One of these theories comes directly out of the inherent structure of a markup language such as SGML (discussed in the Markup Languages and Text Modelling section), and its derivative languages, including HTML and XML.  SGML structures data hierarchically.  In other words, elements must be nested within one another like a series of Russian dolls. Moreover, there are rules languages  (Schema and DTDs, discussed in the Modelling with XML section that follows) that govern for example, where, how frequently, and if  elements can be used one or more times within another element.

Consequently a meta-language like XML or SGML enforces a particular way of thinking about texts. In the late 1980s and into the 1990s a group of scholars including Allen Renear, Elli Mylonas, and David Durand developed a theory for the practice of text encoding specifically, and text more generally.  It was called an ‘Ordered Hierarchy of Content Objects’ (OHCO). They proposed that although different kinds of texts may be organised differently, they still have a hierarchical organisation. In their first iteration of the OCHO model, they asserted

Within the lowest level subsections are objects like paragraphs, sentences, prose quotations, verse quotations, equations, proofs, theorems, and so on. Many of these objects can be decomposed further. This structure is hierarchical because these objects “nest” inside one another like Chinese boxes. It is ordered because there is a linear relationship to objects – for any two objects within a book one object comes before the other. (Renear)

These nesting objects are more familiar, in books, for example, chapters, sections, paragraphs, lists, and so forth. Like the Russian doll example, these content objects fit neatly into one another, from the smallest (a letter or a word) to the largest (a book or monograph), with a myriad of other nested units in between (sentences, paragraphs, chapters, sections, etc).

While this theory was instrumental in the development of the TEI, it never fully accounted for the problem of overlapping hierarchies. Overlapping hierarchies breaks the neatly nesting pattern described above. For example, a metaphor in a poem may cut across two or more lines (marked by the tag <l>). It might seem like a purely technical issue that a language like XML requires one element to close before another opens, as in the following:

• <l>text text text</l>
• <l>text text <metaphor> text text</metaphor></l>

as opposed to the following:

• <l>text text <metaphor> text text</l>
• <l>text text </metaphor></l>.

The creators of the OHCO theory concede that this may be more than a technical issue and that it may point to some of the thorniest issues surrounding text encoding as an intellectual endeavor. Textual editors such as Jerome McGann, D.F. McKenzie and Peter Schillingsburg had been developing different understandings of text. They found the OHCO model was too simplistic a model to represent complex works, particularly literary. Essentially they argued that texts can have multiple structures and diverse meanings, some outside the text itself. These vying approaches to textual editing cumulated  in  ‘mock confrontation’ between Renear and McGann at the Digital Humanities conference at the University of Virginia in 1999. McGann argued for the complex and overlapping structures of texts using poetry as an example:

poetry is not organized in a determinate hierarchy. TEI and SGML markup, therefore, while reasonably adequate vehicles for expository and informational texts, fails to render those features of poetic text that are most salient for its makers and users. Poetical texts are recursive structures built out of complex networks of repetition and variation. No poem can exist without systems of ‘overlapping structures (Hockey)

For example, metaphors may span many lines or stanzas of verse. Narrative events may span many paragraphs and indeed may overlap. Verse drama contains dialogue lines (speeches), metrical lines, and sentences. But these sentences and metrical lines overlap in the case of enjambment or when a character begins talking and another interrupts (Renear). All these hierarchies have equal claim to representation.

Thus the theory of text that SGML (and hence XML and TEI)most eloquently expresses is what one might term the editorial or bibliographic; that is, representing the text in terms of sentences, paragraphs, chapters, front and back matter, and so on. This is not surprising given SGML’s roots as a language written to publish documentary texts in electronic form. From this point of view, one might deduce that the documentary view of text can be read as its only structure.

While Renear and his colleagues  argued for the OHCO model through the the mid-2000s in a series of articles that further refined and developed their original OCHO proposition, within the TEI community, most academics favour McGann’s view (while necessarily having to conform to an OCHO model as that is what the language most natively facilitates) in which texts encoding is approached from the perspective of complex text structures that do not necessarily all fit into a rigid hierarchy.

 

Further Reading

Hockey, Susan, Alan Renear, Jerome McGann. 'Panel: What is text? A Debate on the Philosophical and Epistemological Nature of Text in the Light of Humanities Computing Research. Web. 29 August 2016.

McGann, Jerome. Radiant Textuality: Literature After the World Wide Web. London: Palgrave, 2001.

Renear, Allen, Elli Mylonas, and David Durand. “Refining Our Notion of What Text Really Is: The Problem of Overlapping Hierarchies.” N.p., 6 Jan. 1993. Web. 17 Sept. 2012.

Scholen, David and Sandra Scholen. 'Beyond Gutenberg: Transcending the Document Paradigm in Digital Humanities'. Digital Humanities Quarterly. Vol. 8 No. 4. 2014.

Shillingsburg, Peter L. Scholarly Editing in the Computer Age: Theory and Practice. Ann Arbor: University of Michigan Press, 1996.