Evolutionary Ecology of Words

2025-03-17

2025 IEEE Symposium on Computational Intelligence in Artificial Life and Cooperative Intelligent Systems (ALIFE-CIS)

AI Summary

This study addresses the limitation of classical evolutionary game theory—its reliance on predefined strategy sets—by modeling the open-ended, unbounded evolution of lexical strategies within agent populations. We propose the first framework that deeply integrates large language models (LLMs) into evolutionary game dynamics: the LLM serves both as a dynamic semantic arbiter to evaluate interaction outcomes and as a mutation engine to generate novel lexical strategies. Coupled with spatially embedded agent-based modeling and population-level simulation, our approach enables co-evolution of lexical strategies and emergent ecological diversity. Experiments successfully reproduce canonical evolutionary patterns—including gradualism and punctuated equilibrium—and spontaneously yield environment-specialized subpopulations (e.g., terrestrial vs. marine adaptations) and multistable equilibria. Over large-scale, long-term simulations, the system sustains high-dimensional semantic diversity. Our core contribution is establishing an LLM-driven paradigm for open-ended strategy-space evolution.

Technology Category

• Machine Learning: Evolutionary Learning
• Multiagent Systems: Agent-Based Simulation and Emergent Behavior
• Natural Language Processing: (Large) Language Models

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions

Abstract

We propose a model for the evolutionary ecology of words as one attempt to extend evolutionary game theory and agent-based models by utilizing the rich linguistic expressions of Large Language Models (LLMs). Our model enables the emergence and evolution of diverse and infinite options for interactions among agents. Within the population, each agent possesses a short word (or phrase) generated by an LLM and moves within a spatial environment. When agents become adjacent, the outcome of their interaction is determined by the LLM based on the relationship between their words, with the loser's word being replaced by the winner's. Word mutations, also based on LLM outputs, may occur. We conducted preliminary experiments assuming that “strong animal species” would survive. The results showed that from an initial population consisting of well-known species, many species emerged both gradually and in a punctuated equilibrium manner. Each trial demonstrated the unique evolution of diverse populations, with one type of large species becoming dominant, such as terrestrial animals, marine life, or extinct species, which were ecologically specialized and adapted ones across diverse extreme habitats. We also conducted a long-term experiment with a large population, demonstrating the emergence and coexistence of diverse species.

Problem

• Modeling word evolution using LLMs in agent interactions
• Studying emergence of diverse species through word competition
• Exploring ecological specialization in simulated evolutionary dynamics

Innovation

• LLM-driven evolutionary game theory model
• Spatial agent interactions with word mutations
• Diverse species emergence and coexistence