What the paper is about: The author builds a four‑dimensional lattice model made only of bosons that reproduces a familiar quantum anomaly normally associated with fermions. This is the Adler–Bell–Jackiw (ABJ) or vector–vector–axial (VVA) anomaly — the same formal feature that, for example, helps explain the fast decay of the neutral pion in quantum chromodynamics. The unusual claim is that the model keeps this mixed anomaly exactly even at finite lattice spacing, not only after taking a continuum limit.

What the researchers tried: The construction uses a modified Villain discretization, a lattice setup that represents a compact boson field with an extra integer-valued link variable. That discretization gives exact integer topological quantities on the lattice and makes certain dualities exact. The key extra ingredient is a “no‑intersection” constraint that forbids intersections of vortex world‑sheets. This constraint can be implemented with a Lagrange multiplier, and when it is imposed the lattice model has two global U(1) symmetries with a mixed ’t Hooft anomaly of the vector–vector–axial type.

How they explored the model: Because forbidding intersections by a hard constraint is hard to simulate directly, the author studied two ways through numerics. First, they examined the standard (unconstrained) Villain model to locate its transition and to measure how often vortex intersections occur there. Second, they introduced a softened version of the constraint by adding a fugacity (a penalty) for intersections. Tuning that penalty interpolates between the unconstrained model and the hard no‑intersection model, which the author then used to probe the phase structure.

What they found so far: The unconstrained Villain model develops a substantial density of vortex‑worldsheet intersections right around its critical coupling. That observation implies the no‑intersection constraint should materially change the phase diagram. In preliminary simulations the unconstrained and constrained models agree at large coupling but diverge once intersections become common. The paper reports measurements of several observables (action density, various susceptibilities, specific heat) and two kinds of two‑point functions: the ordinary spin correlator and a correlator tied to the defect (Lagrange multiplier). The defect correlator’s tail grows with lattice size, and the spin susceptibility shows behavior typical of a symmetry‑breaking phase at large coupling, but the author has not yet seen a clean, long‑distance plateau or an unambiguous signature of a new continuous critical point.