With the rise of the digital humanities has come the promise of new methods of exploring literary texts on an unprecedented scale. How does our approach to literature and literary history change when the canon expands to include millions of texts—all of them immediately analyzable by cutting-edge methods? What previously undetectable trends, long-term shifts, and patterns within and across cultures are revealed when we study texts at this scale? Can software help us transcend language barriers to enable a truly global perspective for comparative literature? In recent years a diverse group of scholars at the University have taken up these research questions, exploring the possibilities and challenges that digital technology has introduced to the field of literary studies. The Textual Optics project brought these scholars together in a lab-like environment to consolidate and expand the scope of their work. The goal was to create a permanent home at the University for scholars pioneering this new approach to literary research.

The project centered on the concepts and practices associated with new scalable reading methods, many of which are imported from the sciences and enabled by recent software innovations—everything from data mining and visualization to machine learning and network analysis. The research team employed a set of tools and interpretive methods that allowed them to read textual archives through multiple lenses and scales of analysis, from single words up to millions of volumes. In particular, the project considered how readers might move between close and distant readings of texts, alternating from a qualitative mode that involves traditional exegesis to a quantitative mode that involves extracting statistically significant patterns from huge amounts of data.

To date, this shift between modes has not been well understood. The most sophisticated software tools in this area focus more on the distant reading aspect of text analysis, and therefore create a disconnect between the results of the algorithmic processing of texts and the texts themselves. The abstraction afforded by distant reading risks erasing the particularities of cultural and intellectual production that humanists value so greatly. By integrating tools that enable high-level pattern detection in large corpora with more traditional philological methods of text analysis, the research team developed a system of scalability that allowed for both close reading and distant reading within a unified digital workspace. Textual Optics ultimately aimed to demonstrate the value and extraordinary potential of literary scholarship at the intersection of computation and humanistic inquiry.

The potential of Textual Optics as a method largely depends on three components: a technical infrastructure that can facilitate the movement between distant and close reading; access to a wide range of corpora; and opportunities for scholars to work collaboratively on projects that cross linguistic and cultural domains. During the course of this project, the research team leveraged extensive technical experience in conceiving, developing, implementing, and using high-performance text analysis tools. Participants produced a set of technical interfaces and models for digital literary study that expand access to this kind of work for humanities scholars and facilitate a critical awareness of where these methods stand in relation to other analytic frameworks. By creating a massive collection of digitized texts spanning multiple cultural and linguistic contexts and developing the tools and methods to extract insights from the data, the Textual Optics Lab aspires to become a global destination for digital humanities scholars.