A natural language processing approach to complexity assessment of 18th-century health literature

Abstract

In this paper, we present an experiment for complexity-level analysis of Portuguese texts from the 18th century using NLP tools. The 18th century was the time for the realization of a new world that had been built since the Renaissance, it was the period of consolidation of many of the current sciences. One of its characteristics is the presentation of scientific written records in national languages, rather than Latin, and the expressed wishes that the specialized texts could be more understandable to people of lesser erudition. As such, we intend to collaborate to identify if and how these wishes were fulfilled. To achieve this goal, we resort to an NLP supporting methodology to detect degrees of complexity of a medical work of this time period, and compare it with two other works that have hypothesized lesser and greater complexities. By using NILC-Metrix, we intend to identify features of a continuum of complexity in this kind of documents.