Digital Stages for Old Plays: A Review of Shakespeare’s Language in Digital Media: Old Words, New Tools

Shakespeare’s works and their relation to the language and texts of his culture are hardly new arrivals in online digital scholarship or a new focus in the development of digital methodologies and tools; indeed, the editors of Shakespeare’s Language in Digital Media argue that “the state of computing in Shakespeare studies” functions as “a benchmark of our scholarly digital literacy in general: now, as ever, Shakespeare is the test bed for our latest remediation technologies” [Jenstad, Kaethler, and Roberts-Smith 2020, 2]. As a case in point, the editors relate that twenty years ago, in 1997-8, a sort of precursor to their collection appeared in the form of a pair of special issues of Early Modern Literary Studies. The essays there focused on two new/nascent digital scholarly resources: the Early Modern English Dictionaries Database (EMEDD), under the editorship of Ian Lancashire, and Internet Shakespeare Editions (ISE), under the editorship of Michael Best. Today, EMEDD has been superseded by Lexicons of Early Modern English (LEME), edited by Lancashire, and ISE (now under the editorship of Janelle Jenstad) has been joined by sibling projects like Queen’s Men Editions (QME), Digital Renaissance Editions (DRE) and the Map of Early Modern London (MoEML). While the essays in the present collection bring together text editing and historical linguistic analysis in digital scholarship, the editors point out that a divergence persists between these two fields that is mirrored in the differing directions that LEME and ISE have taken towards their texts: the former along a “dynamic text” path facilitating lexical research across a corpus of period glossaries, the latter along a “hypertextual edition” path facilitating enhanced reading and study of individual Shakespeare plays. The editors point to signs of convergence of these divergent paths, and one of the sources of interest in this collection for the digital humanist (Shakespearean and non-Shakespearean) is the ways the essays suggest how – in the digital humanities test bed that is digital Shakespeare – such diverging scholarly activities and foci (the word and the text, the micro and the macro) might be integrated in or across future digital scholarly environments.

On the whole, the essays in this collection focus on Shakespeare’s (or Shakespeare-adjacent) plays and how they can be newly understood and newly edited with primary resources such as LEME, Early English Books Online (EEBO), digital editing/publishing platforms like ISE and QME, and some common digital tools like Voyant Tools, Python scripts, and spreadsheet software. In addition, Laura Estill and Andie Silva’s essay “Storing and Accessing Knowledge: Digital Tools for the Study of Early Modern Drama” provides a critical if somewhat cursory survey of some other key online resources that are currently available to scholars of early modern drama: the English Short Title Catalogue (ESTC), the Database of Early English Plays (DEEP), Early Modern London Theatres (EMLoT), Patrons and Performances, and the World Shakespeare Bibliography (WSB).

Shakespeare’s language/word usage is the focus of the essays of Part I, “Old words through new media: Re-reading Shakespeare with EEBO-TCP and LEME.” As the editors point out in their prefatory marks to this section, LEME and EEBO-TCP (and their interlinking) now compel editors of early modern play texts to rethink how they gloss words, given that they can now not only explain words that contemporary readers might find obscure, but also identify words that Shakespeare’s contemporaries would have found unfamiliar, and which therefore prompt closer examination of the speeches and scenes in which these unfamiliar (or “hard”) words are present.

The first essay here, Valerie Wayne’s “Beyond the OED Loop: Digital Resources and the Arden 3 Cymbeline,” begins by pointing out the reliance of editors of Shakespeare on the Oxford English Dictionary (OED) in determining the uniqueness of Shakespeare’s deployment and development of English vocabulary and word usage, and the reliance of the OED on Shakespeare (and his editors) as a source for new usages of words and new coinages — the so-called “OED Loop.” One way around this loop is the Early English Books Online-Text Creation Partnership (EEBO-TCP), comprising (as of July 2016), 60,331 texts coded for searching, which Wayne uses in conjunction with LEME and the OED to “assist in the recovery of words once thought to be errors in early texts, correct claims for coinages, clarify occluded meanings, and detect the most prevalent forms and spellings” [Jenstad, Kaethler, and Roberts-Smith 2020, 15]. (As the following essay by Lancashire and Tersigni notes, as of 2018, LEME comprised about “665,000 word-entries in 199 lexical works from about 1475 to 1702,” with LEME and the OED “overlap[ping] by less than 5% in the quotations they offer” [Jenstad, Kaethler, and Roberts-Smith 2020, 29]). Wayne proceeds to demonstrate how such documentary editing is done using specific examples from Cymbeline, in the process — even if only provisionally, given the material still forthcoming on these resources — revealing doubtful editorial emendations, clarifying cruxes in the interpretation of phrases, calling into question Shakespeare as the coiner of certain words and word usages, recovering often downplayed or not fully appreciated sexual meanings, and determining the usual spelling of a proper name (it’s Innogen, not Imogen). As Wayne points out: “Both databases enable forms of recovery and argument that were barely possible a decade or two ago and are of immense help to an editor” [Jenstad, Kaethler, and Roberts-Smith 2020, 24] who wants to do a historically sensitive analysis of Shakespeare’s use of language.

Ian Lancashire and Elisa Tersigni’s “Shakespeare’s Hard Words, and Our Hard Senses,” like Wayne’s essay, shows how LEME can provide more historically accurate and more coherent glosses on Shakespeare words. Noting that “no editors have yet annotated Shakespeare’s works for the relative difficulty of their words to his contemporaries” [Jenstad, Kaethler, and Roberts-Smith 2020, 27], they use LEME to explore how Shakespeare’s contemporary audiences understood his language and how that affected the perception of his characters, focusing on Richard III and The Tempest. They note a marked difference in what words contemporary editors gloss and what words were considered “hard” by those who created glossaries in the early modern period, and compare the differing deployment of hard words in the speeches of Richmond and Richard to their troops before the Battle of Bosworth, which, they argue, Shakespeare employed innovatively to “illuminate his characters” [Jenstad, Kaethler, and Roberts-Smith 2020, 28]. Using the Hard Word Annotator, a LEME tool that accepts plaintext versions of modern-spelling Early Modern texts and identifies the hard words in them, they analyze the entirety of The Tempest “to test whether hard words are used consistently as part of the characterization,” concluding that the “percentage of hard words used by characters reflects their personality, occupation, race, gender, and social status” [Jenstad, Kaethler, and Roberts-Smith 2020, 32]. The bookish Prospero, perhaps unsurprisingly, emerges as the leader in the use of hard words, which Lancashire and Tersigni cleverly suggest explains Prospero’s repeated interruption of his recounting of his past life in Act 1, Scene 2: to confirm that Miranda (and the audience) are still paying attention to a speech peppered with baffling words. Their analysis also supports the decision by some editors to reassign to Prospero a speech (also in Act 1, Scene 2) assigned to Miranda in the First Folio, given that the frequency of hard words is more in character with Prospero’s vocabulary than his daughter’s.

The remaining two essays of Part I, Daniel Aureliano Newman’s “Terms of Art in Law and Herbals,” and Elizabeth Bernath’s “‘Strangers enfranchised’: Shakespeare’s Hamlet and the Mother Tongue,” use LEME and EEBO in conjunction to explore, in the former, the use of herbal and legal “terms of art” (or what we would call jargon) in Hamlet, King John, and The Winter’s Tale (with a particularly compelling contribution to, if not resolution of, the vexed issue of Perdita’s “streak’d Gilly-vors”), and, in the latter, a comparison of hard words in Hamlet against the diachronic evidence of the “enfranchisement” of hard words in the English “mother tongue.” As with Lancashire and Tersigni’s analysis of Richmond’s speech in Richard III, Bernath argues that, in Hamlet’s soliloquies, Shakespeare provides glosses, often synonyms, for (still) hard words within the speeches themselves.

The essays of Part I together offer a compelling case for the important insights that can result from textual databases whose encoding, data structures, and search tools can enable sophisticated and wide-ranging linguistic research queries. The remaining four essays (besides the Estill and Silva essay previously mentioned) broaden the scope beyond words and language, shifting, in effect, from a ‘LEME orientation’ to an ‘ISE orientation.’ Andrew Griffin’s “Text, Performance, and Multidisciplinarity: On a Digital Edition of King Leir,” describes how the capaciousness of digital editions enables the Queen’s Men Editions (QME) to avoid the prescriptivism that invariably manifests in print versions of performance editions, given their limited ability to incorporate a full range of production examples. Instead of a transhistorical or ahistorical or ideal conceptualization of a performance of a play, the QME of King Leir, by collaborating with a team who staged a production of the play, can link to its specific production details (in the form of images and videos) within the context of production notes that detail possible performance options considered, discussion about these options, and what option was eventually chosen (and why). Toby Malone’s “A Digital Parallel-Text Approach to Performance Historiography” describes a prototype parallel-text digital edition (created using spreadsheet software) enabling comparative analysis of the Quarto 1 and Folio versions of Richard III, along with 10 historical promptbooks and performance editions ranging from the 18th to 20th centuries relating to historically significant Shakespearean actors such as Colly Cibber, David Garrick, John Philip Kemble, Edmund Kean, William Charles Macready, and Henry Irving. Malone argues that “[o]ne simple and ubiquitous digital tool — a spreadsheet — makes the systematic, large-scale, and flexible comparative close reading of performance texts possible ” [Jenstad, Kaethler, and Roberts-Smith 2020, 122] in ways that print parallel texts cannot. While some digital humanities scholars might scoff at the use of a general-use spreadsheet program for a (prototype) digital edition, such a tool has the advantage of realizing and ensuring some degree of longevity to an digital edition than one created through a digital humanities-specific comparison and collation tool like the now-defunct Juxta Commons, a web service that was until recently offered by NINES (Networked Infrastructure for Nineteenth-Century Electronic Scholarship), developed at the University of Virginia.

Michael Ullyot and Adam James Bradley’s “Past Texts, Present Tools, and Future Critics: Toward Rhetorical Schematics,” can be regarded as occupying a transitory place between a focus on word analysis and a focus on textual editing, given its focus on rhetorical figures. In the essay, the authors discuss their creation of a Python script designed to process plaintext corpora like Martin Mueller’s Shakespeare His Contemporaries (comprising Shakespeare’s plays along with most of the plays written within a generation before and after his active career as a playwright) to find a rhetorical scheme called gradatio, which is a series of anadiploses. The essay cites an example of gradatio from As You Like It (Act 5, Scene 2): “For your brother and my sister not sooner met but they looked; no sooner looked but they loved; no sooner loved but they sighed; no sooner sighed but they asked one another the reason; not sooner knew the reason but they sought the remedy; and in these degrees have they made a pair of stairs to marriage.” While the Python script itself is not included, a fairly clear description of it is: its goal was “to break text files into sentences, to tokenize their words and lemmatize their tokens, to insert clause breaks based on punctuation, and finally to use Regular Expressions to look for matching lemmas…within four words of the clause breaks” [Jenstad, Kaethler, and Roberts-Smith 2020, 147]. The script identified 114 instances of what it was told was gradatio in 400 plays for analysis and evaluation, leading the researchers to contemplate not only the efficacy of the script, but their critical understanding of graditio itself. The authors concede that there might be instances of gradatio this script did not identify but argue that their modest purpose was to obtain “more complete results than we could gather without the algorithm.” They also concede that such scripts appear to only be effective at identifying rhetorical schemes “of direct repetition and variation” [Jenstad, Kaethler, and Roberts-Smith 2020, 147] of words or phrases, and so cannot be used to locate a broad spectrum of rhetorical figures in digital texts corpora.

In envisioning practical future solutions for “a more nuanced, multivalent, tool-enhanced criticism” that would extend their script and integrate the textual edition and the database (or, in other words, ISE and LEME), Ullyot and Bradley they propose “an unobtrusive plug-in for ISE that makes [its editions] interoperable with databases” and that “preserve[s] close reading while promoting distant reading, balancing our attention to local phenomena with our ambient awareness of comparable phenomena” [Jenstad, Kaethler, and Roberts-Smith 2020, 152–3]. They argue that such interoperability requires that textual corpora have APIs, as Folger Digital Texts does: “An API for the ISE would allow its texts to offer limitless new questions, which these interoperable databases can illuminate” [Jenstad, Kaethler, and Roberts-Smith 2020, 154].

The idea of limitlessness that the interoperability of digital resources promises is clearly in evidence in Diane K. Jakacki’s “Internet Shakespeare Editions and the Infinite (Editorial) Others: Supporting Critical Tagsets for Linked Editions,” which also takes up the topic of integration/interoperability, albeit in a less speculative and more grounded (and consequently wary) fashion. As the editor of the ISE’s edition of Henry VIII, as well as the Technical Editor of ISE, Jakacki is alert to both the possibilities and the pitfalls of interoperable, networked digital editions. The key challenge the essay addresses is “to what degree we should engage in acts of editorial disruption that allow us to move forward towards infinity [the infinity represented by other potential editions] while maintaining editorial stability across digital projects” [Jenstad, Kaethler, and Roberts-Smith 2020, 158]. As she considers her Henry VIII edition beyond the additional TEI encoding she has used and that she demonstrates can be valuable for aiding an editor’s critical interpretation, and starts to consider it as part of a network that ties it in with digital resources like MoEML and DRE, a digital edition of a single play starts to stretch out towards an unfinishable infinity: Jakacki makes reference to how a non-peer reviewed edition of Samuel Rowley’s When You See Me You Know Me that she intended to include in her edition of Henry VIII (presumably as an appendix) was, with the launch of DRE, re-envisioned as a peer-reviewed edition of Rowley’s play for DRE that would serve as a companion edition to her ISE of Henry VIII. She also mentions the idea of creating a shared Personography of all the characters that appear across ISE, QME, and DRE. The challenge Jakacki points out, with legitimate concern, is that considering the creation of an edition in light of how it can be interoperable with other digital resources can significantly add to an editor’s task and timelines to completion. Indeed, such work can potentially take an editor far from their initial interest: the text(s) of a Shakespeare play. Can editing ever be said to 'end' in an interoperable digital environment?

There is much more to the essays in this collection that I have not foregrounded and that would be of value to scholars primarily interested in historical linguistic analysis, Shakespeare studies and scholarly editing. While readers primarily interested in the technical side of digital humanities may find the essays less than forthcoming (Ullyot and Bradley, for example, do not provide the actual Python script they created, and so one cannot judge whether it could be modified or extended to detect similar or other textual structures) and vague on future directions and modes of development, especially regarding cyberinfrastructures for interoperability, there is solid content and provocative ideas in this collection about the possible paths (or perhaps path?) to making Shakespeare’s texts digitally accessible in all their complexity as well as early modern English texts and culture more broadly. If the editors’ claim that Shakespeare is the test bed for new developments in digital humanities is true, we should continue to look to the digital remediation of the author’s old texts for inspiring new digital tools.