Springe direkt zu Inhalt


A digital tool to identify free verse prosody

Funded by the Volkswagen Foundation in the Mixed Methods in the Humanities programme; 2017-2020.

At least 80% of modern and postmodern poems have neither a rhyme nor a fixed meter such as Jambus or Trochäus. But does this mean that they are completely free of rhythmic structures? The American theory of free verrhythmics claims the opposite: modern poets like Walt Whitman, imagists like Ezra Pound, beat poets like Alan Ginsberg or contemporary slam poets have developed a "post-metric" verse that is shaped by prosar rhythms, everyday language or musical styles like jazz or hip hop.

Our project, which has been funded by the Volkswagen-Foundation since 1 January 2017 with a funding volume of 450,000 euros, aims to test this theory on the basis of a digital sample analysis. We are a team of computer scientists and literary scholars from both the FU Berlin and Carnegie Mellon University in Pittsburgh. Our partner is "lyrikline", probably the most important internet portal for international poetry readings worldwide, hosting contemporary international poetry as audio (read by the authors) and text (original versions & translations). Our goal is to use the very effective prosody recognition of current speech technologies in computational linguistics for a digital analysis of rhythmic patterns in this corpus of poetry readings.

The philological subproject first identifies specific rhythmic patterns of modern and postmodern German-American poetry, such as syncopes, variable feet or cadences. The digital subproject develops a digital pattern recognition based on machine or deep learning. Our goal is to develop a method and software for digital prosody recognition and formal corpus analysis of (post-)modern hearing poems.

Traditional theories of poetry used metric patterns such as the pentameter or the hexameter to identify literary genres such as elegy or literary-historical influences such as the one that Greek poetry had on 18th century German poetry. In a similar approach our analysis of rhythmic patterns will enable students, teachers and researchers to capture poetic forms of free-verse poetry, or to identify the influence of the US American free verse prosody on the German-language poetry of (post-)modernism. During the next three years, the research results of our interdisciplinary toolset are to be placed on lyrikline.org so that they can be used sustainably for university teaching and research.