Abstract
This study addresses how societal and linguistic changes can be detected using historical corpora, with the topics of poverty and industrial revolution as a case study, based on large historical corpora, in particular EEBO, and CLMET3.0. The results, based on a rich array of state-of-the art statistical approaches (such as kernel density estimation), show how poverty, industrial revolution, and urbanization are associated through, for instance, the associations of war, religion, family, poverty, and suffering. The study also discusses the importance of data size and cleanness, the temptations of distant reading and the necessity for validating the discovered patterns in close reading and distant reading in interaction.