One Journal, One Decade, 3,797,592 Words: Computational Analysis of HaTzfira's Discourse (1874-1883)

In this study we use computational tools to analyse the dynamic nature of the journalistic discourse as reflected in one decade (1874-83) of one Eastern European Hebrew weekly - HaTzfira. To identify latent themes in HaTzfira's discourse we apply algorithmic topic-modelling analysis to an upgraded optical character recognition of this periodical. The analysis demonstrates two turning points in the history of HaTzfira's discourse. The first, in the mid-1870s, reflected a shifting of balance from scientific topics to topics related to world politics. The second, and more significant, change occurred in 1881: world politics lost its dominance and there was a steep rise in topics relating to anti-Semitism. We argue that computational analysis helps reveal the life cycles and changing dominance of the overarching themes in HaTzfira and identify the main internal and external networks and influences that shape the journalistic text.