Breakthrough in Graph Theory: Bunkbed Conjecture Proven False After 40 Years

For nearly 40 years, the bunkbed conjecture sat quietly in graph theory. Proposed by physicist Pieter Kasteleyn in 1985, this hypothesis claims that the probability of traveling between two points on the ground level of a graph is greater than or equal to traveling between one point on the ground and another on a higher level.

Recently, a group of mathematicians claimed they have disproved this conjecture. Their paper, currently on arXiv and yet to be peer-reviewed, has stirred significant interest in the mathematical community.

Understanding the Bunkbed Conjecture

A graph consists of vertices connected by edges. The bunkbed graph consists of two identical graphs stacked, connected by posts. Imagine friendships in a social network as a simple analogy. The conjecture suggests that if you want to travel from point u to point v on the lower level, it should be easier than getting to point v’ on the upper level using a post.

The Disproof Journey

Disproving a conjecture can sometimes be simpler than proving it. A mathematician only needs to find one counterexample to disprove a statement. Despite the conjecture’s intuitive appeal, no one was actively searching for a counterexample until recently. Igor Pak and his team began with computer experiments on small graphs, later employing AI tools. However, they struggled to find counterexamples. Concerns arose that even if they found something, it might not definitively disprove the conjecture.

In June 2023, another paper appeared on arXiv. Authored by Lawrence Hollom, it explored hypergraphs and implicitly suggested that the bunkbed conjecture was false. Pak’s team used this work to create a graph with 7,222 vertices and 14,442 edges. They found a minute difference in probabilities, which was enough to disprove the conjecture.

Implications of the Result

This result disappoints some applied mathematicians and physicists. Had the conjecture been true, it would have supported theories about fluid movement through solids. More significantly, it raises philosophical questions about mathematical proofs. Should mathematicians trust probabilistic proofs?

Noga Alon of Princeton emphasizes the need for caution. Mathematicians should avoid accepting conjectures simply because they seem likely. The bunkbed conjecture serves as a reminder of this principle.

For more details, the paper is available on arXiv.