Cancel a boundary anomaly with bulk inflow

Use natural units and a normalized compact U(1) background connection a, with unit charge derivative $d-ia$ and curvature $f=da$ locally. For an integer Hall phase on an oriented spin three-manifold M, the leading parity-odd response is $k\int a\wedge da/(4\pi)$ with integer k. The spin assumption and connection normalization are part of the quantization statement; bosonic non-spin systems or fractional Hall phases need their own global definition and additional structure.

Under $a\mapsto a+d\lambda$, the bulk action changes by a boundary integral. Stokes' theorem supplies its sign using the boundary orientation induced from M. An edge effective action with the opposite consistent gauge variation makes the total variation vanish. The cancellation is between functionals under the same gauge transformation, not between two unrelated choices of current. A boundary theory that is anomalous in isolation can thus be a valid component of the full higher-dimensional system.

Functional variation of the bulk action in its interior produces a current proportional to $k f/(2\pi)$. Its coefficient is twice the boundary gauge-variation coefficient. This is not a contradiction: the edge consistent current obtained from its effective action differs by a local Bardeen–Zumino contribution from the covariant current used in a full local charge-balance description. In two dimensions the Abelian covariant anomaly is twice the consistent one, rather than the factor three encountered for a four-dimensional chiral current.

For matter of physical charge q, write the normalized probe as $a=qA_{\rm phys}$ in natural units. The bulk response then gives Hall conductivity $kq^2/(2\pi)$, or $kq^2/h$ when Planck's constant is restored. The two-dimensional consistent boundary coefficient alone would undercount the physical Hall inflow if mistaken for the complete transported charge. Including the bulk-induced local contribution restores the correct balance. The net response is fixed in the specified Hall phase, while an edge's velocities, extra counterpropagating pairs, and detailed reconstruction are not determined by k alone.

The response description applies to weak probes varying slowly compared with the bulk gap and correlation scales, with temperature and frequencies low enough to avoid bulk excitations. Flux insertion on an annulus then pumps charge between its edges while conserving the total charge of the full sample and reservoirs. The edge itself can be gapless; adiabatic here refers to the bulk response, not a claim that every finite-size edge level remains in a unique isolated ground state through an entire pumping cycle.

Inflow establishes the required anomaly carried by the boundary description, but does not manufacture a unique microscopic spectrum from a topological coefficient. Interactions may reorganize boundary degrees of freedom while retaining the same net anomaly. Nor does attaching an unspecified bulk cure a stand-alone anomalous gauge theory without changing its physical definition. One must identify the actual bulk term, its quantization, the boundary conditions, and the combined conserved current. Only then is apparent boundary nonconservation an accounted-for transfer rather than an unresolved inconsistency.