Problems with Digital Scholarly Editions

There are two major problems with digital scholarly editions that threaten to reduce the relevance of many  current digital publishing projects for future humanities research and one day turn them into a curio of the  early web: the absence of machine learning techniques designed to aid curation and analysis and editing  practices which preclude born-digital sources. For example, it is reasonable to expect that future historians  will want a critically curated edition of the former President Donald Trump’s tweets, contextualised using the  sea of online political, media and social discourse that his messages either responded to or prompted. But  how would such an edition be created, annotated, presented, used, and sustained? 

The majority of text-centred activities in the digital humanities (DH) fall into several broad categories:  the assemblage of textual resources (digital scholarly editing), the sharing of such resources (digital  publishing), and the use of computer-assisted techniques to quantitatively analyse cultural materials (distant  reading). As Katherine Bode puts it, there are separate “curatorial and statistical” dimensions to DH (Bode, 2019). Each of  these practices is essential to the arts and humanities: scholarly editing brings the perspective and care of  established experts to bear on preserving and shaping cultural records, publishing turns such knowledge into  usable material, while the computer-assisted analysis and interpretation of such materials allows for the  interrogation of ourselves and our societies using novel methods. That these practices continue to develop  in isolation has only slowed the progress of each: digital scholarly editions would benefit greatly from  integration with data mining techniques and machine-assisted insights, digital publishing needs to free itself  from its obsession with the codex form, and methods for computer-assisted analysis are only as good as the  data on which they operate. Furthermore, we live in an age where many of the cultural records of note are born-digital: public  discourse, political exchanges, and creative expression are all shifting from print to digital, to social media  platforms and virtual spaces. Reflecting on the future of digital scholarly editing, Susan Schreibman notes  that, in the present moment where born-digital materials are “more frequently the literary or cultural  artifact”, scholarly editing can no longer be solely or typically “about migrating the analogue into the digital  or about re-presenting print norms in digital format” (Schreibman, 2013). And yet, that future is upon us, and there is still no  reproducible, low-cost and sustainable model or platform for the production and publication of those born digital materials so central to culture, society and expression in the twentieth and twenty-first centuries,  editions which utilise and capitalise on machine-assisted insights in their production and use.