Every decade brings changes in the perceptions of normal in mental health, as well as in how abnormal is labeled, understood, and dealt with. Neurosis, hysteria, and homosexuality are just a few examples of such changes. The shifts in terminology and classifications reflect our continuous struggle with social representations and treatment of the “other.” How could we best understand mental illness categorizations and become aware of their changes over time? This presentation addresses this and other questions by applying an automated dictionary-based classification approach to the analysis of relevant research literature over time. We propose to examine the domain of mental health literature with an iterative workflow that combines large-scale data, an automated classifier, and visual analytics.