other. Why Matter Won found what survived the boil; The Two Ledgers found that it was neutral; The Quiet Majority counted the remainder and found the CMB reading it; The Forge built the elements; Early Structure explained why the galaxies came early; and the Erratics looked for a stone from the cycle before. This chapter follows the same populations forward until they become the previous cycle for someone else: the light stretched to the horizon and gone to ground, the vacuum split into held islands and emptied doldrums, the matter condensed to indestructible remnants around a compactor drinking on the hourglass clock. The boil chapter said the kettle is always simmering. This chapter says how long the simmer is — at least forty thousand Hubble times, on the framework’s own numbers — and what the kitchen looks like when it finally boils: one island, alone, in a sea it has emptied, popping into a cycle that will find our black dwarfs in its moraine and wonder what they are.
Footnotes
Krauss, L.M. & Scherrer, R.J., “The return of a static universe and the end of cosmology,” General Relativity and Gravitation 39, 1545, 2007. The framework adopts the timescale; the mechanism — exponential redshift in a de Sitter phase — is the same in both pictures.↩︎
Adams, F.C. & Laughlin, G., “A dying universe: the long-term fate and evolution of astrophysical objects,” Reviews of Modern Physics 69, 337, 1997. The stelliferous, degenerate, and black-hole eras and their timescales are taken from this paper; where the framework departs from it — proton decay and black-hole evaporation — the departure is stated.↩︎