If the centre of a black hole is the informational endpoint of collapse, the Big Bang is the informational starting point of expansion. This chapter asks what happens when we begin at the zero-entropy state and allow entropy to increase. The answer is a universe — not assumed, not hardcoded, but derived from the combinatorics of patterns in a mutating bitstring.
Standard cosmology treats space as a pre-existing background: an empty arena into which matter and energy are placed, and which the Big Bang caused to expand. This framing inherits a hidden assumption — that space exists prior to the structures it contains. The framework developed here discards that assumption entirely.
Space is relational. It is not a container. It is what an internal observer perceives when distinguishable structures exist at varying informational distances from one another. When no structures exist, there is no space. There is no distance. There is nothing extended.
This makes the zero-entropy initial state the natural starting point. A completely uniform all-zero bitstring contains no detectable patterns, no distinguishable substructures, and no internal variety of any kind. Under any relational decoding, an observer embedded in this state perceives a single unextended point. The universe begins not with an explosion into a pre-existing void but with the first departure from absolute uniformity.
We evolve the system by a sequence of random bit-flip mutations. No equations of motion are specified. No external clock drives the evolution. The only input is the bitstring and the rule that bits can flip.
As bits flip from zeros to ones, the Shannon entropy of the configuration increases. Minimal detectable patterns — short binary subsequences that match a specified target — begin to appear. Their density across the string grows. An internal observer perceives this proliferation of distinguishable structures as the expansion of space.
This identification is not analogical. The geometric and informational descriptions are, by the representational equivalence of Chapter 1, two readings of the same object. The statement that the universe is expanding and the statement that the entropy of the underlying configuration is increasing are not two correlated facts. They are one fact, expressed in two languages.
The implication is sharp. The thermodynamic arrow of time — systems moving from order to disorder — and the cosmological arrow of time — the universe expanding — are not two independent phenomena that happen to point in the same direction. They are the same phenomenon. The second law of thermodynamics and the expansion of the universe are one feature of reality, not two.
To extract matter from the mutating bitstring, we pass the configuration through a hierarchy of recursive filters. Each filter looks for patterns at a different scale, using the output of the layer below it as its input.
When the abundance of each structural tier is plotted against the system’s entropy, a consistent pattern emerges across all levels. Each tier follows a non-monotonic rise-and-fall trajectory: zero at the start, a peak at an intermediate entropy, and a slow decay toward the high-entropy limit as the underlying bitstring approaches uniform randomness and the patterns that define each structure become rare again.
Every one of these curves is well-described by a lognormal distribution.
The immediate question is whether the lognormal shape is an artefact of the particular filter definitions chosen. The answer is no — and establishing this is the most important result of this chapter.
We tested more than two hundred distinct filter configurations. We varied pattern lengths, density thresholds, hierarchy rules, and coordinate decoding maps across linear, logarithmic, and randomly chosen projections. We counted occurrences of completely arbitrary patterns — a fixed binary string chosen at random, with no geometric interpretation whatsoever — as a function of the system’s entropy. We applied the analysis not to the simulated configuration but to the raw execution trace of the host computer running the simulation.
In every case the result was the same: zero structure at zero entropy, a lognormal emergence curve, and a long-tailed decay at high entropy. The peak location and width vary with the filter definition. The lognormal form does not.
This representational independence is the key result. It establishes that the lognormal distribution is not a property of any particular physical model, geometric embedding, or particle definition. It is an intrinsic combinatorial property of how distinguishable patterns proliferate in an entropy-increasing binary string. The lognormal is what you get whenever you define any notion of structure and ask how its abundance varies as the underlying information content grows.
What we call a particle spectrum is a coordinate choice — a gauge selection — overlaid on a universal informational phenomenon. Different filters yield different particle definitions. The statistical distribution beneath them is invariant.
Standard cosmology requires the early universe to have been extraordinarily smooth and low-entropy — a constraint so severe that Roger Penrose proposed the Weyl curvature hypothesis to explain it geometrically. In the present framework, no such explanation is needed. The zero-entropy initial state is not a finely tuned miracle. It is the unique shortest-description configuration in the entire space of 2L bitstrings. It is the natural starting point, not a special one. The framework does not require smooth initial conditions; it derives them.
As established above, the thermodynamic and cosmological arrows of time are the same feature read from different perspectives. No additional mechanism is required to align them. They cannot point in different directions because they are not different directions.
The entropy saturation curve of a randomly mutating bitstring relaxing toward equilibrium takes the form
Einstein’s De Sitter vacuum solution — the standard model for a universe dominated by a cosmological constant — has a scale factor a(t) ∝ eHt. Under the substitution t → lnt, this becomes a ∝ tH, which matches the entropy saturation curve in the bit-flip parameterisation. Empty space accelerating due to dark energy and a closed statistical system relaxing toward equilibrium are the same curve, written in two different languages. The cosmological constant is not a new ingredient. It is the geometric reading of thermodynamic relaxation.
Three results follow from the simulations and the filter-independence analysis.
First, cosmic expansion is the geometric reading of entropy increase. No cosmological constant, inflaton field, or equations of motion are required. Expansion is what an internal observer perceives as the underlying bitstring moves away from its zero-entropy boundary state.
Second, hierarchical matter structures emerge natively from recursive pattern filters applied to the mutating bitstring. Their abundance follows a lognormal distribution that is invariant across more than two hundred filter definitions, coordinate systems, and representational choices.
Third, the De Sitter expansion curve and the entropy saturation curve are the same mathematical object. The accelerating expansion of the universe is a thermodynamic phenomenon, not a dynamical one requiring new physics.
What this chapter does not yet establish is how these emergent informational structures give rise to quantum mechanics. The lognormal tells us how many structures exist at each entropy level. It does not yet tell us how an observer inside those structures experiences them, or why that experience follows the rules of quantum theory. That is the question the next chapter answers.