Grant-in-Aid for Transformative Research Areas (B)Foundation of In Silico Linguistics
via Large Language Models

My name is Yoshifumi Kawasaki, and I serve as the Principal Investigator of this research area. From FY2026 to FY2028, we will conduct the project “Foundation of In Silico Linguistics via Large Language Models” with support from the Grant-in-Aid for Scientific Research (KAKENHI) Transformative Research Areas (B) program.

Large Language Models (LLMs), such as ChatGPT, have attracted significant attention as artificial intelligence systems capable of understanding and generating language as if they were human—and sometimes with speed and accuracy that surpass human performance. In this research area, In Silico Linguistics, we regard LLMs as a new experimental environment for linguistic research, and we seek to leverage them to their fullest potential. By adopting an in silico approach that treats the “language” simulated within LLMs as the object of investigation, we explore questions that are difficult to examine directly in the real world.

Using LLMs as a common research platform and integrating insights and methodologies from Linguistics, Natural Language Processing, and Neuroscience, this project aims to develop new approaches to fundamental questions: What is language, and how do humans understand and use it?