‘On þe nis bute chatering’
Cyberpragmatics and the paratextual ‘anatomy’ of Twitter

Dr. Johanna M. E. Green
University of Glasgow
Follow @Codicologist

Abstract

1. Introduction

Attention to the single manuscript as historical artefact has become the subject of much recent scholarship (see, inter alia: Robinson and Zim 1997; Kelly and Thompson 2005; Minnis and Roberts 2007). This scholarship goes beyond traditional textual criticism arguing that the individual manuscript,

contextualises the text(s) it contains in specific ways [and seeks to] analyse the consequences of this relationship on the way these texts may be read and interpreted. More particularly, it postulates the possibility that a given manuscript, having been organised along certain principles, may well present its text(s) according to its own agenda […] Far from being a transparent or neutral vehicle, the codex can have a typological identity that affects the way we read and understand the texts it presents. (Nichols and Wenzel 1996: 1–2)

Thus, a manuscript can offer insight into the way these texts were read, or were intended to have been read by the patron or audience with whom it was shared. Owing to work by bibliographers and textual editors, such as McGann (2001), McKenzie (1986) and Greetham (1994), the printed text has equally emerged as a ‘cultural artefact […] fit for historical analysis [and constitutes] material evidence of the past.’ (Mak 2011: 6) The concept and mechanisms of what Gérard Genette terms paratext have been central to the studies of both these artefacts; peri- and epi-textual information found within manuscripts and printed texts – such as layout, script, spacing, punctuation, entitling, textual division and so forth – that surround and mediate a text, can be understood to act as within threshold of interpretation between reader and page. The anatomy of a given page is, therefore, central to an understanding of the text contained within the limits of that page. Further, a page acts as an interface between designer and reader on which the intended reading practices are expressed and the actual reading practices of the audience are performed (Mak 2011).

The recent rise in the number of digitisation projects and the increase in accessibility to digital images of medieval manuscripts and early printed books have brought with them a renewed awareness of the significance of the material object as the first paratext to a text. Information provided by mise-en-page, textual division, marginalia and decoration impact on an understanding of how that text was, or was intended to be negotiated by its readers. What was always meant to be seen in the material manifestation of an individual text – so often lost or obscured in modern print editions of that text – is now once more visible to the modern audience via the digital image. This visibility has brought with it a renewed awareness of the importance of the page and the paratextual mechanisms employed therein. Paratext is at once both a visual cue to aid, and a threshold to mediate an audience’s experience of that text.

This article considers the mediation of the written word on Twitter, both via the vehicle through which it is communicated (a Tweet) and the author who creates its text. Using Bonnie Mak’s definition of the page, and her analysis of the digital page as cultural artefact (2011), combined with Genette’s concept of paratext (1997), this article offers what Peter Barry (2007) terms a ‘context-saturated’ approach to Twitter: it considers the anatomy of Twitter as a new digital page, and provides a suggested analysis of the paratextual mechanisms employed therein, arguing these are a) central to mediating our reading practices of that digital page and b) more firmly rooted in the past that we might expect. In so doing, it contributes to an emerging discussion of the digital page as worthy of analysis, similar to the manuscript or printed book.

1.1 Context

In response to emerging digital editions and CD-Roms of literary works dating from the late 1990s and early 2000s, Peter Robinson laments the near-replication of the printed book in the digital page, arguing:

Scholarly electronic editions up to 2003 have rarely extended beyond the model of print technology, either in terms of product (the materials included and the ways they are accessed) or process (the means by which they are made and by which they may be manipulated) […] up to now, almost without exception, no scholarly electronic edition has presented material which could not have been presented in book form, nor indeed presented this material in a manner significantly different from that which could have been managed in print […] As for hypertext: even before print, scribes created manuscript pages which surrounded the text with all kinds of extra-textual material – commentaries, variant readings, indices, cross-references, glosses, pointers to every kind of matter. Almost all we have done, in the first ten years of electronic scholarly editions, is find ways of mimicking on screen elements long present in print and manuscript (2003: 125–6).

His lamentation, though now 10 years old, is an important one, emphasising the evolutionary links – and the limitations – that exist between the digital environment and parchment or paper. As Stoicheff and Taylor similarly argue:

Today digitization has opened up endless possibilities for visual and acoustic innovation, but our understanding of what constitutes a text remains rooted in the traditions of the medieval page. The architecture of the page has not changed significantly since then, a result of its tremendous economy and functionality […] the traditional features of the page are still highly functional because they developed along with our understanding of what constitutes information […] If web sites still tend to reproduce the features of medieval page design, they do so because these features have become fully integrated with our habits of thought and with the structures of academic publishing (Stoicheff and Taylor 2004: 11–12; emphasis mine).

The digital page is very much that: a page. However, it is important to remember that although it may ‘imitate the look and behaviour of counterparts in parchment and paper, it has its own distinct materiality’ (Mak 2011: 62). Robinson’s arguments make this point clear: in the digital environment one is not constrained by the same materiality of the analogue page; this freedom should be celebrated but, more importantly, should explored fully. In his seminal work, Radiant Textuality, Jerome McGann (2001) celebrates this evolution of the hypertext environment, which frees authors, publishers and readers from the traditional ‘constraints of the codex’; that is, those codicological limitations of the analogue page which demarcate how a written text can be conveyed to its audience. The flexibility of the hypertext environment allows us to break free of these traditional ‘constraints of the codex’, or the limitations of the page, offering the opportunity to reimagine the myriad ways in which we can navigate, interact with and mediate digital written information.

Yet despite this increased ‘flexibility’, some of the most popular methods of communicating in the digital environment bring with them strict constraints. Twitter, a micro-blogging site founded in 2006 to provide an alternative to SMS messaging, is characterised not by its digital flexibility, per se, but rather by its limitations: a single Tweet can famously be no longer than 140 characters – 20 characters less than a standard SMS. Users have the option of including hyperlinks to other sites, and images or videos if they wish, though the 140-character limit remains. Accounts may be public or private, and the user has control over accounts they choose to follow and which Tweets they see. One also has the option of reposting (or retweeting) the Tweets of other users, and favouriting Tweets, either as a means of filing Tweets for future reference or, perhaps with the influence of the ‘like’ function on social media sites such as Facebook, to indicate their agreement with, or enjoyment of a particular Tweet. The result: short bursts of information that form single, communicative acts. Each Tweet is displayed chronologically within a Timeline, running vertically from most recent (top) to oldest (base) on screen. In effect, one has a continuous digital scroll of information, chronologically divided into posts authored by individuals. Despite the character limitations imposed by Twitter, it has fast become one of the most popular micro-blogging sites available (see, inter alia, Java et al. 2007). Perhaps in part due to the semi-spoken nature of the information contained in individual Tweets, the limits on size of content seem key to its success. Brevity necessitates speed, speed assists sharing. News headlines, celebrity gossip, academic discussions can spread more easily across a worldwide digital community in a matter of minutes – nay, seconds – facilitated by the ease of use and accessibility of the Internet on PDAs and other mobile devices.

1.2 Academic Use of Social Media

For the academic, (here a book historian), Twitter is emerging as a central – possibly an essential – communicative space in which images of rare, restricted or sometimes unknown manuscripts and printed books are uploaded, posted and shared instantaneously with the community, in a manner not available elsewhere, and via which live information on scholarly events and discussions can be shared and contributed to with relative ease. One only has to consider the manuscript image Tweeted in 2012 by book historian Erik Kwakkel (University of Leiden) to comprehend how quickly information can travel using the Twitter format. Kwakkel shared an image of a manuscript decorated with the inky paw prints of a (contemporary?) cat, an image originally posted to Twitter by Emir O. Filipović (University of Sarajevo; both Tweets shown below):

The Tweet initially produced an extremely respectable response (Filipović at 196 retweets and 111 favourites; Kwakkel at 53 retweets and 32 favourites respectively), but in February 2013, the image suddenly went viral. The Tweet was brought to the attention of the public by ‘America’s Veterinarian’ Dr. Marty Becker who, having seen the image on Twitter, reposted it to the social media site Facebook; his actions resulted in 8,499 likes and 30,428 shares: [2]

The result of this popularity was reflected back on Twitter as medievalists and cat-lovers alike retweeted the image from their own accounts as interest grew. Both Kwakkel and Filipović saw a significant leap in Twitter followers, which in turn led to a marked increase in interest in their academic work. As Filipović later Tweeted: ‘[...] The cat has certainly influenced the traffic to my academia.edu page [...]’:

This sharing of information from one social media site to another and back again shows the impact and potential reach of communicating via Twitter. The site offers instant publishing, and instant access to information, images, videos, and websites of all kinds, not just academic.

Twitter is thus being established as a key facilitator enabling scholarly discourse. As Jeffrey Jerome Cohen stated on the blog In the Medieval Middle:

[…] a conference is no longer considered a closed or private space where you impart an argument in its almost-article form, just before you publish a citable (and non-dialogic) version within the cement of a paper journal. Despite the fact that someone will check your nametag at the door to ensure you’ve registered if you want to sit at session, conferences in the digital age have become networked, public forums with potentially immediate and wide impact. That is especially true in digital humanities, which embraces that flow of information in subjective and multiple forms. […] Tweeting a conference creates a more embodied space for scholarship to transpire within […] Scholarship unfolds in a world, and I like to experience what I can of that complicated unfolding. It deepens my understanding of how knowledge works, and increases the likelihood that I’ll retain what I’ve learned. (Cohen 2012)

His blog post discusses the growing trend in academia to use Twitter to communicate not simply about a particular conference event, but the detail of an individual session or conference paper. No sooner has the academic speaker made reference to an article or a new online resource than they are shared by audience members via Twitter, providing direct links to works cited as well as offering a live critique to the material being presented. [3] One could easily talk at length on the morals and professionalism of live-tweeting someone’s unpublished research, but the fact of the matter remains: to a member of the audience following discussions on Twitter, a new platform for scholarly discourse is available. [4] Almost instantaneously one can learn not only what the other audience members think on a particular presentation, but one can interact directly with them and those following the discussion online from elsewhere in the world. Truly, the age of Twitter has brought with it a new way of conferencing altogether.

This ever-growing popularity of Twitter, particularly in an academic environment, shows that the limitations imposed by an individual Tweet do not restrict communication; on the contrary, the 140-character limit is central to Twitter’s success. Information is instantaneous and succinct. It is live to the event and so one’s response to subsequent information can be relayed in real time, as if it were a traditional face-to-face conversation. [5]

1.3. Social Media and the Material Object

As a book historian, when one uses Twitter primarily for the discussion of material objects (namely, books) and views Tweets such as Kwakkel’s paw print manuscript in isolation, one is struck not only by the content and reach of the information they contain. Engaging in digital discussion of analogue material objects via Twitter also prompts one to reconsider the relationship between Work and Text (cf. Barthes 1979) as visual aids to literacy. Medievalists are exploiting the digital to analyze and communicate the medieval. However, at the same time, Twitter users are being mediated by many of the same visual cues they examine in medieval manuscripts and early printed texts. It is only when we take a step back, that we can appreciate how the digital page, here twitter, is constructed, laid out by the writer, published by the medium and, ultimately, read by the audience. By engaging in the study of book history via Twitter, one can’t help but consider how scholars are using a new kind of digital page on which to discuss the analogue page. As book historians preoccupied with the codex, how do we conceive of Twitter? When we read the information contained in any one Tweet, how are we interacting with the medium that brings us that information and in turn, how is that information mediated by the design of that medium? The analogue page, as McGann and Robinson lament, is one that is in its very nature constricted – by space, by size, by technology, by the imagination, time, financial and material resources, and the skill of the author, scribe or publisher. These limitations are reconceived by Gérard Genette (1997) as mechanisms existing within a textual threshold, mechanisms that Genette terms paratext; that is, that information which creates a boundary between written text and reader, acting as a threshold of interpretation, mediating the reading practices of a particular audience via various visual cues. Each codicological, palaeographical, and textual element in script and print has a literary function, and can be viewed as a paratextual mechanism making up the anatomy of any given codex, or page contained therein. Rather than being superfluous to textual interpretation, therefore, these bibliographical elements are central to an audience’s understanding of a written text and contain information key to navigating the reading practices intended by that codex. Within the medieval manuscript, paratext can manifest itself in many forms, from simple mise-en-page to more imaginative interactions with the text. A number of recent articles seek to analyse these mechanisms; Kisor (2006) and Fulk (2006) (inter alia) reconsider the numbered sections found in the Beowulf manuscript, for example, while earlier articles such as those by Dahood (1988) and Caie (1988) reconsider the hierarchical use of litterae notabiliores in manuscript copies of Ancrene Riwle and Corpus Christi College Cambridge MS 201 respectively. Recent work of my own (Green 2012) has identified similar hierarchical use of textual division within the Exeter Book, arguing that such divisions in a manuscript, differentiated variously by size and decoration of litterae notabiliores, for example, actually function as an aid to visual literacy, offering the reader a visual means of navigating the text in front of them. Other medieval manuscripts can be seen to employ more imaginative means of mediating their text. A recent article by Thomas White (2012) discusses the use of marginal bracketing in manuscript copies of Chaucer’s Tale of Sir Thopas as an example of paratext not found elsewhere in other of Chaucer’s Tales, used to visually link the rhyming structure of the text.

Similarly, the manuscripts BL Harley 4866 and BL Arundel 38 can be seen to employ similar paratextual means of correcting miscopied text as one another, detailing a man with a lariat pulling the text into the correct position on the page: [6]

Figure 4. BL Harley 4866 and BL Arundel 38, with thanks to @Sarah_Peverley on Twitter for bringing these images to my attention, and the British Library’s recent Open Access Policy which allows me to reproduce them here.

Each of these select examples, though separated by author, text, date, provenance and audience can all be understood to actively employ various visual means of presenting the written word to a reader, each existing within the threshold identified by Genette, and each having a direct effect on the reading practices and reading experience of the audience, both contemporary and modern.

The digital page is no different; written text found online is also bound by many of the traditional mechanisms of paratext that mediate reading practices. On the digital page, attention is drawn to specific elements of information by a variety of these paratextual features that provide the digital environment with a materiality reminiscent of the analogue page. Just as the manuscript or printed book can be considered a cultural artefact worthy of analysis, existing within a continuum of textual transmission and mouvance, the digital page is equally ‘open to critical examination; it is this materiality from which meaning is made’ (Mak 2011: 63).

What follows, therefore is a consideration of the materiality of Twitter: the reading practices involved in Twitter, the anatomy of the digital page, and the paratextual mechanisms central to the construction and cognitive mediation of the Twitter Timeline and individual Tweet. It is my contention that the paratextual mechanisms present in Twitter can be understood to be rooted more firmly in traditional codicological practices than one might expect from such a new and evolving form of social media; that is, Twitter is more bound by those traditional constraints of the codex and of the analogue page that the digital hypertext environment seeks to obviate, and that it is these mechanisms that allow effective communication across such a limited and constrained written space.

2. Methodology

In order to consider the reading practices involved in Twitter, it is first necessary to clarify a number of key concepts. Traditional investigations of paratext and materiality have centred on the codex – either manuscript or print. Central to these discussions is the concept of the page and the definition of text.

2.1 What is (a) page and what is text?

What exactly do we refer to when we use the term page and why is it important? In her recent monograph How the Page Matters (2011) Bonnie Mak defines the page as the standard interface for transmitting knowledge, the graphic communication of thought, and a dynamic device readily transformed to suit the needs of contemporary readers; it is, she states,

a powerful interface between designer and reader, flexible enough to respond to a variety of demands while remaining comprehensible and communicative [and plays] an important role in the transmission of thought. […] The page is more than a simple vehicle or container for the transmission of ideas; it is a part of those ideas, entangled in the story itself. […] the page sets the parameters for our engagement with ideas. (2011: 3; 9)

In her summary of the history of the page, from papyrus to parchment, and paper to pixel, Mak reminds us that the earliest papyrus scrolls organised their written information by use of columns, known as paginae, or pages, visually dividing the long roll of written material into shorter sections for the effective transmission of ideas; ‘[…] designers understood the pagina as an organizational device that set important controls on the transmission of ideas and facilitated successful communication.’ (Mak 2011: 12) This organisation of information on the scroll, which later developed into the traditional page as we now recognise it, can be seen re-emerging in the organisation of information on Twitter. Twitter can be understood as a continuous scroll of information, consisting of paginae (Tweets) that allow that information to be organised into transmittable units. The Twitter timeline can therefore be seen as a vertical scroll rather than a horizontal one, which can be navigated to both conceal and reveal text. it is bound, as are the written texts contained on papyrus, parchment and paper by the medium itself; that is, Twitter and the individual Tweet are subject to the same limitations, the same constraints of the codex that we find in the traditional paginae of a scroll, or page of a manuscript or printed text. To paraphrase Mak, ‘the pagina [Tweet] thus emerges in the scroll [Timeline] as a conceptual structure by which information could be organized; it visually divides the long roll [screen] of writing material into shorter sections [Tweets] for the effective transmission of ideas’ (Mak 2011: 12, paraphrasing mine).

I take the term text to refer to that central written information, knowledge and ideas conveyed within an individual Tweet. In adopting this definition, I place the concept of a central communicated message within the framework set out by Genette (1997); text is defined as it relates to other literary information contained within that (digital) page, here Twitter. Genette’s concept of paratext organises the anatomy of the page into zones of transition and transaction: the text, the fringe between that text and the reader, and the zone that lies beyond the text. The Twitter Timeline can be analysed in a similar way, and I set out an initial consideration of these zones of interaction below to contextualise further paratextual analyses contained within this article.

3. Text, Peritext and Epitext: The Paratextual Anatomy of Twitter

The following image, using the Twitter feed of the British Library’s medieval manuscripts department (@BLMedieval) as an example, marks out the three zones of: text, fringe and beyond text that are defined in Genette’s work. Text as I define it above for the purposes of this article, exists in direct comparison to that epitextual written information contained in the fringe that can be found existing between the Twitter Timeline and what lies beyond the edges of the digital screen. An example is included below, where that information contained within red can be considered epitext and that within green as text; that which exists outwith the orange demarcation can be considered to be beyond the boundaries of the text:

However, as well as considering the Twitter Timeline as a digital page in its own right, the individual Tweet can also be seen to operate using these three zones. Using the original Tweet of the paw print manuscript from Emir O. Filipović, discussed earlier, one can see these three zones of the page existing within the individual Tweet:

Here, the text i.e. the central message of the Tweet is marked out in green, epitextual elements of the page are marked in red including:

1. The avatar of the user
2. Their Twitter handle (i.e. name)
3. The other users they reference
4. The links and images they include
5. Epitextual elements included in the Tweet by the Twitter platform, rather than the author, such as options for interaction (or annotation) as the ‘reply’, ‘retweet’ or ‘favourite’ links
6. Information showing interaction with the message as contained under the image in this example.

All things which exist beyond the boundary of the Twitter page are again outwith the orange demarcation. As with analogue and digital page alike, form is often dictated by content and both in turn by medium. The type of writing material, the skill of the scribe, the necessity of including paratextual elements within a text are governed by those agents producing that text and the needs of the audience intended to read it. Within an individual Tweet, the author, (or digital scribe?), is afforded only limited autonomy in terms of their scribal repertoire; that is, they are restricted not only by the amount they are able to write (140 characters), but also in the breadth of paratextual mechanisms that are made available for them to employ.

As described above, the Tweet is famously characterised by the character limits imposed on the author: users must create their text to fit within 140 characters. This limitation of the writing space has an obvious effect on the way the user constructs their message; much like the medieval scribe who communicates his text via extensive use of abbreviations, the author is forced to contract their prose to the essential elements of the message they intend to author, or deliver their message over a series of otherwise unconnected Tweets. Authors are further bound by spacing limitations: they are not always able to add meaning to their message by means of exploiting white space within their text; creating a ‘return’ uses valuable characters, and can often fail to produce the required effect. Further limitations are found in the script: authors have no means by which they can italicise or embolden their text and, indeed, have no control over the type font in which their Tweets appear.

The only area authors are afforded free reign is those words they choose to write, the image(s) they choose to include, the links they choose to show, and the punctuation they choose to employ. It is the contention of this article that Twitter not only displays information using key paratextual features that can be traced back through printed text, to manuscript, to scroll, but also that authors can be seen to consciously exploit those limited features of paratext available to them in order to mediate the information their Tweet contains. The designer at work here is both the author and Twitter itself; mediation of the text is controlled by the author of a Tweet and the medium in which that Tweet is displayed. The following discussion therefore considers the influence of these two designers: Twitter vs. the author; this consideration is further organised by discussion of each individual paratextual mechanism the reader encounters.

3.1 The Avatar, Username and Name

Within a Twitter Timeline, Tweets are organised chronologically from most recent at the top of the screen to oldest at the base. Within a Timeline, Tweets are first visually signalled by means of an avatar, or picture of or relating to the author. This image serves not only to indicate a new written message has begun, but also visually distinguishes individual authors from one another. This visual cue indicates to the reader where to start and stop, and thus acts as an aid to visual literacy, in a manner reminiscent of subheadings in a printed text or litterae notabiliores in a manuscript. Genette (1997) notes that section headings are rarely accessible to anyone other than the readers themselves. In other words, the content of these divisions relays information the audience is both at once familiar and engaged with already. On Twitter, these images serve a further function of granting authenticity to an individual message. Users, having selected which Twitter accounts to follow and expecting the contents of those accounts to appear on their Timeline, are reminded and reassured of the identity of the author when they see the avatar. The avatar brings with it an expectation of the content of the message, messages reflecting the preoccupations of their author. Familiarity breeds expectation: the message is ‘trusted’ because familiarity with the author exists. Similarly, the username and name of the author, found to the immediate right of the avatar work together with the avatar to provide a sense of authority to the content of the Tweet. Like the avatar, the design of the Twitter platform prioritises this information to appear at the head of the message; these are the first paratextual mechanisms the reader encounters and the first point of mediation between reader and text: it provides a kind of ‘internal title’ to the Tweet. As Mak argues ‘Internal titles are directed towards the audience and therefore might be considered unambiguous evidence of the communication between designer and reader’. (2011: 29) The same can be said of avatars and usernames on Twitter: these subsectional paratexts give the text authenticity and place it in a conceivable narrative, allowing the reader to navigate and interact with the information given in that Tweet in a different way than if it were from an unfamiliar user (i.e. one is more likely to understand the general context of a Tweet if it is from a recognisable author, than if it is a message retweeted from an author with whom the reader is not familiar). This paratext can therefore be understood to be a fundamental part of the communicative act of Twitter; it functions as the initial mediation between audience and text and marks out the boundaries of that text which is to be read.

3.2 The Writing Space and Text

Following Genette’s terminology of paratext, any paratextual mechanisms that function within the written space can be referred to as peritext. In a digital medium such as Twitter, however, which elements constitute peritext and which constitute text? The mechanisms available to the author are limited at best, including capitalisation, basic spacing, hyperlinking, image, video and punctuation, each existing directly around and within the text. Twitter, acting as a designer, limits the use of these mechanisms considerably when compared to the relative freedom of the printed page or the manuscript folio – or even in contrast to other digital pages online. The outermost limitation is that of the writing space itself. Not only does the Tweet provide the four, clear demarcated edges of the traditional page in which the writing space exists, but as with any writing medium it directly affects the text produced upon it. Any one use of these mechanisms listed above uses vital characters from the 140 available that are already so limited within a Tweet. To include an image might use up to 23 or more characters, while including a hyperlink to an external site can reduce the remaining characters by at least 24. Bearing in mind the impact of these elements on the writing space available, it could be therefore be argued that Tweets containing images or hyperlinks intend these elements to form the central part of the message they convey; that, like a full-page image in a manuscript or early printed book, the image extends the written word. In this sense, hyperlinking and the inclusion of images and videos can perhaps be better understood to form part of what we here consider text: rather than mediating a reader’s reception of a central message, these elements exist within that central message. In this case, the remaining mechanisms available form the peritext: capitalisation, spacing, punctuation.

As discussed above, space is limited in a Tweet; not only is the writing area constrained by the character limit, but the ability to influence the layout of these characters by including paragraph spacing is not always realised. Blank areas, argues Mak, ‘[…] are spaces to remember where one exists as a reader.’ (2011: 32) Like punctuation (which is discussed further below), blank areas create a visual cue to the reader to pause, to consider. On Twitter, space is at a premium. Single Tweets in a Timeline can be unfolded, or expanded, to reveal further detail, paratextual information, and interactions with the text – if the user so desires to view this information. In many ways, Tweets in the Timeline are reminiscent of this folded almanac, brought to my attention in 2012 via Twitter, where each page is viewable by the reader at their leisure:

Like the papyrus scroll, which is simultaneously rolled and unrolled to reveal further paginae of information, we might consider the Twitter Timeline; we scroll down the Twitter page hiding and revealing information as we go. Each Tweet also has this transformative and revelatory quality, having the ability to be viewed collapsed purely conveying the central message of the text, or expanded like the almanac shown above, revealing further scenes of information (such as the paratextual elements outlined in green on the example Tweet discussed previously, including retweets, favourites, and interactions from other users). The reader has the ability to both reveal and conceal paratextual information. As well as being mediated in their encounter with a text, therefore, they mediate their own reading experience by the manner in which they chose to view the paratextual elements that accompany a Tweet.

3.3 Peritext: The Bounded Asterisk and the Hashtag

However, despite this freedom of reading practice, there are also a number of paratextual elements that exist within a Tweet that the reader is not able to conceal. As discussed previously, the ability to italicise or embolden text in a Tweet is not available to the author. Traditional italicisation and emboldening have differing typographical roles, though both signal forms of emphasis. Traditional typographical features such as these align closely with ideas of paratext; like other paratextual mechanisms, typographical elements aid the composition of a text to create a readable, navigable and comprehensible whole, with little conscious awareness on the part of the reader that these elements are mediating their reading experience. As Mak explains, discussing such paratextual mechanisms as entitling, ‘[…] readers rarely regard the title of a book with suspicion, or interrogate its chapter divisions. The paratext says, ‘The text is thus,’ as if it were a statement of fact. Authorized by its own presence, the paratext is trusted because it exists’. (2011: 34) Without typographical features available to the author on Twitter, other means of ‘marking up’ or emphasising text are sought from those paratextual mechanisms that are on offer in the medium, namely capitalisation and punctuation.

In a recent post to the Humanist list, Willard McCarthy posed the following question:

A question, perhaps a matter of a viral stylistic perversion, perhaps a merely technological change. Within the last month or so I have noticed that some messages to Humanist come adorned with asterisks, often suggesting that a tag for emphasis is intended but used so often that only annoyance results. I delete them but would rather not have this work to do. Hoping for their sudden disappearance is likely to be in vain, but an explanation would provide at least temporary comfort. (McCarthy 2013)

His question taps into a number of issues relevant here, relating to the creation of text online – and so the responses explained. One particular favourite response describes this development in use of punctuation for emphasis rather well. It read:

I’m a frantic over user of *s and .s. I apologize immediately because I *know* it’s *highly* annoying. Like putting! exclamation! marks! EVERYWHERE. Or like SHOUTING in *all* caps. Another seriously. annoying. problem. is using the period between words to stress (has the effect of stop motion ‘listening’ to the text). I have annoyed thousands no doubt with my bloody use of the.. The .. is somewhere between full stop and ellipsis. That pause and expectation feeling exists, and it’s not a ; –we miss a glyph for it. Anyway, more to the point: I’m using *emphasis* in plain text if there’s no bold. My use is to put in bold the most important pointers in the text (important dates, absolute bottom line requirements etc.). There is some evidence (in Dutch research in any case) that this supports readers in efficiently mining the information from text […] I would advice to that as well. Sometimes I just use it to annoy people –I know, it’s mean, totally uncalled for. Usually I give James Cummings the occasional emoticon to grind his teeth over ;-) (van Zundert 2013)

So it is with Twitter: authors are being asked – nay, made – to compose their Tweets in plain text; the typographical features they usually rely on to create emphasis and produce paratextual guidance are not available, and so they seek means of visual emphasis elsewhere. McCarthy’s original question to Humanist links particularly well with a discussion of Twitter, as the use of asterisks, or more specifically the bounded asterisk, has become a recognizable paratextual feature within the Twitter medium.

The forms of punctuation we see most popularly used on Twitter are not simply choices made at random. Aside from the standard period (full stop), comma, semi-colon, colon, apostrophes, and quotation marks, two additional punctuation marks have developed particular meaning on Twitter, and these are the bounded asterisk *…* and the hash #. The hash, traditionally used as a metadata tag to denote keywords, was adopted by Twitter to help readers navigate particular topics. According to the advice given by Twitter’s own Help Centre

BBy authoring Tweets and including an appropriate keyword marked with a hash (e.g. #paratext) other users are able to find related posts simply by using a traditional search, contribute to ongoing conversations on a popular topic, or raise awareness, all by the inclusion of a specifically constructed hashtag. Latterly, the hashtag has evolved beyond this simple function as a marker of keywords; authors, perhaps repurposing this punctuation mark because it is already recognizable and well-established in the Twitter medium, are now using the symbol as a paratextual mechanism to mark out not only individual words, but also phrases. These phrases can be seen to act as a sort of self-gloss, or a stage direction, or an aside to the audience, providing a detached comment on the central message of the Tweet.

The asterisk has developed a similar purpose in social media. Traditionally, the asterisk was used in conjunction with other marks as signes de renvoi to link passages in the text with side-notes and footnotes, as well as to mark omissions in a text (Parkes 1993). On Twitter, it is used as a bounded asterisk to mark out both verbal and non-verbal items: e.g. *paratext*. In an interesting blog on the topic, Zimmer (2013) discusses the use of bounding asterisks and early ‘“Peanuts”-style starburst characters’ in comics, among other media, to ‘signal non-verbal noises or actions as a kind of self-describing stage direction’.

In his blog, Zimmer builds on recent work by Yus (2011) which charts the cyberpragmatics of internet-mediated communication, to demonstrate the verbal and non-verbal history of these marks in print and digital comics dating to the 1930s, and argues that what appears to be an emerging performativity of their verbal use within social media can be found to be more firmly rooted in print media:

In comic strips of the early to mid-20th century, cartoonists often needed to represent expressive non-verbal noises in the characters' speech balloons. This was typically done with the use of such words as sob, sniff, cough, hack, wheeze, gasp, gulp, choke, and gag. Such interjections came to be used as a form of onomatopoeia, alongside more directly onomatopoetic expressions like ahem, ugh, ooh, whew, and zzz […] These asterisked stage directions and asides have moved well beyond the onomatopoetic coughs, gulps, and sighs of the comic strips into more complex actions stated in the third person […] This type of asterisking has thoroughly infected Usenet posts, blog comments, tweets, and anywhere else online that people feel the need to describe real-world actions in a virtual space. So, by giving yourself your own stage directions enclosed in asterisks, you treat your own words as lines in a play, and then step outside of your character to give the perspective of the playwright in the play you're acting in. It's all so meta (Zimmer 2013).

In many forms of digital media such as Instant Messaging, Word and so forth, the bounded asterisk was adopted as a shortcut to prompt conversion of those words within asterisks to appear as bold. This development in the use of asterisks in less traditional textual environments (comics) and for particular typographical functions online (emboldening) provides a context for its use as a visible bounded asterisk on Twitter – emphasis. The familiar presence on Twitter of bounded asterisk and hashtag alike, therefore, can be seen a continuation of repurposed punctuation marks, used to mark out text that authors wish to draw attention to – an aside, a self-gloss, a stage direction. Though is it not the intention of this article to analyse the linguistic content of words or phrases marked out within such bounded asterisks, or by hashtags, it is certainly an area worthy of future research; here, the function of the asterisk and hash as established paratextual features recognizable by author and by reader is our primary concern. These pragmatic marks are therefore noticeable by their inclusion in a Tweet (after all, they use up vital characters), but simultaneously are becoming less obvious to the reader as they begin to be absorbed into the typographical system of digital writing. As authors use, reuse and read these marks in online environments and become to recognize their alternative or repurposed function within digital text, these marks become less immediately visible in the reading process. As authors and readers of the digital page, these marks are perhaps beginning to appear as italicization and emboldening are in print: present, but also silent. Familiarity with their form and function means the reader makes unconscious decisions as to how texts are meant to be read that employ these features.

The need for writers to have typographical features available to them in order that they might paratextually and consciously mediate their text before it reaches their intended reader is commonplace, despite the medium and despite the age. A recent Internet meme 8 New Punctuation Marks We Desperately Need makes this point rather well, albeit comically, proposing marks such as the ‘I’m not angry’ mark, to indicate a lack of perceived anger that may be misconstrued from a brief message; the ‘sinceroid’ mark, to indicate true sincerity; the ‘sarcastices’, the opposite of the ‘sinceroid’ which indicate sarcasm; and the ‘andorpersand’ which offers a single symbol for ‘and/or’. The meme emphasizes the need to mediate speech-like written text that exists without the availability of key verbal cues to denote such things as tone.

These suggestions are of course intended to be humorous, but the creation of the meme is founded on a real need for authors of any text not simply just to write, but to create – to express meaning, to mediate. The bounded asterisk and the hashtag are therefore peritextual elements: they exist within the body of the text, but nevertheless mediate its meaning from author to audience. To use the peritextual elements of the bounded asterisk and hashtag at all when Tweeting draws specific attention to the written text they surround or mark; I contend these mechanisms can be understood as being used purposely by the author to demarcate the most salient parts of a given Tweet, and their standardisation of use in Tweets shows evidence of their absorption into the typographical features expected from the digital page that help mediate subsequent reading practices. In this way, one can understand Twitter functioning on a dual level of reader interaction: the designer is at once both Twitter and the individual author. Further, the digital page of Twitter can be understood both as existing in a single Tweet and holistically via a Timeline. Each designer and each page have their own mechanisms of paratext available enabling the reader to navigate their subsequent anatomy, which assist in the mediation of the constrained textual creations that are Tweets.

3.4 Textual Interaction and Annotation

Outwith the central written text of the Tweet we see the final paratextual elements of annotation. Here, the designer is Twitter, rather than the individual author, who collects and displays information pertaining to interaction at the base of a Tweet. In manuscript or printed text terms, these marks are the ‘thumbprints’ left on a text by various readers, the dirt found collected in the corner of a page evidencing reader interaction, and the annotations made on a page that remain for others to see – in Twitter terms, these are the number of retweets, favourites and replies. Consider for example the expanded version of our paw-print Tweet used earlier, which details below the central text of the Tweets the numbers of retweets and favourites, the avatars of those who have interacted in this way, and the subsequent Tweeted replies in chronological order:

There is little difference between these interactions with the text of Twitter, and those currently being investigated by Kathryn Rudy (St Andrews) evidenced by dirt on manuscript folios; the page we encounter bears witness to the act of being read, and being read repeatedly. [7] It reflects the reading practices and the mediations experienced by those readers. Just as our avatar and Twitter username lent an air of authenticity to a Tweet, so this paratextual evidence of reader interaction reinforces the authenticity of the message contained within a Tweet. Where paratext authorizes text, here evidence of readership authorizes the text. The reader, therefore, encounters the digital page of Twitter as they would a traditional, analogue page: from top to bottom, mediated by a variety of paratextual and pragmatic mechanisms. The digital page is therefore bound to the traditional requirements of paratext to mediate reading; these requirements differ little to those found, by contrast, in the earliest manuscripts and printed texts. Attention is drawn to specific pieces of information by a variety of paratextual features, namely visual references to further information, images, spacing, and punctuation. The single Tweet as well as the holistic Timeline can each be considered as different types of digital page, functioning as a means of communicating knowledge in written form where the material form, design, layout and decoration are integral to how a particular audience accesses the meaning of the text contained within.

4. Conclusion

By using traditional methodology employed by codicologists and palaeographers to analyse the codex and the individual page, this article has considered how we might approach a ‘reading’ of the page constructed by Twitter. It has posed the question: can we regard an individual Tweet as a page and if so, do the traditional parameters, limitations and thresholds of the page apply. Further, it has offered a preliminary examination of whether Tweets require or employ paratext to mediate their content and, if so, how this paratext is conceived within the limitations imposed. This consideration of the paratexts of Twitter is by no means intended to be exhaustive or conclusive, but it is hoped the above discussion can serve perhaps as an introduction to the kinds of conversations that can be seen to occur between designer and reader on the digital page, to help place the digital page firmly within the context of being a cultural artefact worthy of analysis. Each individual Tweet on a Timeline can be read both alone as an individual text in its own right, and as part of a continual scroll of information. Each communicative act is marked out by an avatar (acting perhaps as a littera notabilior) and a username (acting as a digital diminuendo, or further authentication of the reading space) before the text proper begins. Paratext is present in the inclusion of spacing, capitalisation and punctuation, and the retweets, favourites and replies can be seen as annotations by the reader. Each of these elements contributes to a zone of transition and transaction between page and reader, what Genette has previously referred to as the unsettled limits that enclose with a ‘pragmatic halo’ the literary work (Genette 1997).

As Mak herself explains:

Digital pages are complex interfaces that provide a point of contact between designer and reader. Like their analogue counterparts, these pages communicate verbally, graphically, aurally, and tactilely, and are constructed in a material way that influences how they are rules and understood. Each page is a unique combination of visual expression and physical platform, and it is with this dynamic relationship that the page transmits information and contributes to the making of meaning […] the entanglement between the material and mattering of the page continues to be of central importance. (2011: 62)

The structure of the Tweet and Timeline celebrate engagement and interaction with text and actively encourage a practice of continuous reading. Rather than being a new development in the evolution of the book or page, Twitter can be seen to be governed more closely by traditional codicological limitations and expectations than we might first realise. The means of transmitting this knowledge graphically is rooted more in the earliest transmission of knowledge – the scroll and the paginae. It is these traditional elements of written text, page, and paratext that are central to the success of such a fast-paced method of written communication: they are necessary in order to allow such short bursts of information to be mediated effectively within the space provided. Traditional paratextual mechanisms so familiar within the analogue page make the new, digital page instantly navigable. Though capable of evolution via digital media, our reading practices demonstrated on Twitter are rooted in that analogue page. To conclude,

[…] these principles of layout set fundamental conceptual parameters that prevail to this day, shaping our conceptions of textual authority […] Just as modern academics continue to work within these medieval structure, we continue to work in the intellectual framework of the medieval page. The present digital hypertext environment […] with its elaborate hierarchy of information and privileging of visual over acoustic data, is a continuation of this tradition […] The material page carries with it medieval notions of reading and quantifying information. The electronic writing space, seldom as yet responsible for revolutionizing those notions, nevertheless holds out the possibility for doing so. Such a possibility can only be realized if not just the differences but the similarities between the material page and its electronic counterpart are recognized. (Stoicheff and Taylor 2004: 12–19)

Notes

[1] It should be noted that each Figure used in this article was originally brought to my attention via Twitter.

[2] Information correct as of 23.06.2014.

[3] This author once experienced a start-up project event which ran the event’s live Twitter feed on a screen behind the speakers, sharing the views of the audience in real-time to everyone in the room, as the papers were being delivered.

[4] See Roopika Risam’s 2012 blog for a summary of the debates that surrounded #Twittergate and links to other posts of interest: http://www.emory.edu/ACAD_EXCHANGE/whats_new/roopikatweetingconferences.html.

[5] A number of studies are now emerging which engage with varying aspects of communication at work via Twitter, examples of which include, inter alia: Buzeck, M. and Müller, J. (2010); Celik, I., Abel, F., and Houben, G. (2011); Hurlock, J. and Wilson, M. L. (2011); Java, A., Song, X., Finin, T., Tseng, B. (2007); Stoicheff, P. and Taylor, A. (2004).

[6] Two manuscript copies of Hoccleve’s Regiment of Princes dating after c.1411 and between 1411 and 1432 respectively.

[7] See Rudy (2010).

Sources

Digital Image of The Great Isaiah Scroll, the Israel Museum, Jerusalem: http://dss.collections.imj.org.il

References

Barry, P. 2007. Literature in Contexts. Manchester: Manchester University Press.

Barthes, R. 1979. “From work to text”. Textual Strategies: Perspectives in Post-Structuralist Criticism, ed. by J. V. Harari, pp. 73-81. Ithaca: Cornell University Press.

Buzeck, M. & J. Müller. 2010. “TwitterSigns: Microblogging on the walls”. Proceedings of the 18th International Conference on Multimedia 2010, Firenze, Italy, October 25–29, 2010, ed. by A. D. Bimbo, S. Chang, & A. W. M. Smeulders, 819–22.

Caie, G. D. 1988. “The vernacular poems in MS CCCC201 as penitential literature”. A Literary Miscellany Presented to Eric Jacobsen ed. by G. D. Caie & H. Nørgaard, 72–77. Copenhagen: University of Copenhagen.

Celik, I., F. Abel & G. Houben. 2011. “Learning semantic relationships between entities in Twitter”. International Conference of Web Engineering (Lecture Notes in Computer Science 6757), ed. by S. Auer, O. Diaz & G.A. Papadopoulos, 167–181.

Dahood, R. 1988. “The use of coloured initials and other division markers in early versions of ‘Ancrene Riwle’”. Medieval Studies Presented to George Kane, ed. by E. D. Kennedy, R. Waldron, & J. S. Wittig. Woodbridge: D.S. Brewer.

Fulk, R. D. 2006. “The origin of the numbered sections in ‘Beowulf’ and in other Old English poems”. ASE 35. 91–109.

Genette, G. 1997. Paratexts: Thresholds of Interpretation. Repr. in translation M. Mclean (1987) from Seuils with permission of Georges Borchardt, Paris.

Green, J.M.E. 2012. Judgement Day I, Resignation A and Resignation B: A Conceptual Unit in the Exeter Book. Ph.D. dissertation, University of Glasgow.

Greetham, D. C. 1994. Textual Scholarship: An Introduction. New York: Routledge.

Humphries, L. 2010. “Historicizing microblogging”. CHI Workshop on Micro-Blogging 2010 April 10–15, 2010, Atlanta, Georgia. 1–4.

Hurlock, J. & M. L. Wilson. 2011. “Searching Twitter: Separating the tweet from the chaff”. Association for the Advancement of Artificial Intelligence.

Java, A., X. Song, T. Finin & B. Tseng. 2007. “Why We Twitter: Understanding Microblogging Usage and Communities”. In Proceedings of Joint 9th WEBKDD and 1st SNA-KDD Workshop 2007.

Kelly, S. & J. J. Thompson, eds. 2005. Imagining the Book. Turnhout: Brepols.

Kisor, Y. 2006. “Numerical composition and ‘Beowulf’: A reconsideration”. ASE 38. 41–7

Mak, B. 2011. How the Page Matters. Toronto: University of Toronto Press.

McGann, J. 2001. Radiant Textuality. New York: Palgrave.

McGann, J., ed. 2007. The Rossetti Archive. http://www.rossettiarchive.org/

McKenzie, D. F. 1986. Bibliography and the Sociology of Texts. London: British Library.

Minnis, A. & J. Roberts, eds. 2007. Text, Image and Interpretation: Studies in Anglo-Saxon Literature and its Insular Context in Honour of Éamonn Ó Carragáin. Turnhout: Brepols

Nichols, S. G. & S. Wenzel. 1996. The Whole Book: Cultural Perspectives on the Medieval Miscellany. Michigan: University of Michigan Press.

Parkes, M.B. 1993. Pause and Effect: An Introduction to the History of Punctuation in the West. Univ. of California Press.

Robinson, P. R. & R. Zim, eds. 1997. Of the Making of Books: Medieval Manuscripts, their Scribes and Readers. Essays Presented to M. B. Parkes. Aldershot: Scholar Press

Robinson, P. 2003. “Where we are with electronic scholarly editions, and where we want to be”. Jahrbuch für Computerphilologie 5: 125–146.

Rudy, K. 2010. “Dirt books: Quantifying patterns of use in medieval manuscripts using a densitometer”. Journal of Historians of Netherlandish Art 2(1–2). http://www.jhna.org/index.php/past-issues/volume-2-issue-1-2

Sanz, A. & M. Goicoechea. 2012. “Literary reading rituals and practices on new interfaces”. Literary and Linguistic Computing 27: 331–346.

Schulz, C. M. 1974. Peanuts. Comic originally published April 19, 1974.

Stoicheff, P. & A. Taylor, eds. 2004. The Future of the Page. Toronto: University of Toronto Press.

Watterson, B. 1989. Calvin and Hobbes. Comic originally published September 24, 1989.

White, T. 2012. “The dark is light enough: The layout of the tale of Sir Thopas”. Dark Chaucer: An Assortment, ed. by M. Seaman, E. Joy & N. Masciandaro. New York: Punctum Books.

Yus, F. 2011. Cyberpragmatics: Internet-Mediated Communication in Context. Philadelphia: John Benjamins.

Blogs cited

Cohen, J. J. 2012. “The best MLA is the one I didn’t attend: Tweeting the conference.” In the Medieval Middle. http://www.inthemedievalmiddle.com/2012/01/best-mla-is-one-i-didnt-attend-twitter.html

Kwakkel, E. 2013. “My first year on Twitter: How I became @erik_kwakkel”. Medieval Fragments. https://medievalfragments.wordpress.com/2013/01/18/my-first-year-on-twitter-how-i-became-erik_kwakkel/

Zimmer, B. 2013. “The cyberpragmatics of bounding asterisks”. Language Log. http://languagelog.ldc.upenn.edu/nll/?p=4466

Discussion lists cited

McCarthy, W. 2013. “Why the asterisks?” Humanist 26.832. http://lists.digitalhumanities.org/pipermail/humanist/2013-March/010729.html>

van Zundert, J. 2013. “Why the *********?” Humanist 26.838. http://lists.digitalhumanities.org/pipermail/humanist/2013-March/010735.html