Converted from an OASIS Open Document
In his 2003 article, P.M.S. Hacker confronted Thomas Nagel’s reasoning in “What Is It Like to Be a Bat?” (Nagel, 1974). Nagel had argued that consciousness has a unique individual subjective quality that defies reduction to the materiality of body and mind. Hacker’s refutation of Nagel’s reasoning is almost exclusively based on the observation that Nagel’s argumentation in key places is syntactically faulty (Hacker, 2003:170). For Hacker—a Wittgensteinian philosopher of language—erroneous syntax implies faulty logic and thus incorrect argument. His counter-argument implicitly reduces consciousness to that which can be understood and explained linguistically. But this move ignores thereby the evidence and arguments put forward in fields beyond linguistics and philosophy that speak against the idea of consciousness as a solely linguistic construct: empirical evidence from the neurosciences (cf. Koch, Massimini, Boly, and Tononi, 2016), theories on consciousness that relate consciousness to mind and body such as came forward from integrated information theory (cf. Tononi, 2012), and ideas surrounding embodied cognition (cf. Shapiro, 2012). Sources from many disciplines thus seem to suggest we cannot understand consciousness solely as a linguistic construct, or one that must be explained exclusively by linguistic means. A concrete understanding of consciousness is likely to integrate types of knowledge and experiences that are distinctly non-linguistic and hard to properly linguistically express.
Arguably then, mind and consciousness are better understood as emergent properties of a body orchestrating multiple types of sensations and information. Explaining consciousness through linguistic properties alone would be akin to trying to explain plastic arts by words solely. As early 20th century artist Vernon Blake (1926) put it in his phenomenological treatise on art: “By means of words one cannot hope to attain to precise transmission of ‘plastic’ thoughts. The plastic arts exist because they are the natural and only way to transmit that species of thought from one human being to another. They necessarily deal in thought factors which are inexpressible in words.”
We observe a parallel in digital textual scholarship, which is the field’s preoccupation with the digital mimetic reproduction of printed text. This is the prevalent paradigm (comprehensively described in e.g. Pierazzo, 2013), and most digital scholarly editions that have been produced follow it meticulously. This mimetic philosophy has historically been expounded through the socio-technical system known as TEI-XML and its supporting community. The vocabulary and syntax of the TEI is geared exclusively towards the structural, documentary, and content description of non-digital texts. Its philosophy and technology confine text to an existence as a book or as a digital reproduction thereof; itsvocabulary and syntax treat textuality only by means of the non-digital idiom of print, manuscript, score, script et cetera. The application of this idiom consequently leads to the reification of text as that which was contained in a document.
Just as it is an illusion to think that consciousness can be explained entirely by means of linguistics, so it is an illusion to think that text and a full understanding of it can be achieved by digital means that reduce the text to a print paradigm. Barring some eccentric exceptions there are, to our knowledge, no scholarly editors that hold that only what is on the page is the text, that the text exists solely as a semiotic representation. It is in reading that the text becomes to exist. Umberto Eco (1981) regarded a written text as a series of instructions for the reader to create meaning. The meaning of a text is a co-creation that emerges from the interaction between the text and the situated knowledge and embodied cognition of the reader. The uniqueness of this interaction—that stems from two uniquely situated sites of knowledge, namely the text and the reader, mingling—gives rise to the indeterminacy Jerome McGann (1991) calls the textual condition.
Like consciousness cannot be reduced to linguistics, textuality cannot be reduced to documentary description. We may never be able to experience the uniqueness of another reader’s mind. Consequently we may never be able to fully explain what text is. But the salient point is that we can apply digital technology to understand text beyond its being as a sign on the page. We contend that digital technologies can (and should) be used to explore the properties of these different beings of text. Analogous to what Alan Kay (1997, 2007) said about the computer revolution, we may perhaps say that the digital text has not happened yet. The medium which differentiates the digital from linear, non-interactive text is still in the process of being imagined.
So we textual scholars must become more imaginative, and ask ourselves: is there anything it is like to be a text? For us this question serves to displace an understanding of digital text as an inanimate series of electronic signs with an understanding of text as a more dynamic and interactive agent. Being a text in a digital environment involves formal complexities and computational procedural elements that are alien to print or manuscript text. Furthermore, being digital allows a text to transgress the boundaries of the document and to connect to other texts and information in ways that reify connections usually only present in human cognition. Lastly there is an element of code to the existence of text in the digital environment. Software code is text’s animated computational twin. Text will often interact with code, to the point of symbiosis. Our use cases demonstrate these properties of digital text.
Our first use case focuses on Codex (Neill and Kuczera, 2019), a system for combining text annotation with graph database networks. A standoff property text editor is a central tool allowing users to create multiple overlapping annotations (avoiding ‘OHCO impedance’). The graph network can be constructed out of entities derived from within texts as well as managed independently of the texts themselves. The aspiration of Codex is not so much the reproduction of a manuscript within a document (like a single cell), but the emergence of a multi-cellular network of texts linked by the connective tissue of the graph database. The goal of Codex is to enable scholars to create a kind of ‘meta-text’ formed from out of the relations existing within texts across the corpus.
Our second use case focuses on the creation of a digital critical edition as a computational model of the text and its witnesses, where a documentary model of edition was considered insufficient to represent not only the individual texts, but the relationship between the text witnesses themselves as well as to the information the text carries. The process of transcription, transformation to valid TEI-XML, collation of the chosen layers of witness text, and analysis of the variation is done with available tools, such as tpen2tei (Andrews, Veigl, and Kaufmann, 2018) and CollateX (Dekker and Middell, 2011), that are as general-purpose as possible. Even so, the editors had to write custom code ‘plugins’ in order to facilitate this process, for example to specify a normalization string for collation, to handle the expansion of abbreviations, or to convert certain elements of TEI markup to HTML display. We consider that this custom software code, insofar as it carries an editor’s interpretation of the text and the significance of its parts, is itself part of the edition that is produced, and thus part of the edited text.
Our third use case considers the object-oriented modeling of a text using a general-purpose computer language. The objective in this case is to demonstrate that adequate transcription as the scholarly community has performed it up to now does not necessarily encompass adequate re-mediation of that text. Computer languages facilitate the meticulous modeling not only of the linguistic layer of a text, but also of many dimensions of the text beyond that. We may actually program a digital scholarly edition so that, when the code is run, the result is the scholarly edition. But we may also model the objects, events, and relations that exist in the narrative world of the text, in which case running the code becomes a re-enactment of the text’s narrative. In this way digital technology calls into question the limits of transcription and how digital transcription may add to understanding in novel ways.
These use cases show that to imagine what it is like to be a digital text invites us to reconceive of text in a specifically digital fashion that points to affordances we fail to see and use were we to regard digital text as a mere mimesis of print and manuscript texts. This, we contend, adds to our understanding of what being a text is, and ultimately to understanding what text is.