The concept of “justice” is at once universal or near-universal through time and societies, but the meanings given to the term are historically fluid. What is considered “just” in one society may be profoundly unjust in another. In addition, conceptualizations of justice can change dramatically within a generation.
We used the digitized proceedings of the Dutch parliament (Handelingen der Staten-Generaal) and a digital method in our approach to solve this problem. We trained Word Embedding Models (WEMs) using the Google Word2Vec algorithm to compare discursive spaces of different terms over time. By combining WEMs with Cosine Similarity measures, we developed a workflow to compare discursive spaces of words diachronically.