Computational identification of versification metric structures in Euclides da Cunha's prose

Versification structures written in Portuguese language prose are still an unexplored phenomenon to literary theorists and historians, and the automatic mining of such structures is still unseen in Computational Linguistics. The MIVES (Mining Verse Structure) system was developed for computational scansion of metric versification structures in Portuguese language prose. It is able to identify, classify and compare, frequency, density and dispersion of heterometric structures of versification, distributed at different scales of observation, from a work or author, to historical periods and movements. We present the system, and a preliminary validation analysing three works by Euclides da Cunha (Os Sertoes, A Margem da Historia, Contrastes e Confrontos). They were selected because they constitute the main corpus of one of the most important Portuguese-language writers, and because Os Sertoes was the object of what Augusto de Campos (2010, p. 14) called "verse-spectral reading", an operation able of revealing "more than 500 decasyllables in the book, among sapphic and heroic verses, and more than two hundred dodecasyllables. MIVES identified metric structures in 48.18% of the sentences of Os Sertoes, 48.96% of A Margem da Historia and 37.48% of Contrastes e Confrontos, an unexpected rate when compared to the partial results of Augusto de Campos. It can be said that MIVES inaugurates new patterns of observation and analysis of structures not yet investigated in Portuguese language literary prose.