Measure Geometry and the Origin of Probability

Additivity, Invariance, and Stability of Measures

In the pre-physical state space $\mathcal{O}$, physical predictions depend on a measure $\mu$ defined over subsets of ontic states. For this measure to have physical meaning, it must satisfy three structural requirements: additivity, invariance, and stability.

Additivity ensures that mutually exclusive alternatives combine consistently. If two disjoint subsets $A,B \subset \mathcal{O}$ represent incompatible outcomes, then

Invariance requires that $\mu$ be unchanged under symmetry transformations of the state space that preserve physical equivalence, such as basis changes or unitary transformations in quantum sectors. Finally, stability demands that small perturbations of the state space or its decomposition do not lead to discontinuous or radically different measure assignments. Measures violating stability lead to physically ill-defined or observer-dependent predictions.

Projective Decomposition of State Space

In regimes where quantum descriptions are valid, the relevant structure of $\mathcal{O}$ reduces effectively to a complex Hilbert space $\mathcal{H}$. Physical alternatives correspond not to vectors in $\mathcal{H}$ but to rays, since global phase has no physical significance. The natural decomposition of $\mathcal{H}$ is therefore projective: outcomes correspond to orthogonal subspaces or projection operators $\{P_i\}$ satisfying

A measure on outcomes must therefore be defined on these projective components. Any admissible measure must respect the orthogonality structure and be compatible with coarse-graining, meaning that combining subspaces corresponds to summing their measures.

Uniqueness of the Squared-Norm Measure

Under the combined requirements of additivity, invariance under unitary transformations, and stability under refinement of projective decompositions — together with the additional hypothesis that $\dim\mathcal{H}\ge3$ (essential to the argument; the conclusion is known to fail at $\dim\mathcal{H}=2$, see Appendix A, § A.2, and Gleason 1957) — the measure on $\mathcal{H}$ is uniquely fixed up to normalization. Explicitly, for a normalized state $\ket{\psi}$ and a projection $P_i$, the only structurally stable assignment is

This squared-norm measure arises naturally from the geometry of Hilbert space: it corresponds to the induced volume measure on the unit sphere projected onto orthogonal subspaces. Any alternative assignment either breaks unitary invariance, fails additivity under refinement, or is unstable under infinitesimal perturbations of the state.

Derivation of the Born Rule from Stability

The Born rule is thus not postulated but derived as the unique measure compatible with structural stability. Outcome weights are fixed by the requirement that physically equivalent decompositions yield the same predictions and that small perturbations of the measurement interaction do not drastically alter outcome statistics.

From this perspective, probability is not a primitive concept but an emergent property of measure geometry on the pre-physical state space. Frequencies observed in repeated experiments arise from measure concentration: almost all admissible realizations consistent with the state $\ket{\psi}$ are distributed according to the squared-norm weights.

Failure of Alternative Probability Assignments

Alternative proposals for outcome weights, such as linear amplitudes, higher powers of amplitudes, or context-dependent rules, fail one or more structural criteria. Linear measures violate additivity under superposition. Higher-order measures are unstable under decomposition and lead to basis-dependent predictions. Contextual assignments depend explicitly on the measurement setup and thus lack invariance.

These failures are not merely aesthetic but structural: such measures lead to ambiguous or contradictory predictions when applied across different experimental arrangements. Structural stability therefore singles out the Born rule as the only physically admissible probability assignment within the unified framework.

In summary, probability emerges as a geometric consequence of measure structure on the pre-physical state space, unifying its role in quantum theory with the selection principles governing gravitational configurations.