What Computers Can't Read: Computational Graphology and Literary Manuscripts
The interpretation of manuscripts remains important in literary scholarship, but today’s software cannot reliably decipher handwriting, let alone detect the nuances that help critics date and authenticate manuscripts, assess an author's mood or health, and prepare new editions for publication.
Existing digital-humanities tools for analysis of visual artifacts such as paintings and advertisements do not support the interpretation of handwriting. This project uses forensics software, a pen-wielding robot, and other equipment to read manuscripts from Emily Dickinson and Edgar Allan Poe to recent Instagram poetry. It situates these methods within an interdisciplinary history of graphology to show how technological challenges help us to rethink the value of seeing literary manuscripts in the first place.
Seth Perlow teaches American literature and media studies at Georgetown University. He is the author of The Poem Electric: Technology and the American Lyric (Minnesota, 2018) and editor of Gertrude Stein's Tender Buttons: The Corrected Centennial Edition (City Lights, 2014). His writing about poetry and public culture has appeared in the Washington Post, the Los Angeles Review of Books, Public Books, and elsewhere.