The Landis Conjecture For Nonlocal Elliptic Operators: Polynomial Decay

In the vast landscape of mathematical physics, there is a fundamental question about how things behave when they get very far away. Imagine a wave or a field spreading out across an infinite space. If that wave fades away quickly enough as it travels toward the horizon, does it mean the wave was never there to begin with? For decades, mathematicians have studied this idea, known as the Landis conjecture. It asks whether a solution to certain equations that disappears at an exponential rate—meaning it shrinks incredibly fast, like a light dimming to nothing in a fraction of a second—must be zero everywhere. For a long time, the answer seemed to be yes, but then a counterexample was found showing that under specific conditions, a non-zero solution could still exist even with that rapid decay. This left scientists wondering exactly how fast a solution must vanish to guarantee it is truly gone. The rules change when the equations involve "nonlocal" effects, where a point in space is influenced not just by its immediate neighbors, but by the state of the entire system at once. Understanding these long-range interactions is crucial for modeling phenomena in materials science, finance, and biology, where distant parts of a system can affect one another directly.

A team of researchers has now tackled this question for a broad class of these nonlocal equations, specifically those involving fractional derivatives, which describe systems with long-range connections. Their work reveals a surprising shift in the rules of decay. While the original conjecture for standard equations relied on exponential decay to prove a solution was trivial, the researchers found that for these nonlocal operators, the threshold is much more relaxed. They proved that if a solution decays according to a polynomial rate—meaning it fades away like a power of the distance, such as one over the distance raised to a certain power—then the solution must be zero everywhere. This is a significant discovery because polynomial decay is much slower than exponential decay; it is the difference between a light vanishing instantly versus a light dimming gradually over a long distance. The fact that a solution must be zero even when it fades this slowly highlights a unique property of nonlocal systems: their influence is so pervasive that even a slow fade is enough to force the entire system to collapse into nothingness.

The researchers focused on a specific type of mathematical operator that generalizes the fractional Laplacian, a tool used to model these long-range interactions. They considered equations where the solution is influenced by a potential, a function that can vary across space, and they assumed the solution was a "viscosity solution," a robust way of defining solutions that works even when the functions are not perfectly smooth. The team established that if the system satisfies certain maximum and minimum principles—essentially rules that prevent the solution from behaving erratically in bounded regions—and if the solution vanishes at infinity at a rate faster than a specific polynomial, then the solution is identically zero. This result holds true even without assuming that the potential itself decays, which was a major limitation in previous studies. The proof relies on a sophisticated technique involving a weak Harnack inequality, a tool that provides a lower bound on how large a positive solution must be in a region based on its values elsewhere. By adapting this tool to work over arbitrarily large distances, the authors showed that a positive solution cannot exist if it decays too fast, because the nonlocal nature of the operator would eventually force a contradiction.

One of the most striking aspects of this finding is that it changes the expected behavior of these systems from exponential to polynomial. In the context of the original Landis conjecture for standard equations, a solution had to vanish exponentially fast to be proven trivial. Here, the researchers demonstrated that for nonlocal operators, the requirement is far less stringent. They showed that a decay rate of the form one over the distance raised to the power of the dimension plus twice the order of the operator is sufficient to force the solution to be zero. This finding is new even for the fractional Laplacian, a specific and well-studied case of these operators. The authors also noted that this polynomial rate appears to be the sharpest possible limit; if the solution decays any slower than this specific rate, non-trivial solutions can exist. This suggests that the nonlocal nature of the operator fundamentally alters the landscape of unique continuation, making it easier to prove that a system is empty based on its behavior at a distance.

The work builds upon earlier attempts to solve the Landis conjecture for fractional equations, which had previously required solutions to decay exponentially or to satisfy complex integral conditions involving exponential weights. By relaxing these conditions to allow for polynomial decay, the researchers have opened the door to a wider understanding of these operators. Their proof involves constructing a positive solution that serves as a benchmark and then using comparison principles to show that any solution decaying faster than this benchmark must be zero. This approach avoids the need for the potential to have specific decay properties, making the result more general and applicable to a wider range of physical and mathematical models. The study confirms that the nonlocal character of these equations imposes a strict discipline on their solutions: they cannot linger in a state of slow decay without being forced to vanish entirely. This insight deepens the understanding of how information and influence propagate in nonlocal systems, offering a clearer picture of the conditions under which a system can be considered truly at rest.