MARGENTO: Poetry Computational Graphs

The SSHRC (Social Sciences and Humanities Research Council) awarded project Poetry Computational Graphs aims to read, categorize, classify, and analyze poetry by means of computer programming and graph theory applications. Me--Chris Tanasescu (MARGENTO)--and my team--Professor Diana Inkpen and our graduate and undergraduate students--will digitally define, compute, represent, expand, and evaluate graphs—networks of nodes and edges, where the nodes shall be the poems, and the edges, links between the poems defined according to various criteria—as computational tools to (re)organize Canadian poetry and to discover new relevant commonalities and paradigms among various poems (and therefore poetries and poets) to be used and developed upon in further poetry criticism, poetics, and Digital Humanities by the participants in this project, as well as other academics and researchers interested in such results and approaches.

The research shall discover and study by means of computational graphs new facets of North American and world poetry (in terms of themes, motifs, form, and diction) while looking deeper into already established categorizations and assessments, and developing more rigorously structured archives and critical/historical classifications, using the graphs for (and making them available for even further) trend evaluation and predicting future evolutions in North American poetry, and last but not least, finding and assessing new relevant connections and aspects of North-American poetry within global and transnational contexts.

So far, me and my team have processed over 11,000 poems from the Poetry Foundation website and the analysis so far regards subject categorization through the code written by our Research Assistant, and we have two more undergraduate students in Computer Science who received UROP fellowships to work on our project, one dealing with poetic forms and the other with the syntax of poems.

Our resulting papers and presentations (two of them forthcoming) will include critical commentaries combining the computer science and literary considerations and findings involved and will open new perspectives on poetry and North American literature and culture in general, and thus attract and encourage further similar initiatives in literary, interdisciplinary, and digital humanities writing, editing, publishing, and data(base) (re)configuration. They will also provide tools for (re)assessing poetry collections, anthologies, poets’ oeuvres, poetry schools’ publications, and poetry-related phenomena (on-line publishing and blogging, as well as popular music and hip-hop lyrics) in certain (sub)cultures, regions, or periods.

The resulting poem graphs and publications will also be later employed in studying literary communities and commonalities across linguistic, national, regional, and racial divides, and also as expressive of cohesive or centrifugal forces within cultural (Aboriginal, for instance), historical, and linguistic communities in a globalized world.

The software and computational graphs we are generating will be usable (by us and by others) for learning purposes and curriculum enhancement in areas such as creative writing, machine learning, and new media.

Outside the academic community as well, both (e-)poets and literary critics or the poetry audience in general may use our results for (e-)poetry composition or for (re)searching (through) poetry and literature for various purposes.

Close to the of the completion of the project, we shall start working on a critical North American poetry anthology assembled by means of computational graphs, which will attract the interest of writers, anthologists, publishers, and Digital Humanities scholars, setting an example for future anthologies or collective writing and publishing.

Our project shall also provide digital tools to be further used in poetry criticism and literary history, also applicable to other genres, literatures, and various areas of Digital Humanities and interdisciplinary studies. Still, all such further research, studies, and applications (within as well as outside of the academia) shall have to acknowledge and/or develop on our results in the field of North American literature and Canadian and US Digital Humanities, and therefore also acknowledge the place of Canadian literature and literary research and criticism in North-America and the world in general in a new ever-shifting and challenging global context.


Help type
Kinds of collaborators
Graduate students
Undergraduate students
Help description
Entering published mainstream and innovative contemporary North-American poetry in digital/digitized format
Help needed