Failure of disjoint additivity under magnetic Gauss constraints

In the quantum world, the rules that govern how particles interact are often written in the language of symmetry and constraint. Imagine a grid of tiny magnets, where the direction of each one is not free to point anywhere but must align with its neighbors to satisfy a strict local rule. This is the essence of a lattice gauge theory, a framework physicists use to model the fundamental forces of nature. In these systems, the "physical" states are only those configurations where every local rule is obeyed. A central question for decades has been how to describe the behavior of a large system by looking at its smaller, separate parts. If you know everything that can be measured in one isolated region and everything that can be measured in a neighboring, non-touching region, can you combine those two sets of knowledge to know everything about the combined area? For many years, physicists believed the answer was yes, provided the regions were far enough apart or the underlying rules were simple enough. This idea, known as disjoint additivity, suggests that the whole is simply the sum of its independent parts.

A recent study challenges this assumption in a specific but profound way. The researcher, working with a model based on a finite group of mathematical symmetries, constructed a scenario where two separate regions on a grid obey all the local rules individually and can be measured independently, yet their combined measurements fail to capture a specific observable that exists only when the two regions are considered together. This failure occurs even when the two regions are placed within a small, flat area that does not wrap around the grid, meaning the problem is not caused by the global shape of the space. The study demonstrates that for any non-Abelian group—a type of symmetry structure where the order of operations matters, much like putting on socks before shoes versus shoes before socks—this breakdown in additivity is guaranteed to happen under certain geometric conditions.

The researchers focused on a regular square torus with dimensions at least 8 by 8 and constructed counterexamples for every finite non-Abelian group. They examined a fixed eight-edge loop as a reference point, but the core of the work involves a family of peripheral supports placed around this loop, rather than the loop itself being the sole focus. They then tested whether the algebra of physical observables generated by these two separate regions could reproduce the algebra of the combined region. The results showed a clear divide: if the group governing the system was Abelian, where the order of operations does not matter, the additivity held true. However, for every non-Abelian group, the additivity failed, but only if the second region was positioned to "cover" all four inner faces of the loop in a precise way. Specifically, the failure occurred when the second region included at least one edge from each of the four pairs of edges that border the inner faces of the loop. If even one of these inner faces was left uncovered, the additivity was restored, and the separate regions could once again describe the whole.

The mechanism behind this failure is subtle and relies on how the system responds to changes. In the failing configuration, the researchers found that a specific operation could be performed on the combined system that swapped two distinct physical states without violating any local rules. However, this operation could not be built from the separate operations allowed in the two individual regions. The separate regions could each perform their own valid moves, but when combined, they lacked the ability to coordinate a specific type of transition that required a shared, non-central shift in the symmetry values. This coordination was only possible because the second region was "locked" to the first in a way that selected specific physical backgrounds, allowing a transition that would otherwise be forbidden. The study proves that the joint system possesses a type of coherence between different symmetry sectors that is invisible to the separate parts, effectively hiding a piece of the physical reality from any observer limited to the individual regions.

This finding directly contradicts a recent proposition that claimed disjoint additivity should always hold for these types of magnetic constraints when regions are vertex-disjoint. The new work shows that this proposition is incorrect for non-Abelian groups, even in simple, contractible geometries. The researchers established that the failure is not a rare anomaly but a structural feature of non-Abelian systems, occurring in a specific fraction of possible geometric arrangements. They also identified a "buffer" condition: if the second region is moved slightly further away, such that it does not touch the immediate neighbors of the loop, the additivity is restored for all groups. This suggests that the breakdown is a local phenomenon, dependent on the immediate proximity and the specific way the regions cover the inner faces of the loop.

The implications of this work are significant for how we understand the structure of quantum information in constrained systems. It reveals that the physical space of a system is not always built up from the independent physical spaces of its parts, even when those parts are far apart in terms of direct contact. The study provides a concrete counterexample where the whole is strictly larger than the sum of its parts, not because of some exotic global topology, but because of the local interplay between non-Abelian symmetries and the specific geometry of the supports. By pinpointing exactly when and why this happens, the research refines the rules of locality in quantum systems, showing that the ability to reconstruct a whole from its parts depends critically on the algebraic nature of the symmetries involved and the precise arrangement of the regions. The work leaves no doubt that for non-Abelian groups, the assumption of disjoint additivity must be abandoned in favor of a more nuanced view that accounts for these hidden correlations.