AI in EvoLang

Since the last few editions of the Evolang conference, artificial intelligence (AI) has had an increasing impact on scientific research, and this has influenced research on language evolution as well. Evolang has a long tradition of using AI techniques (some of the papers at the first Evolang in 1996 were based on AI) and most recently, this was reflected by a workshop on machine learning at the Kanazawa conference (2022) and by a workshop on large-scale computational approaches at the Madison conference (2024). Since this time, AI methods have become even easier to use.
AI techniques can for instance be used in the form of large language models in corpus analysis, as chatbots to interact with human participants in experiments, as part of agent-based models in simulation and as tools for sound- and image analysis when analyzing animal behavior. Different applications require different techniques, and perhaps more subtly, have different constraints on how existing AI models can be applied without introducing biases and artefacts. Exactly because modern AI tools are so easy to use, there is a definite risk that they may be applied incorrectly. However, there are as yet no textbooks that help researchers decide on how to use AI techniques appropriately for their research.
This workshop therefore proposes to bring together recent and ongoing research that uses AI techniques. The focus of the workshop is methodological: it is not so much about the linguistic/evolutionary questions that the work addresses but is meant to help exchange techniques and ideas on how to use AI techniques productively and correctly in evolution of language research.

Submissions

All deadlines have passed and the programme has been finalized. Prospective participants are encouraged to join the evolang mailing list to receive emails about practical information.

Practical Information

The workshop will be organized in conjunction with Evolang 2026 in Plovdiv, on the 10^th of April from 9:00 to 13:30. For travel and other practical information, there are sections on the main conference web site.

Venue

Ramada Trimontium Hotel. Information about the precise room will be added once it becomes available.

Programme

9:00 Opening

9:05 Invited lecture: Willem Zuidema
Learnability and evolvability of grammar: What the AI revolution teaches us about language evolution

9:35 Axel Ekström & Runhui Song
Learning from a hundred thousand tube models: Exploring the role of vocal tract size in shaping phonetic output

9:55 Rodrigo Manríquez, Sonja Kotz, Andrea Ravignani & Bart de Boer
Spiking Neural Networks trained on Acoustic Data

10:15 Break

10:30 Andres Karjus, Sophie-Marie Ertelt & Muhammad Okky Ibrohim
Evolutionary idea dynamics: drift and diffusion in LLM-annotated historical corpora

10:50 Bach Phan-Tat, Kris Heylen, Dirk Geeraerts, Stefano De Pascale & Dirk Speelman
From Parsers to Prompts: Combining AI and Linguistic for Interpretable Language Evolution

11:10 Jonas Gillain, Katrien Beuls & Paul Van Eecke
Exploring Open-ended Neural Concept Representation for Language Emergence

11:30 Nathaniel Imel & Noga Zaslavsky
Evolution and compression in LLMs: On the emergence of human-aligned categorization

11:50 Break

12:05 Anna Jon-And & Jérôme Michaud
Modeling Cognition with Minimal AI

12:25 Yuqing Zhang, Ecesu Ürker, Tessa Verhoef, Gemma Boleda & Arianna Bisazza
NeLLCom-Lex: A Neural-agent Framework to Study and Simulate Lexical Semantic Change

12:45 Roman Miletitch & Limor Raviv
Swarm Robotics: Embodiment and Spatial Constraints in Emergent Communication Models

13:05 - 13:30 Discussion

Proceedings

Extended abstracts have been bundled in this pdf. Possibilities of publishing a special issue for the Journal of Language Evolution will be discussed at the workshop.