Gérard Genette’s theory of the paratext is usually applied to some form of manuscript or printed material, but why not digital material? In this post I want to explore how we might think about paratextuality in the relation to Eighteenth-Century Collections Online (ECCO). Briefly, ECCO is an online database published by Gale-Cengage. First published in 2003, it gives access via subscribing libraries to 184,536 titles of material printed between 1700 and 1800, comprising the searchable text and digital page images. It is currently accessible via four different user interfaces (UIs):
- Gale’s original, standalone, interface (2003)
- Gale’s two cross-collection platforms, Gale Primary Sources (2016), and Gale Digital Scholar Lab (2019)
- JISC’s platform Historical Texts (UK only, 2014)
The UI has obvious affinities with Genette’s notion of paratext as a space wherein reader and text interact: ‘a zone not just of transition, but of transaction; the privileged site of a pragmatics and of a strategy, of an action on the public in the service, well or badly understood and accomplished, of a better reception of the text and a more pertinent reading.’ The paratext is a technology that aims to aid the apprehension of the text, but significantly, Genette allows for paratexts to be misread – readers are not passive to the paratext – nor are the intentions of the author (or printer or publisher) always successful. It’s just as useful to consider these same hits and misses to apply to the interface. As Johanna Drucker notes, ‘[i]nterface is not a thing’, it is rather ‘a dynamic space of relations’; she argues that ‘[g]raphical features organize a field of visual information, but the activity of reading follows other tendencies.’ How ECCO’s interfaces embody the presence – to borrow Genette’s term – of eighteenth-century books depends on the dynamics between our acts of reading and navigation, but also upon the intentions of Gale’s designers and the compilers of metadata.
ECCO’s UIs privilege page images, although all but Gale’s original interface also enable the reader to open a pane to display the OCR’d text (could we still call this a paratext? Even if it is literally presented parallel to the book images, it is the text). In addition, each UI surrounds the page image with various reader-controlled windows for displaying information about these books – in other words, their metadata. This includes information such as author, title, publisher, date of publication, printer, or other ‘peritexts’ gathered from the text itself. But it also includes ‘epitexts’ ingested by Gale and created by the ESTC, such as a list of holding libraries, format, variants, pagination, errata, Library of Congress subject headings, and ESTC ID number. Taken together this metadata is largely bibliographical – it concerns the book itself. Metadata also includes information created by Gale, such as a document IDs, citation information, and their own subject category.
However, each UI presents metadata in different ways and to varying degrees of completeness. In the standalone interface you’ll see some limited metadata presented at the bottom of the screen: full bibliographical metadata is only available when the reader clicks to open a completely different window, replacing the page image view. Is this new window a paratext at all?
In JISC’s Historical Texts book viewer, readers have a wide choice of paratexts (or none at all): I’ve opted to display thumbnails on the left and bibliographical metadata on the right:
The metadata, however, is partial: presently JISC have chosen not to display the source copy and holding libraries, nor Library of Congress subject headings; aside from reproducing Gale’s subject category (‘Social Sciences’), there is no other metadata added by JISC, arguably obscuring some aspects of the book’s material existence.
In Gale’s newer platforms, there is a similar facility for the reader to control the paratextuality of each book. In Gale Primary Sources, the reader can choose to prioritise page images or text, or have the page image surrounded by text, a table of contents, an ‘explore’ window (a facility to search the text of the book), or bibliographical metadata (‘full citation’). Curiously, metadata is split between two different paratextual views. The first, in the ‘explore’ view, contains a link that opens up a pop-up window listing the holding libraries, a brief note about pagination, and what to many must be an obscure reference, ‘Moore, 68.’
There’s also an option to see the Essay alongside an expanded set of metadata (‘View Full Citation’):
It’s not clear why different facets of the book’s metadata is split across separate windows. Nor is clear for whom this selection of metadata is intended: the Gale ID seems to look towards internal systems; some of the bibliographical metadata requires specialised knowledge (e.g. ‘Moore, 68’), while some is suppressed (no LoC subject headings); and the document type ‘monograph’ might be of utility to a library cataloguer but likely not a scholar.
Finally, the way in which the Digital Scholar Lab presents a book’s metadata is again different:
It only includes metadata of author, title, imprint; and Gale’s document ID, a link path to ECCO, and date accessed. All material traces of the book – the bibliographical metadata accumulated by the ESTC – have not been included.
Genette theorized that the paratext accompanies the text ‘to make it present, to assure its presence in the world, its “reception” and its consumption, in the form, nowadays at least, of a book.’ Much the same can be said of the aims of a bibliographical record, which is to assure the presence of a book’s material life – ‘its ancestry and its offspring’, as the early editors of the 18thC STC put it. The interfaces to ECCO arguably accomplish this presence, this service to the reader’s apprehension of the book, but to varying degrees of success.
Digital archives and their interfaces are neither static artefacts nor transparent. Each interface was developed against a different set of needs by its users and changes in scholarship. The original interface was designed specifically and solely for that collection. But JISC’s Historical Texts, Gale Primary Sources, and Gale’s Digital Scholar Lab are designed to offer libraries access to cross-searchable packages of different digital collections, each of which were created differently and with different metadata. Standardising how metadata was presented in the paratexts of these platforms was an understandable and challenging priority for JISC and Gale. Moreover, the age of quantitative analysis has shifted the priorities of how a book’s ‘presence in the world’ is to be assured and ‘consumed’ – the decreasing coherence and detail of bibliographical metadata and an increased emphasis on analysing the text in these platforms speak to these shifting priorities in the paratextuality of ECCO.
– – – – – – – – – – – –
Stephen H. Gregg is a lecturer at Bath Spa University, specialising in eighteenth-century literature and digital book history. A History of Eighteenth Century Collections Online is forthcoming with Cambridge University Press.
 For example, Dorothee Birke and Birte Christ, ‘Paratext and Digitized Narrative: Mapping the Field’, Narrative, 21.1 (2013), which examines DVDs, e-books, and born-digital narratives.
 Gérard Genette, ‘Introduction to the Paratext’, trans. by Marie Maclean, New Literary History, 22: 2 (1991), 261–72 (261-62).
 Johanna Drucker, ‘Humanities Approaches to Interface Theory’, Culture Machine, 12 (2011), 3, 6 <https://culturemachine.net/the-digital-humanities-beyond-computing/> [accessed 16 February 2020].
 There’s no space here, but paratextuality might also extend to the underlying data files, written in XML (eXtensible Markup Language).
 Text produced from image files by Optical Character Recognition software.
 Genette, ‘Introduction to the Paratext’, pp. 263-64.
 Gale packaged ECCO into different subject categories such as ‘Literature and Language’ or ‘Fine Arts and Antiquities,’ replicating the microfilm packages upon which it is based.
 Thanks to Catherine Nygren for the ‘Full Citation’ screenshot.
 This is a remnant of the methodology behind the ESTC: where possible a reference to a scholarly bibliographical study was included; in this case, J. R. Moore’s 1960 A Checklist of the Writings of Daniel Defoe, p.58.
 Genette, ‘Introduction to the Paratext’, p.261.
 Robin Alston and M. J. Janetta, Bibliography, Machine-Readable Cataloguing, and the ESTC (London: The British Library, 1978), p.29.