Original title: The rise and evolution of a referential code in populations of bee-like agents
Grzegorz Chrupała · August 2026
Animals need shared rules to understand one another. If a signal changes without the listener changing too, communication breaks down. Honeybees face this puzzle. Bees on flat, horizontal combs point straight toward food during their dance. Bees on upright, vertical combs cannot point directly, so they translate directions relative to gravity and the sun. Researchers simulated populations of bee-like agents to see how both systems evolve under colony-level selection. First, direct pointing readily evolves when food is moderately hard to find by random searching, whether sites are few and large or many and small. It fails if food is too rare or too easy to find. Next, adding an external bonus for vertical combs lets bees switch to the gravity code reliably without miscommunication, driven mainly by the mutation rate and the size of that bonus.
Imagine you and a friend are trying to meet at a restaurant. You agree to use landmarks to guide each other. At first, you can point directly to the restaurant because you can see it. But then you move to a city with tall buildings blocking the view, so you have to use the direction of the sun instead. This change requires both of you to learn the new system at the same time, or you'll get lost. In this paper, the bees face the same problem: they must switch from pointing directly at food to using gravity as a reference, and the model shows how they can do it without losing communication.
This matters because it explains how a complex communication system like the honeybee waggle dance can change its coding scheme without losing function. It shows that evolution can find a path from direct pointing to gravity-referenced codes, which helps us understand the origins of symbolic communication in nature. The findings also suggest conditions under which such transitions are possible, relevant to any system where communication codes must be coordinated.
The abstract does not specify limitations explicitly, but the model is a simplified representation of bee communication. It assumes specific conditions for the evolution of codes and does not account for learning, individual variation, or other environmental factors. The transition relies on an exogenous benefit for vertical combs, which may not always exist in nature. The results are parameter-dependent and may not generalize to all communication systems.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
The rise and evolution of a referential code in populations of bee-like agents — Grzegorz Chrupała
Abstract licence: CC0
Explanation generated by Synthesis from the abstract — it is not the original paper. This is a preprint, so it hasn't necessarily been peer reviewed.