Shakespeare Quarterly

Shakespeare Goes Digital: Three Open Internet Editions

Andrew Murphy
University of St. Andrews

0 In 1853 J. O. Halliwell-Phillipps began issuing, through the publishers C. & J. Adlard, an extraordinarily elaborate edition of the works of Shakespeare. It ran to 16 volumes in oversized folio format and it was copiously illustrated. The edition was intended for a strictly limited market — just 150 copies were printed. The price also reflected the restricted readership that it was aimed at: 25 copies were issued with the illustrations on India paper, at 150 guineas per set; the remainder sold for 80 guineas per set. As the final volumes of the edition were starting to appear, in the mid-1860s, Halliwell-Phillipps began discussing a very different kind of Shakespeare project with Adlards. His idea was to strip his edition to its absolute bare bones in order to produce a popular text that would sell for just 1s, or .06% of the cost of the 80 guinea folio edition.[1]

0 Halliwell-Phillipps subsequently explained to the publisher John Camden Hotten what his vision for the shilling edition had been, observing that ‘One of the chief objects in the original design was the distribution of copies by employers amongst the working classes’.[2] The editor was motivated here by the progressive emergence of a working-class readership in the early to middle decades of the nineteenth century — a readership which, Halliwell-Phillipps sensed, might take to Shakespeare if the playwright’s works were made readily available, either at a low price, or, ideally, provided free of charge by paternalistic employers. His essential instinct was correct, even if his own project never came to fruition: by the end of the 1860s at least three publishers were issuing shilling editions of Shakespeare and the texts sold extraordinarily well. The era of ‘mass Shakespeare’ had well and truly arrived.

4 Halliwell-Phillipps’ double project anticipates, in interesting ways, developments in Shakespeare publishing from the end of the twentieth century, as the playwright’s text has moved from the page to the computer screen. When Thomas Nelson issued the Arden Shakespeare CD ROM in 1997, the initial (pre-sales tax) selling price was £2,500 in the UK and $3,995 in the US. A certain Prof Pyper at the University of St Andrews had, in the nineteenth century, managed to subscribe to Halliwell-Phillipps’ edition, but it seems highly unlikely that any academic of average means could ever have afforded to buy the electronic Arden. The same could be said of other early computer-based Shakespeare resources (such as Chadwyck-Healey’s Editions and Adaptations of Shakespeare) which were also priced well beyond the budget of individual purchasers. But, just as Halliwell-Phillipps had a twin vision of, on the one hand, a highly elaborated package sold at the highest possible cost, and, on the other, a far more basic offering provided either at a minimal price or wholly free of charge, so too did the digital world split between high cost packages and cheap or, more commonly, free-to-access offerings. So, early in the 1990s, for example, the text of the plays had been rendered into a form which could be presented on screen, and this version of the text — the ‘Moby Shakespeare’ — was distributed through various websites, the most enduring being those established by Matty Farrow in Australia (http://www.it.usyd.edu.au/~matty/Shakespeare/index.html#list) and by Jeremy Hylton at MIT (http://shakespeare.mit.edu/).

2 The Moby text is based on the Globe Shakespeare, first published by Macmillan (in conjunction with Cambridge University Press) in 1864. The reasons for choosing this particular text would appear to be lost in the mists of prehistoric digital time. Copyright on the Globe edition had, of course, long since lapsed — an important consideration. It may simply have been happenstance that led the text’s creator to this particular copyright-free edition, but it remains an interesting choice, nonetheless, from a Shakespearean point of view. The Globe was, of course, itself part of a bifurcated publishing project. Alexander Macmillan commissioned William George Clark and William Aldis Wright (and John Glover) to produce a scholarly edition of the plays, the first to be created by university academics. Just 1,500 copies of the edition were printed, selling at 10s 6d per volume, with the total cost running to almost £5. The nine volume edition was subsequently reduced to a single volume, selling at just 3s 6d. Macmillan had intended the text specifically for the popular market and was rather chagrined to discover that, within a couple of years of the Globe’s appearance, it was competing against the 1s editions mentioned earlier. The Globe, however, held its own and served as the standard reference text for several decades, with many subsequent scholarly editions keying their referencing system to the Globe act, scene and line numbers.

4 In a way, then, it is fitting enough that, out of all readily available copyright-free texts, it should have been the Globe that emerged from the primal haze of the early digital world to serve as the standard Internet Shakespeare. And it persists today, not only in the (now rather primitive-feeling) Farrow and Hylton sites, but also in new offerings such as Eric M. Johnson’s Open Source Shakespeare (OSS; http://www.opensourceshakespeare.org/). One of the key features of Farrow’s site was that it offered users a search function, making the business of tracking quotations and carrying out some basic concordancing tasks much easier than it had ever been before. Johnson has taken this idea and has built upon it very substantially, using the much greater range of programming tools now available to computer specialists. In general terms, the site is attractively laid out and operates intuitively. The plays are segmented by scene, but, by clicking on ‘Complete Play’, it is possible to access the whole text and to read it through simply by scrolling down. Johnson uses TLNs, with the numbers appearing after every five lines (including stage directions). In ‘Complete Play’ mode, the act and scene markers are relatively unobtrusive (which is to say, less obtrusive than in most print editions, but not as unobtrusive as those used in the Oxford Shakespeare). A nice feature of the site is that Johnson has created a version of the text presented as ‘The Bard for the tiny screens’, intended for use with mobile devices (http://mobile.opensourceshakespeare.org/). Even on a relatively cheap ‘phone handset this produces a perfectly readable text, though one would not, obviously, want to have to read all three parts of Henry VI in this way.

0 Where OSS really comes into its own is with the tools that Johnson has added to the text. Every single word is linked to a concordancing function. Doubleclicking on any word brings up a screen indicating how many times that word appears in the complete canon, with a link to a set of statistics about the word, including the number of instances by play, and a further link which produces a full set of all instances. Mary Cowden Clarke must surely be wonder-struck, somewhere out there in the ether. Clicking on a character name produces a complete set of lines for that character. This is useful for actors, of course, but Johnson offers an even more useful function: once you reach the character lines that have been ‘harvested’, clicking on ‘Show cue lines’ produces a full set of the character’s speeches together with, in each instance, the immediately preceding line of the text. The homepage of the site also includes tools for carrying out text search, concordance search and character search.

0 OSS includes a paper entitled ‘Open Source Shakespeare: An Experiment in Literary Technology’, which provides an account of the process of constructing the site. This is essentially a cutdown version of an MA thesis submitted by Johnson at George Mason University. It makes for interesting reading. Johnson indicates that much of the work on the site was undertaken while he was stationed in Kuwait as a Marines reservist during the most recent US invasion of Iraq — a piece of information that could make critics of a certain age nostalgic for the high days of Cultural Materialism. There are substantial sections on the history of the Globe text and on its modern digital incarnations. This material is solid, well informed and considered, particularly given that Johnson does not have a bibliographic background.

0 In ‘The Editing and Structure of the Open Source Shakespeare’ section of his paper, Johnson describes his seven stage working method, and then notes:

0 This procedure might seem very complex, and indeed it took many hours to perfect. However, the last fifteen or sixteen plays went very quickly, as it was just a question of repeating the same process over and over. I got to the point where I could finish one or two plays an hour, depending on how many discrepancies there were in the texts.

0 Johnson’s ‘one or two plays an hour’ might be contrasted with John Dover Wilson’s comments, as he drew to the end of his almost half-century long stint of editing the New Cambridge Shakespeare: ‘Two or three life-times are insufficient for the proper editing of Shakespeare, and I know it well enough. Had I lingered over Hamlet alone as long as I ought to have done, and should like to have done, I should still be at it.’[3] Of course, Johnson’s project and Dover Wilson’s are very different from each other, and here we get to the heart of the problem of any resource which has as its remit the business of taking the Globe (or any text of similar vintage), standardising it, and adding electronic functionalities: manipulating the text in this way may produce a useful resource, but, as it does not, of course, involve editing the text, the heart of the project will inevitably remain fundamentally (and often fatally) outdated. By way of comparison here, one might say that, while one probably could fit a Model T Ford with power steering, anti-lock brakes and airbags, it would still remain, in essence, a Model T Ford. The Globe was, indeed, a good text in its day — and Clark and Wright were fine scholars — but it predates the New Bibliography, not to mention the various important developments in textual thinking of more recent decades. To take just one small example of the problems with the Globe: Clark and Wright worked on their edition before W. W. Greg and William J. Neidig demonstrated that the Pavier Quartos were all published in 1619, despite the differing imprints provided on the titlepages of some of the texts. As far as the Globe editors were concerned, then, Nathaniel Butter’s King Lear of 1608 and Pavier’s 1619 Lear (with its title page falsely dated ‘1608′) were published in the same year. The editors thought — at least initially — that Pavier’s was the edition first published, and, therefore, the more authoritative. This confusion is written through Clark and Wright’s text of Lear. Problems of a similar nature arise with other plays in the canon, compounded by the fact that Clark and Wright had a tendency to favour eclectic editing methods, sometimes picking and choosing readings from different editions without consistently considering the question of what the exact authority of these texts might be.

4 Are the shortcomings of the Globe wholly fatal to the Open Source Shakespeare? Sadly, for scholars and serious students, the answer is likely to be ‘yes’. One might, of course, make the argument that the Globe was the text of Shakespeare for a very broad range of readers for at least a half century after its first release — and, as Margreta de Grazia has indicated, it was very much the dominant Shakespeare text of Britain’s high period of Victorian imperialist expansion, carried around the world in much the same way that Johnson carried it (electronically) to Iraq and Kuwait.[4] In this sense, OSS does provide an interesting reading text, if we view the Globe as a historical artefact. The trouble is that the concordancing tools — which are, after all, one of the central new resources offered by OSS — are the element of the package likely to be of most interest to scholars, and no scholar could really contemplate using these tools for serious research, when the dataset they are being applied to is almost 150 years out of date. Having said all of this, perhaps the most interesting thing about OSS is that, in terms of its structure and functionality, it provides a good model for how a site of this kind can (and should) work. It is, effectively, a ‘Web 2.0 ready’ site, in the sense that the materials are encoded in database form, rather than as page images — this is what enables OSS to offer a very high level of manipulation of the materials included on the site. Very generously, Johnson has made all of his resources freely available, and he expresses the ‘hope that other people will use the code and database as examples for their own work’. If the architecture of this site could be applied to a more up-to-date text, then scholars really would have a worthwhile resource at their disposal.

0 In a sense, this is the approach taken by David and Ben Crystal, in their Shakespeare’s Words website (SW; http://www.shakespeareswords.com/). At the heart of this package is the New Penguin Shakespeare, with J. M. Nosworthy’s Arden 2 Cymbeline and Giorgio Melchiori’s New Cambridge Edward III added, as neither of these had appeared in the New Penguin at the point when the project was initiated. The age of some of the texts is something of a problem here again: Nosworthy’s Cymbeline was first published more than a half century ago and some of the New Penguin texts are themselves rather old now (the earliest titles in the series date from 1967). Certainly this does raise some problems: to take the case of Lear again, G. K. Hunter’s 1972 edition of the play predates the intense engagement with the multiple text issue that marked the closing decades of the twentieth century. Hunter’s Lear thus offers, as was customary up to this time, a conflation of Q1 and F. But we have at least travelled some distance from the Globe edition with this package. As in the Open Source Shakespeare, SW segments the plays by scene, but here there is no option to access the whole play — one must move from one scene to the next by clicking the ‘next scene’ link at the bottom of the screen. By comparison with OSS’s use of TLNs, SW uses conventional act, scene and line number references, with the full reference being given against every single dialogue line (e.g., ‘Ham I.ii.244′). The light grey text used for these references means that they are not as obtrusive as they otherwise might be. A brief synopsis of each play is provided, together with a set of ‘character circles’ — simple diagrams which map out the relationships among the individual characters. The character circle for The Comedy of Errors is elegantly conceived and would certainly be a godsend to any student struggling to make sense of the play.

0 The look of the page is generally very clean, the text itself being displayed on a central white vertical band, with tabs above for ‘Dramatis Personae & Circles’ and ‘Play Synopsis’. On the right hand side of this band is a ‘definitions’ column, with a dark grey background. There is a ‘Hide definitions’ option, which removes the definitional material while, somewhat oddly, leaving the grey vertical band itself in place. The definitions themselves are the core of this package and the Crystals present here the rich fruits of their impressive ongoing research into Shakespeare’s language. Each word in the text thought to require explanation is clearly defined, generally by offering three related alternatives. Thus, ‘invention’, in the second line of the Chorus to Henry V, is glossed as follows: ‘invention (n.) 1 inventiveness, imagination, creative faculty’. Clicking on the word being defined brings up a list of other instances of its being used in the same sense elsewhere in the canon — in the case of ‘invention’, 17 hits (for an exact match) from seven different plays and from five different sonnets. A particularly helpful feature is that these individual hits, as displayed, include a few words of context, as, for example: ‘AYL IV.iii.35 [Rosalind as Ganymede to Silvius, of Phebe's letter to Ganymede] such giant rude invention’. Clicking on the line reference brings the user to the line in the playtext itself.

0 It is possible to gain direct access to the material that lies at the core of the package by going to the ‘Glossary’ section of the site. Here, all of the words for which definitions are provided are listed in alphabetical order. Clicking on any given word provides a complete list of all of the definitions associated with it. So, for example, ‘absolute’ is defined in 7 separate senses, ranging from ‘perfect, complete, incomparable’ to ‘curt, peremptory, blunt’. Again, clicking on a specific individual definition provides a list of instances of the word being used in that particular sense. Some Glossary entries only cross-reference to other entries. Thus, for example, ‘tail’, which might, one would have thought, have had its own separate entry (given that it is sometimes used as a sexual pun), simply crosslinks to ‘come cut and long tail’.

0 Because SW has been produced in collaboration with a commercial publisher, there is a certain amount of hard sell on the site. A ‘Buy the Book’ link directs users to Amazon, where they can purchase the book version of Shakespeare’s Words, and there are also links to purchase other publications by Ben and David Crystal. A Paypal link solicits contributions for the upkeep of the site, with any surplus funds generated rather nicely being pledged ‘to help theatre companies engaged in Shakespeare productions which receive no government subsidy.’ In fairness, however, Penguin and the Crystals are providing a lot of free material on this site and the central resources — including the Glossary and the text of the plays — are completely free of advertising of any sort, presenting the user with clean, uncluttered screens. Taken all in all, SW is an excellent resource: while some of the playtexts are indeed now rather dated, they do still present a high level of scholarship; the Glossary function is a scholarly and well conceived tool.

0 Neither the Crystals nor Eric Johnson are editors, so, as we have seen, other people’s texts sit at the heart of their websites. Their projects might be contrasted in this regard with the Internet Shakespeare Editions site (ISE; http://internetshakespeare.uvic.ca/index.html), first conceived by Michael Best as far back as 1996, when he had ‘an ambitious vision: . . . to create a website with the aim of making scholarly, fully annotated texts of Shakespeare’s plays freely available in a form native to the medium of the Internet’. Best’s aim has been to provide facsimiles of all the earliest printings of the texts; searchable transcriptions of those texts (essentially diplomatic editions); modern, edited versions of each text; ancillary materials, primarily connected with the plays in performance; materials collected under the general category ‘Life and Times’; and a reference section. The intention is that all of the editions and transcriptions should be peer reviewed.

0 The supporting materials provided on the site are certainly useful. The Julius Caesar pages, for example, include a total of almost 100 performance records for the play, ranging from Georges Méliès’ five minute Shakespeare Writing Julius Caesar (1907), to a 2009 Victoria Shakespeare Society (Canada) performance of the play itself. Inevitably, there is, at times, a slightly arbitrary feel to this material. It seems odd, for instance, that two productions of Julius Caesar by Shakespeare on the Saskatchewan are logged here, but none by the Royal Shakespeare Company. Moving beyond the performance materials, things get patchier still. The Julius Caesar pages offer a collection of images relating to the play, but these are all derived from an early Charles and Mary Cowden Clarke edition. Doubtless more images will be added in time, but it is hard to see why this particular set of illustrations should be the first (and so far the only) one to be included. The problem here would seem to be that materials are not (yet) being collected in a coherent fashion — presumably what turns up on the site is a function of the interests and resources of individual editors (or, perhaps, the greater editorial team). It is clear that a more focussed approach, with clear goals, is needed. There may also be opportunities here for opening the site to the collection of materials through crowd-sourcing, filtered through the editorial mechanisms.

0 These materials are, of course, relatively peripheral: the heart of this project is the set of texts it provides. It is a little difficult to gain an overview of the exact current state of this aspect of the project just by looking at the site itself. Table 1 thus collates information from the individual text pages, to provide a snapshot of ISE as it stood at the beginning of 2010.

0 The breakdown of materials that follows is based on the information in this table. As regards the facsimiles, for each Folio play the relevant pages of the Brandeis University and State Library of New South Wales copies of F1 have been incorporated into the site, as have the New South Wales copies of Ff2-4. As far as the texts with early quarto (or octavo) editions are concerned, facsimiles of at least one of the earliest editions have been provided for about half of the texts. Transcriptions have been included for all F1 texts, though 15 of them are presented without indication of who has compiled them, and 27 of the F1 transcriptions have yet to undergo a process of peer review. For texts with editions earlier than F1, transcriptions of at least one early text have been provided, though only 5 of these 31 transcriptions have been peer reviewed. Only 5 texts in total can be said (at the time of writing) to be fully complete: with all relevant facsimiles loaded, transcriptions prepared and peer reviewed, and a modern edition prepared and peer reviewed. These are: As You Like It (edited by David Bevington), Cymbeline (Jennifer Forsyth), Julius Caesar (John D. Cox), Venus and Adonis (Hardy M. Cook) and The Tempest (Brent Whitted). It is notable, of course, that all of the plays listed here first appeared in F1, and so have no associated early quarto texts — they are, in this sense, among the easier texts in the canon to edit and prepare. A modern edition of one further text — Romeo and Juliet, edited by Roger Apfelbaum — has also been produced and peer reviewed, though some of the transcription materials still await checking.

2 The modern editions are one of the most important elements of this project. Though only six modern texts are currently available, many of the other editors have already been appointed. There are rather more junior scholars on the list than might be typical of a high profile print edition, though this is not necessarily a bad thing in itself — and, certainly, the project as a whole is being overseen by an impressive board of senior scholars. The modern texts that are currently available are well conceived and well presented. As with OSS, plays can be loaded scene by scene, or in their entirety. Again, tlns are used here, with the numbers being provided every five lines, though it is slightly hard to see, from the texts themselves, what is and is not being counted as a line (possibly this may be a function of how the browser handles prose and partial lines). Two sets of tools are provided: annotations (basic or advanced) and collations. Annotated elements in the text are underlined and a single click brings up a text box with the annotation. The notes themselves primarily consist of brief explanatory glosses, though the longer annotations are helpful, as, for example, David Bevington’s 360 word note on ‘Robin Hood of England’ in 1.i of his As You Like It, which discusses, among other things, the source texts for the Robin Hood story.

0 One of the aspects of ISE that indicates the greatest potential of the project is the way in which textual variants are handled. In Bevington’s As You Like It, variants from more than 50 editions are registered, ranging from F1 to Michael Hattaway’s 2000 Cambridge edition. The different editions consulted are colour-coded and they can be accessed selectively: all variants can be displayed; an individual editor’s variants (including those introduced in ISE) can be displayed; any combination of editors can also be selected. It is a wonderfully powerful facility. Or, at least, it would be, if it were used consistently across the range of texts being offered. John D. Cox’s Julius Caesar offers variants for just F1, the 1691 quarto, Rowe, Steevens and Capell — a much more restricted set of texts than Bevington has consulted. Brent Whitted’s Tempest appears not to offer any collations at all.

0 These inconsistencies serve to highlight one of the core problems with ISE. It is a worthy project, but it is very much a work in progress. Using it is a bit like wandering around the Sagrada Familia — a fascinating structure, to be sure, but one despairs of ever seeing it in its finished form. In a sense, of course, there is nothing new in this. Michael Best and his team are attempting to create something that is, in fact, bigger than a standard print edition, and print editions themselves have never been built in a day: Arden 1 took 32 years to complete, the New Cambridge 45. There is a difference here, however: print editions have released individual texts when they have been in their finished state; with ISE whole sections of the site seem to be offered to users in a provisional state and, frustratingly, it is very hard to tell quite how complete any given element of the site is thought to be by Best and his team. In theory, texts included in the ‘Library’ have been fully reviewed, with an ‘Annex’ section being reserved for works that ‘have been carefully proofread, but have not yet had a full scholarly review’. In practice, however, all textual materials appear simply to have been loaded into the Library section of the site, regardless of their state of (in)completion.

0 Taking a broader view of these issues, it is important, of course, to recognise that, in a digital world, all texts are provisional — that is, after all, the great beauty of electronic textuality. But we still need to know whether, for example, Whitted’s Tempest and Cox’s Julius Caesar will at some point in the future eventually have the same wonderfully rich range of collations as Bevington’s text. Even the ‘Editorial Guidelines’ do not really answer this question, as they simply indicate that editors should ‘Collate subsequent editions of importance, particularly twentieth-century editions, whenever, but only when, a reading is offered which you deem worthy of serious consideration along with the one you yourself have chosen.’ Does Bevington’s text, by these lights, represent a particularly generous reading of the guidelines provided by the editorial board, or does his edition indicate a ‘gold standard’ that the other texts will eventually reach? It is very hard to answer this question on the basis of the materials provided on the site. For scholars and teachers, it is precisely the dynamic quality of online resources that makes it so important to have clear statements about the state of progress for a resource such as ISE.

0 Perhaps in the end, though, the moral of the story here may be that, as ever in life, you pay (or don’t pay) your money and you take your chance. In the nineteenth century 80 guineas bought you a sumptuous, limited edition, multi-volume Works of Shakespeare — a thing of real beauty, edited by one of the foremost Shakespeareans of the time. A shilling bought you the complete works in miniscule print, in a paper wrapper — essentially a disposable text, with no indication of who had carried out the editing (or of when the editing had been done). If you want free Shakespeare on the Internet, perhaps you have to put up with texts that are, similarly, somewhat unsatisfactory in one way or another, being (to a greater or lesser extent) out of date, uneven or incomplete. ISE has been supported by the University of Victoria and the Social Sciences and Humanities Research Council of Canada, but it lacks the kind of financial resources that publishers have often, in the past, been willing to invest in large scale Shakespeare projects, such as when Oxford University Press employed Stanley Wells and Gary Taylor as full time editors for several years, as they worked to produce the Oxford Shakespeare. Lacking these resources, it is not hard to see why ISE feels so much like a slightly disorganised building site, and why relatively few texts have been brought to completion over the course of its existence.

2 Does this mean, then, that, if we want a fully useable electronic text of Shakespeare, we must wait for the latest incarnations of the big scholarly editions to be released in electronic form, probably having to pay for access to them (and hopefully paying a lot less than Thomas Nelson thought we should pay for electronic access to Arden 2)? For scholars, again, I think, the answer to this may well be ‘yes’. For the general public, less concerned with textual niceties, the response may be rather different, as it was in the nineteenth century, when vast numbers of the shilling Shakespeares were sold to general readers (it is notable, in this regard, that more than a million users have already connected to OSS, probably untroubled by the antiquated text that lies at its heart). It is important, however, not to lose sight of the real value of all three of the sites under review here. Johnson’s concordancing facility is excellent, as is the glossarial material included in Shakespeare’s Words — and we must bear in mind that neither Johnson nor the Crystals ever set out to provide a wholly up-to-date text. ISE offers us a vision of what editors can now do with the electronic text — providing scholars with something that really goes beyond the limits of the print edition. ISE’s best texts are very good indeed; the problem is that the site is developing slowly and unevenly, and not enough information is being provided to users on the current state of the individual segments of the project. The best elements of these three sites provide a nice indication of where new technologies can take us — and it is a world that even Halliwell-Phillipps, visionary pioneer though he undoubtedly was, could probably never even have dreamed of.

0 General comments on this critical review may be linked to this page.