The aim of this contribution is to discuss the relational or network turn in Digital Humanities, in particular on matters of literary fiction and narrative discourse. For this purpose, after first reviewing Bruno Latour’s theories of hybridization, composition, translation and mobility/mutability (Law & Mol 1995), we are going to focus on certain recent approaches which study relational patterns and network links/connections within large groups of textual objects. What is important to spotlight is that, in Latour’s understanding, such literary objects (or in general, objects of digital humanistic inquiry) are objects of an abstract composed materiality embedded within an evolving cultural and sociotechnical environment. However, the process of debunking, decoding or revealing such a digital materiality is not a gesture to access a (or the) world beyond appearances of words. By contrast, it is an assembling process that involves the composition (through the immanent reality) of these objects as “objects of concern.” Thus, the conditions for the computational enactment of any attempt to model relational or network immanence in literary text requires that the digital-humanist researcher takes a series of decisions about the analytical methodologies to follow. In this respect, as it was argued by Michael Witmore, one might distinguish three required conditions on which several computational options might bifurcate in the process of analysis: massive addressability, feature explicitness and corpus closure.
Based on this context, we are going to discuss the orientation and the methodological positioning of our approach for a network analysis of either distinct literary texts or even “big data” of materials of interest to Digital Humanities. Actually, our discussion is going to elaborate on the underlying theoretical presumptions of our computational experiments of extracting social networks from literary text (to be presented at a separate workshop).