Conformal Cyclic Cosmology

The Program in Brief

Roger Penrose’s Conformal Cyclic Cosmology begins from an observation that is hard to un-see once it has been pointed out. In the extremely remote future of an accelerating universe, after the stars have burned out, the black holes have evaporated, and whatever massive particles survive have decayed or been diluted beyond relevance, there is nothing left that carries a scale. Massless fields do not know how big anything is. And a universe whose content cannot measure distance is, in Penrose’s reading, not really a large empty universe at all — it is a conformal geometry, and a conformal geometry can be rescaled.

The move is then breathtakingly direct. Rescale the remote future so that its conformal infinity becomes a finite spacelike surface, and identify that surface with the initial surface of a new Big Bang. The universe becomes a sequence of aeons, each one’s infinite future glued smoothly to the next one’s beginning. No first aeon is required.

CCC is not only a picture of succession; it is also an answer to what Penrose has long argued is the deepest problem in cosmology — the extraordinary specialness of the initial state. His Weyl curvature hypothesis makes that specialness precise: the gravitational degrees of freedom, measured by the Weyl tensor, were essentially zero at the bang and grow thereafter as matter clumps and black holes form. Gravitational entropy starts near its minimum and rises. In CCC that hypothesis is no longer an extra postulate: the conformal boundary of the previous aeon is automatically Weyl-flat, because that is what a conformal boundary of a de Sitter-like future looks like. The arrow of time is explained by the geometry of the crossover rather than stipulated at it.

It is a serious, elegant, and unusually falsifiable program, and the framework agrees with more of it than one might expect.

Cyclicity Without a First Cycle

The framework independently arrives at a cyclic cosmology, and — like CCC — one with no first cycle. Observer-Centric Cyclic Cosmology begins from an obstruction rather than a rescaling: true de Sitter heat death is axiomatically forbidden, because relational invariants are permanent and complete thermalization would require destroying them. Something must persist below the second law’s drive.

What follows is a four-phase cycle. Complex observers assemble, dissolve, leave behind a floor of one-bit minimal observers that must keep interacting (Multiplicity), and eventually support a new composite. The state on a Cauchy slice is finite-dimensional, of dimension $D \lesssim A/(4\ell_P^2)$; once $D$ independent relational invariants have accumulated, every further invariant is a redundant constraint. Combined with the exclusion of quasi-periodicity — all observer frequencies must be commensurable for a composite to close its loop — this forces exact periodicity rather than mere recurrence. The causal history keeps growing; the state returns identically.

So both programs reject heat death and a singular beginning alike, and both replace “the first moment” with a structure that has no distinguished starting point. Penrose gets there by making the succession of aeons geometrically seamless. The framework gets there by making the cycles numerically identical: “was there a first cycle?” is not merely unanswerable but ill-posed, since nothing distinguishes one traversal from another. The causal DAG is the universal cover of the cycle — a helix that projects onto the same circle no matter how many times it winds.

How Scale Loses Meaning: Conformal Versus Epistemic

This is the point of deepest contact, and the framework says so explicitly. From Cyclic Cosmology: “Penrose achieves scale-stripping through conformal invariance; the framework achieves it through epistemic capacity.”

Penrose’s route requires that the far future contain only conformally invariant content, so that the metric is defined merely up to a Weyl rescaling and “how big” stops meaning anything. The framework reaches the same conclusion from the other side. Duration is a comparison — this interval against that one — and a comparison requires an observer with enough memory to hold one measurement while making another. At the floor phase, no such observer exists: a one-bit loop cycles $0 \to 1 \to 0$ with no room to record that it has cycled before. Spatial extent fails for a parallel reason, since geometry in the framework is constituted by the relational invariants observers can actually access, and at the floor only the substrate level is instantiated. The floor phase is therefore not long or short, large or small. Those predicates have no referent inside it.

The conclusions coincide: to anything present at the transition, the end and the beginning are indistinguishable. The mechanisms do not. Penrose postulates a symmetry of the field content; the framework derives the scale-stripping from the structure of what can be known — and so does not require rest mass to actually vanish in the far future, a physical assumption about particle decay that CCC has always had to carry. The related question of why exact conformal invariance should be fundamental rather than earned is taken up on the Turok comparison page.

The Low-Entropy Past, and Whether the Arrow Resets

Penrose’s estimate that the initial state occupies a fraction $\sim e^{-10^{123}}$ of gravitational phase space is one of the sharpest statements of the fine-tuning problem anyone has produced, and CCC is built to answer it: the Weyl-flat crossover makes the new aeon begin with minimal gravitational entropy.

The framework accepts the problem and rejects the arithmetic. Cosmological Arrow of Time argues that the number compares the early state against a phase space the universe never had access to. The accessible sector is partitioned by bootstrap level, matching rules, and coherence correlations, and it grows as the universe cools and new bound states freeze out. The early universe was a typical member of a small accessible sector, not an absurdly atypical member of a large one. On this reading the past hypothesis does not need explaining because it never described an improbability — and the arrow of time is the direction of hierarchy elaboration, which is why it tracks expansion.

Underneath sits a still more structural claim. In Entropy as Inaccessible Coherence, entropy is defined relative to an observer as the coherence it cannot reach, the second law is a theorem rather than a statistical tendency, and the entropy of the universe relative to itself is identically zero. There is consequently nothing that has to be reset at a transition. Time as Phase Ordering supplies the arrow directly as monotone growth of relational-invariant depth along the causal DAG, and Observer Loop Viability keeps it from saturating: $\Lambda > 0$ strictly, with the framework-computable lower bound $\Lambda \geq 3\pi/(C_0\ell_P^2) \approx 2 \times 10^{-122}\,\ell_P^{-2}$, so the de Sitter bath supplies events forever.

That last point is a quiet convergence worth naming. CCC needs a positive cosmological constant — without acceleration there is no spacelike conformal infinity to glue to the next aeon. The framework derives one, along with $w \geq -1$: phantom dark energy is excluded because a Big Rip would dissolve every coherence carrier (Dark Energy Equation of State). A Big Rip would also destroy CCC’s crossover. Both programs need the same feature of the dark sector, and one of them predicts it.

Where they cannot agree is on the bookkeeping across the transition. CCC must reconcile an enormous late-aeon entropy with a low-entropy new bang, and Penrose’s answer has been that black hole evaporation genuinely removes degrees of freedom, so the phase-space volume itself shrinks. The framework cannot take that route: its Hawking radiation and black hole entropy derivations make horizon entropy observer-relative loop-counting and preserve information through evaporation. It does not need the route either — its second law is never violated for any observer in any phase, because the cycle runs in the space of which observer levels exist, not in any observer’s entropy.

The Pattern: Cyclic by Geometry, Cyclic by Bookkeeping

Across every theme the structure repeats. Penrose reasons from the geometry of the whole spacetime: a conformal structure, a boundary, a rescaling, a curvature tensor whose vanishing defines specialness. The framework reasons from what a bounded observer can hold: how many bits it has, which invariants it can access, whether its loop can close. The two lines of reasoning keep landing on the same propositions — no first cycle, no meaningful scale at the transition, positive $\Lambda$, an arrow that demands a structural rather than a stipulated explanation — while disagreeing about which of those propositions is doing the explaining.

Where They Genuinely Part Ways

The bang itself is the sharpest split. CCC passes through a conformal boundary; the framework denies there is anything to pass through. Singularity Resolution bounds curvature at $K \leq c_K/\ell_P^4$ from the Planck-scale cutoff and then applies contraposition to Penrose’s own singularity theorems: if the conclusion fails, a premise must, and the premise that fails is the energy condition. Loop-closure pressure diverges near $\rho_P$, Raychaudhuri defocusing follows, and the bang is a bounce at finite density. Penrose’s theorems are used, respectfully and literally, to argue that the object CCC rescales does not exist.

Second, the observational split is clean. CCC’s most striking claim is that the previous aeon leaves marks — concentric low-variance rings from black hole encounters, and “Hawking points” from supermassive black holes that evaporated before the crossover. Penrose and collaborators report detections; other analyses, including work with the Planck data and against simulated skies, do not find the significance to survive, and the status of the signal remains genuinely disputed. The framework’s prediction here is unusually stark: no inter-cycle fossils exist, not because they are hidden but because the state at the start of each cycle is identical, and the DAG that does record the difference is inaccessible to any observer within the cycle. If a CCC signature is firmly established, the framework’s exact-closure claim is falsified.

Third, the two disagree about what a cycle is. CCC’s aeons are ordered and physically distinct — information crosses between them, which is what makes the signatures possible. The framework’s cycles are the same cycle, traversed; the succession lives only in a causal record with no dynamical consequence.

What the Framework Takes From It

Penrose identified the right problem and named the right pressure point. That the transition requires scale to become meaningless is his insight, and the framework’s epistemic-capacity argument is a re-derivation of it rather than a rival intuition — the debt is acknowledged in the derivation itself. So is the insistence that the low-entropy past must be earned rather than assumed; the framework’s dissolution of the fine-tuning number is an answer to Penrose’s question, and the question would not have been posed so precisely without him.

There is a methodological lesson too. CCC committed to a concrete, checkable CMB signature and let the data have a say. The framework’s exact closure makes the opposite commitment with equal sharpness, and the two claims cannot both survive. That is the healthiest possible relationship between two cosmologies.