The Tier Below
If a dc1 particle is itself a paired vortex of a finer condensate, its near-perfect elasticity becomes topological protection — and δ, the free universe-decay constant, becomes the substrate ladder read downward, tying the framework’s smallest number to its largest.
Speculative! This chapter rests on a tier the rest of the framework does not require — a condensate beneath dc1. It earns one number and is honest about the one coefficient it still owes.
The framework builds everything on the dc1 substrate: a superfluid of light particles whose paired, counter-rotating vortices tile space into the lattice. Two clues invite a question the rest of the paper never asks. First, the substrate is scale-free — the ladder chapter reads the same \sqrt2 handshake at the sheet, the cell, and the core, a discrete scale invariance with no length of its own. Second, the mechanism that makes a dc1 particle is itself a paired, anti-phase breath. Put the two together and the question writes itself: what if a dc1 vortex is not fundamental, but is itself a paired vortex of a still-finer condensate, breathing anti-phase one tier down? If the substrate’s motif recurs upward without end, does it also recur downward?
We gave this tier a working name — dhe3, “dark helium,” for its ³He-like fermionic, anti-phase-breathing character — and spent a run of sessions asking the only question worth asking of a new tier: can it derive a number the framework leaves open, or does it merely relabel structure dc1 already has? The tempting prize was the big one — the condensation number \nu itself, the electroweak lift (WIP-30). That bet failed, cleanly and instructively, and we record why below. But the same tier, pressed on its scale structure rather than its channel count, pays out a different and smaller coin: it turns one of the framework’s last free parameters — the universe-decay constant \delta — into the substrate ladder read downward. That is this chapter.
A condensate beneath the condensate
The premise is precise. A dc1 particle is a paired, counter-rotating vortex in the dhe3 condensate, whose own quantum has mass m_\text{eff}\approx1.7 MeV and whose cell is the reach of that mass, \bar\lambda_C(m_\text{eff})\approx116 fm. The scale that separates the two tiers is not a new number: it is the condensation number the mass ladder already carries, \nu \;\equiv\; \frac{m_\text{eff}}{m_1} \;\approx\; 8.3\times10^{8}, \qquad \ln\nu \approx 20.54, the same \nu by which the electron’s mass is lifted off the substrate quantum, m_e/m_1=\alpha_{mf}\,\nu (WIP-30). The dc1 cell \xi\approx100\,\mum exceeds the dhe3 cell by exactly this factor, \xi/\bar\lambda_C=\nu. Crucially, the dhe3 cell is pinned by m_\text{eff} alone and never re-imports m_1, so positing the tier introduces no circularity — the scale is honest.
Two disclaimers belong here, up front. dhe3 does not derive \nu: the route that would have — the paired vortex supplying the tower’s coupling through a large coherent channel count C=N_c^2 — was tested and dies, because a dhe3 vortex sitting at its own marginal node presents only C\sim1 coherent doorway, not N_c\approx6.6. (The same doorway width \Gamma\approx0.67\,\Delta_\text{node} that pins the Weinberg angle is as wide as the entire core-state band, smearing the ladder into a single channel; a node gives the sharp Weinberg doorway or a resolved multiplicity, never both.) And dhe3 does not solve Lorentz invariance — it relocates the question one tier down, in the honest Volovik sense, and no further. What follows needs neither of those; it stands on the scale ratio \nu and the tower alone.
The one free number
The material properties of the substrate list a quantity the rest of the framework simply assumes: dc1 collisions are elastic to \varepsilon = 1-\delta, with \delta \sim 10^{-40} \quad\text{(the "universe-decay constant")} a free parameter. This tiny inelasticity is what keeps the universe from being exactly eternal — it is the microscopic reason the substrate can never quite finish relaxing, the seed of the arrow of time and, downstream, of the un-drained disequilibrium that the gravity chapter reads as the cosmological constant. A number that important should not be free. The dhe3 premise is what pins it.
Elasticity is topology, so the leak is an instanton
The framework already teaches that a dc1 particle is stable because it is topological — nontrivial winding cannot unwind by any smooth deformation. Read one tier down, that same statement acquires teeth: a topologically protected paired dhe3-vortex has no perturbative decay channel at all. It cannot shed energy a little at a time, because “a little” is a smooth deformation and topology forbids it. The only way to leak is to change the winding — to nucleate a defect in the dhe3 medium during a reconnection — and that is a barrier-crossing, a tunneling event. So the residual inelasticity is not a small perturbative rate; it is a nonperturbative instanton amplitude, \delta = e^{-S}, \qquad S \gg 1. This already answers the qualitative half of the puzzle. “Why is the universe-decay constant 10^{-40} and not order unity?” has the same shape of answer as “why is the proton stable” or “why is tunneling rare”: a protected object leaks only by tunneling, so of course the leak is e^{-\text{large}}. The smallness stops being a tuning and becomes a structural inevitability. What remains is to find S.
Why a power law, and not a catastrophe
Here is the subtle part, and the physical heart of the chapter. Leaking into dhe3 means exciting a mode of the finer tier — a disturbance of size \bar\lambda_C = \xi/\nu, sourced by a reconnection happening at the dc1 scale \xi. That is a demand for a Fourier component at wavenumber k\sim\nu/\xi: enormously higher than the scale of the process producing it.
For an ordinary, smooth vortex core — a profile that heals over one length like a \mathrm{sech} — the amplitude at such a wavenumber is set by the profile’s Fourier tail, which is exponentially small in the wavenumber: \sim e^{-\pi\nu/2}\sim e^{-10^{9}}. That is not 10^{-40}; it is zero to any precision the universe could ever express. A substrate with smooth cores would be exactly elastic: no leak, no arrow of time, no cosmological constant — a dead, eternal, perfectly reversible medium.
The substrate is not that medium. It is critical — scale-invariant, sitting at the gapless marginal point that is the very condition for the DSI tower to exist and for light to propagate at a single speed. A critical core has no exponential tail; its overlaps fall as power laws. So the leak is not e^{-\nu} but a power of \nu, \delta = \nu^{-p}, \qquad \ln\delta = -p\,\ln\nu = \mathcal{O}(90), which is a number the universe can express. This is the pivot: the same scale invariance that builds the tower is what lifts \delta off zero. The universe-decay constant being tiny-but-finite — rather than exactly zero — is the fingerprint of substrate criticality. No criticality, no leak; no leak, no arrow of time and no \Lambda. The tower does not merely coexist with the arrow of time; on this reading it causes it.
The tower read downward
The exponent p is fixed by the ladder itself. Crossing from the dc1 tier to the dhe3 quantum traverses the tower’s rungs — steps of \sqrt2 in scale, N = \ln\nu / S_M \approx 59 of them, with S_M=\tfrac12\ln2 the per-rung (Majorana) increment. The tower is a motion, a log-periodic (discrete-scale-invariant) breath whose phase advances by the DSI exponent s_0=\pi/S_M=9.06 per unit \ln(scale). By construction that makes each rung exactly \pi of DSI phase (s_0 S_M=\pi): every \sqrt2 step is one half-oscillation of the tower’s breath. The whole tier is therefore N half-oscillations, s_0\ln\nu=\pi N\approx186 radians of accumulated phase.
The leak action is half that phase — one quarter-turn of imaginary action per half-oscillation crossed: \boxed{\;S = \tfrac{s_0}{2}\,\ln\nu = \tfrac{\pi N}{2}, \qquad \delta = \nu^{-\pi/\ln 2} = e^{-\pi N/2} \sim 10^{-40.4}.\;} Against the phenomenological \delta\sim10^{-40} — a quantity the framework previously quoted only to an order of magnitude — this lands within 0.4 dex, from a formula with no adjustable knobs. The exponent is p=\pi/\ln2 = 4.53, the ratio of the per-rung action (\pi/2) to the per-rung mass increment (S_M=\tfrac12\ln2).
The reading that survives all the arithmetic is a single sentence: \delta and \nu are one ladder, read in opposite directions. The mass lift climbs it, \nu = \sqrt2^{\,N}=e^{+NS_M}; the leak descends it, \delta = e^{-\pi N/2}. The framework’s largest number and its smallest number are the same tower — up and down.
| per-rung leak action c | S = cN | \delta \sim | reading |
|---|---|---|---|
| c=1 | 59 | 10^{-26} | one action unit per rung |
| c=\pi/2 | 93 | 10^{-40} | half-oscillation per rung |
| c=\pi | 186 | 10^{-81} | full oscillation per rung |
| observed | 92 | 10^{-40} | substrate-particles.qmd |
The honest debt is exactly one number: the per-rung increment c=\pi/2. It is motivated — half an oscillation per half-octave — but it is not yet derived from the reconnection action, and the bracket c\in[1,\pi] spans \delta\in[10^{-26},10^{-81}]. The robust content is the structure (\delta=e^{-cN}, so \delta is the tower’s own N read as a leak); the coefficient is the owed piece.
The coefficient is an adiabaticity
That owed number has a name, and the route to earn it can be carried out far enough to change what is owed. Read the descent through a rung as a Landau–Zener passage: the protected dc1 level is swept through an avoided crossing with a dhe3 level, and the leak is the diabatic-transition amplitude. For a crossing swept linearly in DSI phase, the exact (Dykhne–Davis–Pechukas) leak action is c = \pi\,\Gamma, \qquad \Gamma = \frac{\Delta^2}{4\,|\dot\epsilon|}, with \Gamma the passage’s adiabaticity — the gap \Delta against the sweep rate of the detuning \epsilon. So the three rows above are not three guesses but three adiabaticities: c=1 is \Gamma=1/2\pi (a nearly sudden crossing), c=\pi is \Gamma=1 (the adiabatic edge), and the chapter’s c=\pi/2 is \Gamma=\tfrac12 — a half-adiabatic passage. “Why \pi/2?” is now exactly “why is the rung crossing half-adiabatic?”
Two things make that reframing more than bookkeeping. First, the two limits of \Gamma reproduce this chapter’s own two regimes — a consistency check, not an input. The adiabatic limit \Gamma\to\infty sends the leak amplitude e^{-\pi\Gamma}\to0: the smooth-core, exactly-eternal universe of the previous section. The sudden limit \Gamma\to0 sends it to 1: no protection at all. The observed \delta sits at the marginal knee between them, precisely where the criticality argument already places it. Second, “half-adiabatic” is not arbitrary: a rung is a half-oscillation of DSI phase (s_0S_M=\pi), and \Gamma=\tfrac12 is the passage that is neither eternal nor unprotected. What the calculation does not do is force \Gamma=\tfrac12 over \Gamma=1 from first principles — both can be called “marginal” — and it honestly surfaces a second factor of two, since the leak amplitude e^{-\pi\Gamma} and probability e^{-2\pi\Gamma} differ by exactly the bracket’s width. So the Landau–Zener reading turns one owed coefficient into one owed number with a mechanism — the marginal adiabaticity \Gamma=\tfrac12 (scripts/dhe3_landau_zener_perung.py).
The crossing lives at a Bragg point
Pushed one step further, into the frame where \Gamma is honestly a pure number, the picture both sharpens and corrects a shortcut (scripts/dhe3_marginal_bdg_perung.py). The correction is dimensional. The doorway width the node BdG already delivers, \Gamma_\text{width}\approx0.67\,\Delta_\text{node}, is an energy — it is what pins the Weinberg angle. The Landau–Zener adiabaticity \Gamma=\Delta^2/4|\dot\epsilon| is a dimensionless exponent. They are not the same number: a genuine sweep needs a clock, and the two differ by the DSI clock rate s_0 — so “the same 0.67\,\Delta_\text{node} the node already gives” is too glib. The frame that carries no hidden clock is the descent read in the log-radial coordinate x=\ln r, where the whole problem is dimensionless: the marginal mode is a plane wave u\sim e^{\pm i s_0 x} of wavenumber s_0, and the per-rung reconnection is a potential periodic in x with period S_M.
There a structural fact falls out for free. That potential’s first Bragg wavevector is \pi/S_M, which is exactly s_0 because s_0 S_M=\pi. The marginal DSI mode sits precisely at the first Bragg point of its own rung lattice — the same s_0 S_M=\pi that makes a rung a half-oscillation, now read as band structure. It gives the criticality hinge a sharper mechanism than “smooth vs critical core”: perfect scale invariance is no reconnection — no gap, the mode propagates freely, \delta=0 (exactly eternal); any reconnection opens a gap resonantly at the mode, which sits dead-centre in it, converting free propagation into a per-rung evanescence c=\mathrm{Im}(q)\,S_M. That the mode sits on the Bragg resonance is why the leak is a power law and not e^{-s_0} — criticality couples the reconnection to the mode maximally. What this still does not do is close the coefficient: the size of the leak is the size of the gap the reconnection opens, and c=\pi/2 needs that gap to be an O(1) fraction of the tower’s own supercritical coupling s_0^2 — natural, not forced. The honest target is now that one clock-free ratio, the reconnection gap of the marginal band, rather than a bare \pi/2.
The obvious move is to source that gap from the node BdG the framework already solves — its doorway resonance supplies a genuine matrix element, the self-energy whose width \Gamma_\text{width}\approx0.67\,\Delta_\text{node} pins the Weinberg angle. Tried, it fails, and the failure is the useful part (scripts/dhe3_reconnection_matrix_element.py). The doorway width is the wrong object twice over. By units, it is an energy — an in-plane resonance width — while the band gap is the dimensionless log-radial evanescence just described; bridging them needs the breath clock, which is exactly the s_0 factor the doorway does not carry. By process, the doorway width is decay into the gapless same-tier (dc1) bulk at fixed winding — an O(1), unprotected channel (\Gamma_\text{width}/\omega_0=1/2.99) — whereas the leak that sets \delta is a winding change: a dc1 particle is a paired dhe3 vortex, and leaking one tier down means changing its dhe3 winding, a core reconnection. An O(1) resonance width can never produce e^{-92}; only topology turns O(1) into e^{-\text{large}}. So the per-rung gap is not a matrix element of any fixed-winding BdG spectrum. The tempting next guess — that it is the imaginary action of the winding-change saddle, the modon-core reconnection action — can now be tested, because that action has since been computed from the tier above (the dc1 modon-core phase-slip barrier, a Gross–Pitaevskii black soliton / vortex nucleation, scripts/reconnection_barrier_from_above.py): \alpha\equiv S_\text{rec}/E_\text{min}\approx0.2–0.6, a clean O(1). Converted into the log-radial band by the DSI clock (V_1=4\pi\alpha, an explicit and factor-robust conversion, scripts/reconnection_band_conversion_phase1b.py) it gives a weak gap V_1\approx3–8 and \delta\sim10^{-1}–10^{-7} — some thirty orders short of 10^{-40}. A reconnection barrier is a fraction of E_\text{min}, hence a weak log-space potential; it cannot be the gap \delta needs. That gap is therefore not any reconnection action but the tower’s own supercritical coupling s_0^2 — the marginal-node pairing of WIP-26 — and the owed coefficient collapses into that one open problem rather than a separate instanton (next section).
Two near-misses, explained
Working the tier down this way dissolves a coincidence the exploration had earlier flagged and set aside. The tower’s supercritical coupling is g=-(s_0^2+\tfrac14)=-82, and \ln\delta=-\tfrac{s_0}{2}\ln\nu=-93; the two looked suspiciously close, and were quarantined as “suggestive, not to be built on.” The resolution is now clean: both are functions of the one DSI exponent s_0=9.06 — one as s_0^2, the other as \tfrac12 s_0\ln\nu — so both land near -85\pm10 without being equal, and neither leans on the other. The near-miss is explained, not exploited.
The mechanism has an established sibling one tier up — but, sharpened here, a sibling and not a twin. A photon below the modon floor is a delocalized winding that cannot reach the core-reconnection that would unwind it, so its residual dispersion is an exponential, \delta_g\sim e^{-\alpha\,\nu_\text{floor}/\nu}, whose coefficient the open-problems ledger once deferred to “the modon-core reconnection action.” That action has now been computed from above — the dc1 modon-core phase-slip barrier, \alpha\approx0.2–0.6 — and it does fix the sub-floor law and the FRB per-cell scattering amplitude (\varepsilon<6.5\times10^{-16}): those two genuinely are one reconnection barrier, read at one tier. But \delta is not the third member of that set. Its coefficient is a band gap, not a reconnection barrier, and the barrier is thirty orders too weak in log-space to be it (previous section). So the tier-below leak and the sub-floor exponential share a mechanism — topological protection forbidding perturbative decay — but not a number: the sub-floor coefficient is the reconnection action (now computed), while \delta’s is the tower coupling s_0^2 (WIP-26). The earlier reading that made all three “one calculation” was too strong, and the computed barrier is exactly what corrects it.
Honest accounting
What is robust. Two things need no indulgence. (i) The mechanism: topological protection forbids perturbative decay, so \delta=e^{-S} is forced — the smallness is structural, not tuned. (ii) The criticality hinge: a smooth core gives e^{-\nu}\approx0 (an exactly-eternal universe), and only the substrate’s scale invariance converts that to a power law \nu^{-p} of the right order. This ties the arrow of time to the same criticality that carries the tower, and it holds regardless of the exact exponent.
What is owed. The per-rung increment \pi/2 is motivated but not derived; it is the single coefficient standing between “a mechanism that lands the right order of magnitude” and “a derivation of 10^{-40}.” The Landau–Zener treatment above narrows it without closing it: it recasts c=\pi/2 as the marginal adiabaticity \Gamma=\tfrac12 (via the exact c=\pi\Gamma) and reproduces the eternal-vs-unprotected limits as a check. Carrying that into the log-radial frame corrected a units conflation — the dimensionless \Gamma is not the 0.67\,\Delta_\text{node} doorway width — and reframed the debt as one clock-free number: the per-rung gap the reconnection opens in the marginal band, whose O(1) size (a fraction of s_0^2) sets c. That gap has now been chased to its source. It is not the node-BdG doorway (wrong by units and process), and it is not a modon-core reconnection action either: that action was computed from the tier above (\alpha\approx0.2–0.6) and, carried into the band, is some thirty orders too weak. What is left is the only object of the right size — the tower’s own supercritical coupling s_0^2. So the owed coefficient is not a separate debt at all; it is WIP-26: whether the marginal-node paired vortex opens a Bragg-resonant gap that is exactly the \pi/2 fraction of s_0^2. The same open problem that owes the tower’s existence — that anti-phase pairing drives the coupling supercritical — owes its \delta coefficient too, and the tower’s rung count N\approx60 rides on it as well. One debt, not three.
What this does not claim. It does not derive the cosmological constant. \delta is the microscopic decay constant — the per-collision inelasticity; the observed \Lambda comes from the state quantity \delta T/T_c, the standing disequilibrium, which the gravity chapter fixes geometrically as (m_1/M_\text{Pl})^2, independent of \delta. The relationship is one of provision, not identity: \delta is what makes the substrate dissipative at all — without it there is no relaxation to fall short of — while the depth of the shortfall is set elsewhere. Nor does it derive \nu; that bet failed, and this chapter is careful to spend only the scale ratio, never to re-earn it.
What would move it out of Speculation
Two handles, in order of decisiveness. Close WIP-26. The owed coefficient has been localized: four passes (Landau–Zener → the marginal adiabaticity \Gamma=\tfrac12; the log-radial band picture → a clock-free reconnection gap at the mode’s Bragg point; the node-BdG doorway → ruled out by units and process; and now the reconnection barrier computed from above → ruled out as thirty orders too weak) converge on the one object of the right size, the marginal-node Bragg gap as a fraction of the tower coupling s_0^2. Showing that the anti-phase paired vortex opens a gap of exactly the \pi/2 fraction — the same WIP-26 Bogoliubov calculation that owes the tower’s very existence — would derive \delta and graduate the chapter in one stroke. The tower’s existence and its smallest number are now the same debt. Consistency downstream. Any independent determination of the relaxation dynamics (WIP-16) that ties the accumulated per-collision leak to the observed disequilibrium would corroborate or break the value. What is no longer a handle: the “shared THz coefficient.” The sub-floor dispersion onset and the FRB amplitude fix the reconnection action (now computed), but that is a different number from \delta’s, so a terahertz measurement constrains the sub-floor law, not this one.
Putting the section in context
The tier below began as a bid for the framework’s biggest open number and did not win it — the channel-count route to \nu is dead, and the chapter says so plainly. What it delivers instead is quieter and, in its way, more satisfying: the smallest number in the framework, the free universe-decay constant \delta\sim10^{-40}, is not free at all but the substrate ladder read downward, \delta=\nu^{-\pi/\ln2}=e^{-\pi N/2} — the same tower whose upward climb is the mass lift \nu. A topologically protected vortex cannot leak perturbatively, so it leaks by tunneling; a critical substrate makes that tunneling a power law rather than a catastrophe; and the tower fixes the power. One coefficient — the half-oscillation per rung — is still owed, but it now has a single address: the marginal-node coupling of WIP-26, the same calculation the tower’s existence waits on. The attempt to earn it cheaply, as a reconnection instanton borrowed from the tier above, was carried out and failed by thirty orders — which is itself the reason the debt is now WIP-26 and nothing smaller. Until that is earned this belongs here, among the speculations. But the shape of the result — the framework’s largest and smallest numbers turning out to be one ladder, and the arrow of time turning out to be the price of the same criticality that carries the tower — is the kind of thing that does not usually stay in the Speculation section for long.