Chapter 50
Smith’s Phase Transition and the Density of Attractors
50.1 Life as a Fourth Geosphere
Eric Smith proposes that life constitutes a fourth geosphere — alongside the lithosphere, hydrosphere, and atmosphere — that emerged from prebiotic chemistry through a phase transition. The key mechanism is autocatalysis: certain chemical reaction networks catalyze their own production. Once such a network is initiated, it becomes self-reinforcing, crossing a thermodynamic threshold beyond which it is stable against fluctuation. The emergence of life is the crossing of this threshold.
50.2 The Density of Recursive Attractors
The connection to the present framework runs through the logical topology of Part II and the fixed-point structure of Section 49.
The set of replicating terms in the Rho calculus, \[\mathrm{Rep} \;=\; \bigl\{ [P] \in \terms(\Rho)/{\bisim} \;\mid\; P \bisim P \mid P' \;\text{ for some }\; P' \bisim P \bigr\},\] is dense in the logical topology \(\mathcal{O}_{\HML}\): every non-empty open set \(\llb\phi\rrb \subseteq \terms(\Rho)/{\bisim}\) contains a replicating term.
Given any process \(P\), one can construct a process \(\hat{P}\) bisimilar to \(P\) on all observable behaviors but additionally capable of replication. The construction uses the concurrent Y combinator to add a replication sub-process running in parallel: \(\hat{P} = P \mid \mathbf{Y}(\mathrm{copy})\) where \(\mathrm{copy}\) is a process that duplicates \(P\) onto a fresh channel. The added component runs in parallel and does not interfere with \(P\)’s observable interactions, so \(\hat{P}\) satisfies all the same HML formulae as \(P\) — except formulae that specifically probe the replication channel. Since every open set \(\llb\phi\rrb\) is defined by a single formula \(\phi\), and since \(\phi\) generically does not probe the replication channel, \(\hat{P} \in \llb\phi\rrb\). Making this argument rigorous requires showing that the added replication machinery can always be hidden from the formula \(\phi\) — a careful but plausible argument deferred to future work.
50.3 Phase Transition as Basin Crossing
Consider a prebiotic population of processes in the Rho calculus, wandering through the bisimulation quotient \(\terms(\Rho)/{\bisim}\) under selective pressure. Interactions that succeed — those with high path amplitude \(W(\gamma)\) and affordable cost \(\vect{c} \leq \vect{A}\) — are reinforced. Interactions that fail deplete the vectorial account. The population drifts under this selection pressure without direction or goal.
By the density of \(\mathrm{Rep}\) (Conjecture 50.1), every open neighborhood in the logical topology contains a replicating term. The wandering population therefore inevitably enters a neighborhood of \(\mathrm{Rep}\). This is not a low-probability event requiring fine-tuning; it is structurally guaranteed by density. The question is not whether the population encounters a replicator but when.
Once the population enters the basin of attraction of a replicating fixed point, the dynamics change qualitatively. A replicating process produces copies of itself, increasing \(\mathrm{CN}(P, \mathcal{N})\). More copies mean more interactions, more reinforcement, higher path amplitudes for the replication pathway, lower effective cost per replication event. The population converges onto the replicating fixed point.
What emerges from the funnel is a population with a qualitatively new feature: hereditary replication with variation — the precondition for Darwinian evolution. This is Smith’s phase transition, made precise in the GSLT framework.
The new geosphere, in our terms, is the compact coherence cluster (Part II, Section 6) centered on the replicating fixed point: the maximal compact subpopulation within which the replicating process and its variants can reach consensus. The phase transition is the crossing of the basin boundary in the logical topology — the moment at which the subpopulation containing the replicator becomes compact, and consensus (here, consensus on the replication strategy) becomes achievable.
After the phase transition, both \(\mathrm{AI}\) and \(\mathrm{CN}\) increase together. Selection amplifies processes that replicate more efficiently, which typically requires more sophisticated catalytic machinery — higher \(\mathrm{AI}\). And more efficient replication produces more copies — higher \(\mathrm{CN}\). The joint biosignature of Assembly Theory is therefore not merely an empirical observation but a theorem about the post-transition dynamics near a replicating fixed point, given density of attractors and account-governed selection pressure. This is a key prediction of the framework: wherever a phase transition of this kind occurs, the joint \(\mathrm{AI}\)-\(\mathrm{CN}\) signature will be observed.
The prediction can be sharpened from a direction into a rate. Proposition 53.9 shows that the copies are not merely a by-product of efficient replication but are forced by depth: an ecology using a tower of height \(h\) on media of reliability \(p\) must hold at least \(m p^{-h}\) learners at that depth, so the two coordinates lie on a frontier whose slope is \(\log(1/p)\). The joint rise is therefore not two findings that happen to correlate; it is one relation, and it has a measurable exponent.