Large-N factorization and the planar expansion

Consider a gauge or matrix theory with adjoint fields and an action convention containing N times a single trace. Hold the dimensionless ’t Hooft coupling fixed as N increases. Drawing each adjoint propagator as two color lines turns a connected graph into an oriented ribbon surface. Index loops give powers of N, vertices and propagators contribute their normalization factors, and the final exponent is the Euler characteristic of the surface.

For a closed vacuum graph the characteristic is two minus twice the genus, so a handle costs two powers of N relative to a planar graph. The coefficients remain nontrivial functions of the fixed coupling and can have different convergence properties. This topology statement concerns the perturbative organization; it does not sum all planar diagrams at strong coupling. Fundamental matter boundaries, nonorientable gauge groups and different large-N limits change the counting and must be specified separately.

An unnormalized single-trace insertion makes a boundary in the ribbon counting. With b such insertions, a connected genus-g contribution scales with two minus twice genus minus b powers of N in this matrix convention. Dividing each trace by N gives observables with order-one expectation values and connected two-point fluctuations of order inverse N squared. Their leading two-observable expectation therefore factorizes into the product of the separate expectations when those expectations are nonzero.

CFT discussions frequently use a different normalization in which the connected two-point coefficient is order one. The corresponding connected three-point interaction is then order inverse N, rather than inverse N to the fourth power for the normalized trace convention. These are compatible statements after rescaling every insertion consistently. Subtracting an operator's expectation value can also change which disconnected term is present, so a slogan about factorization must be translated into an explicit connected correlator.

A weakly interacting bulk interpretation matches the suppression of suitably normalized boundary interactions. A weakly curved local Einstein description needs more: an appropriate spectrum, a gap suppressing additional light higher-spin or string states, and suitable coupling regimes. Some large-N theories have many light higher-spin excitations and do not admit a simple two-derivative gravity truncation. A planar limit at weak boundary coupling is therefore not the same approximation as classical supergravity at strong boundary coupling.

The genus analogy is useful evidence for particular duality conjectures, alongside symmetry and detailed observable checks. It is not an empirical demonstration of extra gravitational dimensions. The probability lab supplies exact independent occupancy statistics as a limited reference for factorization. Independent random outcomes in that instrument have no color indices, ribbon topology, connected gauge correlators or string worldsheet expansion, and cannot establish the dynamical assumptions needed for a holographic spacetime.