Scalable Scholarship

Simply put, scalable scholarship entails the study of objects across varying levels of scale. It encourages scholars to think of what a corpus may look like when writ large. Rather than just examining the relationships between an author’s texts, what happens when we broaden the purview of inquiry to how an author’s text relates to the print output from the year of its publication? More generally, what would it look like to investigate connections within a corpus of 100, 1,000, or even 10,000 texts?

To explore such a sizable corpus requires an interdisciplinary approach that combines humanist and computational approaches. Scalable scholarship synthesizes the desirable elements of close reading and distance reading through automation. Close reading, a humanist method of inquiry, involves deliberate, attentive reading at multiple levels of a text with the goal to interpret how its ideas unfold. Close reading detects patterns at different textual levels, such as syntax, imagery, and structure. Pattern-detection fuels computational text analysis. Algorithms are coded to detect patterns for our inspection. Texts can be thought of as data upon which we perform abstraction, formalization, and statistical inference. Computational text analysis can provide patterns between documents in a corpus to evaluate through close reading. Though the automated reading of computational text analysis cannot replace close reading, it can supplement close reading. Scalable scholarship capitalizes on the similarities of humanist and computational methods.

Scalable scholarship furthers textual analysis by automating pattern identification at a larger scale to augment human reading capacities. In fact, it enables multi-scale research, with connections to be drawn between a constructed corpus and one of its texts, and many intermediate scales in between. These scales facilitated by computational methods can bring a multitude of texts into conversation with current scholarship. Close reading has been applied to a relatively small number of exemplary texts when considering the vastness of English literature. The scale of texts available simply dwarfs what scholars are capable of reading. Computational methods can bring a multitude of texts into conversation with current scholarship and the fraction of exemplary texts, challenging our assumptions of cultural and intellectual history.