Systematically uncovering transition dynamics: An STCA of the Dutch energy system using LLMs

Abstract

While transition studies have produced critical insights into how socio-technical transitions unfold, only a handful of studies have tried to systematically identify long-term patterns of transitions. This thesis addresses this gap by exploring a novel methodological approach for analyzing socio-technical transitions. This novel approach uses Socio-Technical Configuration Analysis (STCA) complemented by a Large Language Model (GPT-4o-mini), allowing automated coding of 9,867 trade journal issues. While the primary contribution of this thesis is methodological, it also provides valuable empirical insights into the Dutch energy system by analyzing changes in its socio-technical configurations. This analysis covers the period from 1880 to 1966, offering new insights into the evolution of technologies, markets, and institutions in the Dutch energy system. The findings reveal five systematic patterns in transition dynamics, including the geographical diffusion of innovation and the co-evolution of selection environments. These patterns support and extend existing theories such as the Triple Helix model, the Multi-Level Perspective, and the Deep Transition framework. By combining theoretical development and methodological innovation, this thesis advances both the methodological tools and empirical understanding needed to guide society into a more sustainable future.