It does not require a particularly savvy reader to parse Richard Hatfield’s intent to insult “all women” in “All women have vertues noble & excelent” (18v of the Devonshire Manuscript, [Baron 1994, 335]). The metrical lines praise women for their fidelity, while the new lines suggested by virgules damn all women, imagining each woman’s sole virtue as her ability to please her husband, a virtue which none possesses. It is a commonplace that this sort of easy dismissal, or indeed outright hostility towards women, often guised as humor, is still endemic to online social spaces, such as the Wikimedia foundation's suite of projects. This paper takes as its starting place Greg Crane’s exhortation that there is a “need to shift from lone editorials and monumental editions to editors ... who coordinate contributions from many sources and oversee living editions”  [Crane 2010]. In response to Crane, the exploration of the “living edition” detailed here examines the process of creating a publicly editable edition and considers what that edition, the process by which it was built, and the platform in which it was produced means for the production of editions that support and promote gender equity. Our study draws on the scholarship about the culture of the Wikimedia suite of projects, the gendered trolling experienced by members of our team in the production of the Social Edition of the Devonshire Manuscript in Wikibooks (see figure 1) [Siemens et al. 2012a][1], and interviews with our advisory group. Wikibooks proved a challenging environment for edition production for both cultural and technological reasons, reasons which might incline scholars to dispense with the platform. The collaborative space opened up by social media while not inherently feminist, and indeed often openly hostile to women is, however, one of the central online spaces where the public turns for information. The Wikimedia suite of projects are so important to the contemporary circulation of knowledge, that the key for feminist scholars is to encourage gender equity in social behavior, credit sharing, and knowledge organization in Wikimedia, rather than abandon the Wikimedia suite of projects in favour of more controlled collaborative environments for edition production and dissemination.

Despite Stephen Nichols’ call to “dismantle the silo model of digital scholarship”  [Nichols 2009], many electronic scholarly editions, like print editions, continue to exist as self-contained units that do not encourage interaction with other resources. Furthermore, many editions do not actively encourage or facilitate interaction among the communities of practice they serve, or even among those who have the most knowledge to bring to bear on the edition.[2] The scholarly community is now producing tools for crowdsourced transcription and annotation, but the community of users that has developed around the Wikimedia suite of projects has anticipated (and, we speculate, inspired) the development of these tools. Acknowledging the dedicated community already engaged in Wikibooks, we sought to discover Wikibooks’ affordances for editors, scholars, and students. As we investigated and participated in Wikimedia’s community, we experienced what research has already suggested: Wikimedia, the go-to resource for many when seeking information, is a disturbingly gendered space.

Wikimedia is a non-profit foundation, most famous for Wikipedia. The foundation itself is very small — it has 117 employees (up from 26 in 2010) (Wikimedia Foundation), responsible for the foundation’s management, fundraising and technological development. The content of the projects is contributed and moderated by volunteer editors. In order for Wikimedia projects, including Wikipedia, Wikiquote, Wikibooks, and Wikisource, to be feminist they need not only address issues of import to women (although Wikimedia’s dearth of information traditionally of interest to women is indeed a feminist issue), but also need to address how behavior and credit in online space structure the creation, design, and content of projects and pages within Wikimedia.

Since Wikipedia and Wikibooks are often a first, and occasionally only, stop for many members of the public when searching for information, it is incumbent upon scholars, as members of a specialist community often supported by public funds, to engage with the platform. In building an edition on the principles of open access and editorial transparency, we have integrated scholarly content into the environments maintained by the editorial communities already existent in the Wikimedia suite of projects, including Wikipedia, Wikibooks, and Wikisource. In an experimental spirit, we extended the editorial conversation into multiple pre-existing social media platforms, including blog posts; Wikibooks discussion pages; dedicated Renaissance and early modern online community space; Skype-enabled interviews with our advisory group; and Twitter. In this paper, we will introduce the Devonshire Manuscript itself and offer a brief overview of the steps that led up to our Wikibook instantiation of the manuscript. Drawing on Jacqueline Wernimont’s argument that textual content is not the only index of a feminist digital resource [Wernimont 2013, 10], we argue that there need not be a text by or about women at the center of a publicly edited edition in order for that edition to be feminist. A social edition’s success as a feminist text in the wiki environment comes from its ability to short circuit personal sexist attacks in the online space of the edition, avoid latent sexism in the structure of information, and resist the deletion of content that is either produced by women or culturally coded as feminine. We conclude by suggesting a method of receiving credit for Wikimedia contributions, which would attract editors who otherwise might be too overextended in the offline world to be able to contribute without getting credit. If widely adopted this method would lead to a more diverse group of editors with the skills and Wikimedia editorial credibility to respond to instances of inter-editor trolling and sexism in the structure of information. By encouraging the ongoing conversation between and across online communities rather than demanding the diminution of gendered markers online, the social edition process sheds light on how digital humanists might leverage existing online platforms to meet broadly feminist goals.

Feminism is understood here as the organized effort to undermine patriarchy, the system in which men, women, and institutions engage in the persistent valuing of things culturally coded as masculine and male over things coded as feminine and female. The burgeoning of sites that store user-generated content need feminist intervention, since “Web 2.0 culture, while clearly not as exclusionary or hostile as the earlier mainframe and hacking cultures, remains at its ideological core, a masculinist culture”  [Bury 2010, 235]. A technofeminist approach, one which combines the recognition that technology comes with cultural freight including gender constructions, is better suited, Rhiannon Bury argues, than liberal, socialist, or radical feminist approaches to the persistence of patriarchal values in the Web 2.0 context [Bury 2010, 235]. Taking its cue from Bury, this paper addresses two ways that Wikimedia projects' spaces maintain patriarchal values: via outright hostility to women represented by personal attack, and institutionally through the organization of information that devalues women’s contributions to knowledge production and dissemination. Personal attacks against female Wikimedia editors on the grounds that they are women (the very definition of prejudice: hostility, dismissal, or violence against an individual based on his or her belonging to a group) often takes the form of trolling, a form of transient, though oft repeated, aggression in the interest of creating a hostile environment or provoking an angry response. The feminist organization of knowledge, attendance to the ways women are represented at the level of code, and dismantling the equity barriers that that representation may erect, can undermine the valuing of things culturally coded as masculine and male over things coded as feminine and female.

Shifting the power away from a single editor is not an inherently feminist act. Conceiving of this shift, however, leads us to speculate on what a feminist method might look like in an open-access Web 2.0 environment. The gendered version of the aphorism that “on the Internet no one knows you are a dog”  [Chow-White 2012, 7], “on the Internet no one knows you are a woman” suggests that it might be possible to avoid sexist trolling and biased deletion of Wikimedia content by concealing markers of female sex or feminine gender in Wikimedia; however, if the goal of feminist editing is to undermine patriarchy on the terms listed above, then obfuscating sex and gender cannot be a feminist editor’s goal. As the attacks against Anita Sarkeesian on Wikipedia suggest (her Wikipedia page was defaced, not for her critiques of male gamer culture (which she had not made yet) but because she was planning to make a documentary critiquing male gamer culture [Consalvo 2012]), the obfuscation of sex and gender, even where possible in the case of Wikimedia editors is not possible in the case of Wikimedia subject matter (i.e. the page about Anita Sarkeesian). Furthermore, beyond attacks against Wikimedia subjects and editors on the grounds of sex and gender, the structure of information is a feminist issue since those structures can perpetuate institutional sexism, as the quiet removal of female authors from the “American Novelists” category in Wikipedia, categorizing them instead as “Women American Novelists” in 2013 attests [Filipacchi 2013]. The problems introduced by harassment and structural sexism in Wikimedia projects is exacerbated by institutional sexism offline. Studies show that women have less spare time than men to devote to Wikimedia contributions and that the content that they do contribute is more likely than men’s to be deleted as trivial [Eckert and Steiner 2013] [Lam et al. 2011]. The social forces that exacerbate female and feminine people’s oppression offline, persist online — reinforcing and building on the behaviors and structural barriers that constitute oppression offline.

Although the social edition project predates the Wikipedia edit-a-thons of 2012 and 2013, the project editorial team endorses the #TooFew and #DHPoco editing drives. As Adeline Koh has pointed out, since the average Wikipedia editor “is a college-educated, 30 year old, computer savvy man who lives in the United States or Western Europe, it is unsurprising that the online encyclopedia has its own unconscious ideological leanings”  [Koh 2013a]. Those leanings, discussed in more detail below, contribute to an environment that is hostile to women resulting in lower quality of the topics that are culturally coded as being feminine or of interest to women [Lam et al. 2011] [Currie 2012]. The edit-a-thons, organized independently and through THATCamp Feminisms conferences, are evidence of scholars’ recognition of the inequitable nature of editorial practice and knowledge organization in Wikimedia projects. Moreover, the edit-a-thons are proof that scholars recognize that the Wikimedia projects are valued by the public as sources of information and so warrant feminist scholars’ engagement.

Scholarly editing for print has always been social, and has already been the subject of feminist analysis. In any reputable edition, acknowledgements disclose the team of graduate assistants, librarians, reviewers, and publishers whose work underpins the edition. Extending this model, Ray Siemens has called elsewhere for scholars to improve our

understanding of the scholarly edition in light of new models of edition production that embrace social networking and its commensurate tools… [to develop] the social edition as an extension of the traditions in which it is situated and which it has the potential to inform productively.  [Siemens et al. 2012b, 447]

The Social Edition of the Devonshire Manuscript has been produced by just such a team. The Devonshire Manuscript Editorial Group (or DMSEG,[3] made up of a core team of researchers at the Electronic Textual Cultures Lab (ETCL) and the University of Victoria, and distributed network of scholars, postdoctoral fellows, graduate researchers, and programmers, working with two publishers,[4] an editorial board,[5] and self-selected members of the public) is motivated by a desire to render transparent the production of an online edition of the Devonshire Manuscript by a method that privileges process over product. The expectation was that with transparency of workload and contribution would come a flattening of hierarchies, since, where there are power imbalances, the door is open to abuse, particularly of the type evinced by women’s higher service load both within the academy and without [Bury 2010]. This is not to say that where there is power imbalance sexism is an inevitable result, but it is certainly easier for personal and systemic sexism to go unchecked in a closed hierarchical editorial environment.

The Social Edition of the Devonshire Manuscript draws on a collaboratively edited manuscript to create a collaboratively edited edition. Edited and circulated by a group of women in Anne Boleyn’s court, the Devonshire Manuscript bears traces of women’s resistance to Hatfield’s jest about women’s sole virtue. Compiled in the 1530s and early 1540s by various sixteenth-century contributors, the manuscript itself is a multi-authored verse miscellany. The Social Edition of the Devonshire Manuscript fills the void that Margaret Ezell notes has been left by the “little effort [that] has been made to catalogue and reconstruct patterns in women’s manuscript texts to provide an inclusive overview of literary activities rather than isolated, individual authors”  [Ezell 1999, 23]. The editorial communities that have grown up around social media sites like Wikibooks indicate a public desire to expand our knowledge communities using the social technologies at our disposal. Using the Devonshire Manuscript as a prototype, we have devised a method that addresses the questions that a social edition raises. In a feminist context, which in this case we take to mean the resistance to patriarchy’s devaluing of all things culturally coded as feminine or female, we must ask whether feminist edition production must have a text written by women at its center in order to be a feminist project. We argue that, while the recovery of women’s history and women’s contribution to literature is a feminist aim, in the context of Wikimedia edition production, feminist methods cannot only be indexed to content, but must instead be evaluated against the behavior and the organization of knowledge during the creation and maintenance of the edition.

A social edition constitutes a collaboratively maintained research environment, in the case of the Devonshire Manuscript Wikibook, one complete with facsimile page images and a comprehensive bibliography: the material that interested users would need in order to make a contribution to the content of the edition. In addition to a general and textual introduction, the online edition includes extensive hand tables that open our paleographic attribution process to public scrutiny, witnesses that reflect the poem’s textual legacy, and biographies and genealogical diagrams that clarify the relationship between the manuscript’s sixteenth century compilers. We have also included the facsimile images of the manuscript (courtesy of Adam Matthew Digital). Providing further information and room for debate, the discussion sections on each page promote conversation on the various aspects of the poem at hand. In this way, the Wikibook edition extends the social context of the Devonshire Manuscript by providing a space for ongoing community collaboration. The Wikibook edition’s features stretch the limits of a print edition — including sheer size. Even if the manuscript facsimile pages and the xml files were excluded, The Social Edition of the Devonshire Manuscript would run to over 500 standard print pages. The social edition is reliant on the infrastructure and affordances provided by Wikimedia.

The DMSEG team had considered hosting the edition on a stand-alone site; however, in response to public interest in the project, coupled with the team’s investment in emerging public knowledge communities, we devised an editorial experiment: as a control, we produced the static authoritative version of the edition and as a variable we moved the same content into Wikibooks. We considered several Wikimedia projects and finally decided to mount our edition in Wikibooks. Even though Wikipedia has more editors, Wikibooks has more affordances that support a book-like form. With a book-like research environment as our end goal, we produced an edition in Wikibooks that is scholarly in a traditional sense but also enables citizen scholars to access, contribute, and annotate material. Wikibooks archives each change to the book, allowing us to track reversions and revisions to the text. Furthermore, under the hood of Wikimedia projects’ pages is a network of editor’s personal pages, talk pages, and edit reports which allow for the specific self-declaration of an editor’s sex or, via social cues, the inference of it.

Content culturally coded as feminine is underrepresented in Wikimedia. Shyong Lam, a computer scientist from the University of Minnesota, has noted that Wikipedia content that attracts male editors is of higher quality (using length as a proxy for quality) than the content that attracts female editors [Lam et al. 2011, 5].[6] He concludes that women are more likely to contribute to Wikipedia’s People and Arts sections than they are to Geography, Health, History, Science, Philosophy, and Religion. But, due to the relatively few female Wikipedia editors, male editors still outnumber female editors in the Arts and People by a ratio of ten to one [Lam et al. 2011, 5].[7] Lam attributes the gap to the culture of Wikipedia: women do not have a critical mass on Wikipedia, and the Wikipedia community treats them with greater hostility than it does men. Only 16% of new editors on Wikipedia are women, and, Lam found, new female editors are more likely to have their edits reverted than their male counterparts [Lam et al. 2011, 4, 3]. Women’s hostile reception has resulted in low female participation in Wikipedia editing. Their underrepresentation has skewed the content, quality, and visibility of the subjects that, due to the enculturated differences between the sexes, are of interest to women. Our experience in Wikibooks confirms Lam’s findings; in the final section we will discuss the treatment of one of our female editor’s experience with an online aggressor (a “troll”).

The Devonshire Manuscript’s structure and content, rather than the culture of Wikimedia projects informed our choice of Wikibooks as the venue for initial publication. Although the manuscript has attracted scholarly attention as an artifact of the first sustained multi-gendered writing community in English, at the time of writing there had been no authoritative critical editions of the Devonshire Manuscript.[8] The manuscript has, however, served as a source for Sir Thomas Wyatt’s poetry. His verses have been transcribed and published by A. K. Foxwell (1914), Kenneth Muir (1947, 1949, 1969), and Patricia Thomson (1969) in their respective editions of his work, but until now no scholar has transcribed the manuscript in its entirety. This “author-centered focus,” Arthur F. Marotti has argued, “distorts [the] character” of the Devonshire Manuscript in two ways: “first, it unjustifiably draws the work of other writers into the Wyatt canon, and, second, it prevents an appreciation of the collection as a document illustrating some of the uses of lyric verse within an actual social environment”  [Marotti 1995, 40]. Scholars focused on Wyatt’s poems in the manuscript until the middle of the twentieth century, when Raymond Southall (1964), John Stevens (1961), Ethel Seaton (1956), Richard Harrier (1975), and Heale (1995) began to assert its value as a record of court life and of women’s editorial practices. These scholars acknowledged the manuscript’s significance as a product of multiple authors representing their private and public concerns in the social context of Henry VIII’s court. While the significance of the manuscript as a source of Wyatt’s poetry was by no means diminished by this new focus, Helen Baron’s identification of the hands in the Devonshire Manuscript (1994) has increased scholarly interest in Mary Shelton, Margaret Douglas, and Thomas Howard’s contributions.

Generically, the Devonshire Manuscript itself is a true miscellany: including all creative textual works — complete poems, verse fragments and excerpts from longer works, anagrams, and other ephemeral jottings — the manuscript consists of 194 items. It is the work of “educated amateurs,” a coterie that included members of Anne Boleyn’s entourage. Margaret Douglas, Thomas Howard, and Mary Shelton entered the majority of the original work in the manuscript. Of the identified hands, Mary Fitzroy, wife of Henry VIII’s illegitimate son Henry Fitzroy, is the only member of the identified group who only copied extant poems into the manuscript [Baron 1994]. Even amongst the copied text, one finds many features that suggest the personal engagement, immediacy, and spontaneity of this group.

Characterized by Nicola Shulman as “the Facebook of the Tudor court”  [Shulman 2011, 142], the Devonshire Manuscript is much more than an important witness in the Wyatt canon; the manuscript is also, in Colin Burrow’s estimation, “the richest surviving record of early Tudor poetry and of the literary activities of 16th-century women”  [Burrow 2009, 3]. Throughout our process, we remained mindful of Marotti’s assertion that “literary production, reproduction, and reception are all socially mediated, the resulting texts demanding attention in their own right and not just as legitimate or illegitimate variants from authorial archetypes”  [Marotti 1993, 212]. The Social Edition of the Devonshire Manuscript has published the contents of the manuscript in its entirety, moving beyond the limitations of an author-centered focus on Wyatt’s contributions in isolation, to concentrate on the social, literary, and historical context to situate the volume as a unified whole.

The advisory group noted the particular way that the Social Edition of the Devonshire Manuscript addresses the role of women in the production of the manuscript itself. As we noted above, the hitherto unpublished manuscript is the first example of men and women writing together in English. A DMSEG advisor commented,

“one of the things [I have been] thinking about [is] again bringing women writers up to the front, and showing them as a part of a network of writers who were all sort of cross-pollinating and doing all this stuff in a way that’s not ghettoizing them, which is what was happening [in scholarship in general].”

The rise of New Historicism and of interdisciplinary studies has sparked new interest in previously overlooked early modern texts, supporting the feminist goal of recovering women’s history via translations, treatises, diaries, memoirs, letters, and even crafted objects. “Women’s translations are being treated as important literary and cultural texts,” explains Micheline White,

and women’s letters, gifts, and needlework are recognised as important objects of study. Work focusing on individual women is complemented by essays and books that position women writers alongside their male counterparts and that incorporate women's texts into larger literary, cultural, and historical narratives about Tudor England.  [White 2000, 488]

Such enquiries have brought many texts written by women to the fore, hastening a shift in scholars’ processes of canonizing writers and works. This new critical focus has encouraged researchers, as Sara Jayne Steen suggests, “to re-imagine a manuscript culture that included writers of both sexes”  [Steen 2004, 147]. As this new focus of literary study drew scholars’ attention to the way court lyrics could reflect the interactions of poetry and power in early Renaissance society, enquiry began to extend beyond the consideration of canonical texts and privileged genres. New Historicist feminist scholarship reclaimed the importance of the Devonshire Manuscript to women’s history. That said, even though its importance to the women’s history canon is reason enough to put the Devonshire Manuscript at the centre of a feminist analysis of social editing practices, the tenants of feminist social editing cannot limit content to women’s history, but ought to be applicable to all editorial content. Any collaboratively produced edition’s equity work originates in affordances of the editorial platform, the structure of information, and the behaviour of collaborators.

Particular platforms may encourage equitable behaviors while others make it easy to persist in personal and systemic sexism. The question of platform and encoding-specific affordances was of particular interest to the DMSEG, since even though the edition resides in Wikibooks, it took a host of social media platforms to coordinate the encoding and review process. The content of stakeholder and partner interactions outlined below was facilitated by multiple social media platforms, including blog post comment threads; Wikibooks discussion pages; Iter’s dedicated Renaissance and early modern online community space; personal interviews via Skype; and Twitter conversations. We have found that each social media platform attracts different stakeholder groups and enables specific types of interaction. Employing and participating in various platforms both alerted us to different priorities across platforms, and forced us to think through how we might create a multispatial experience for safe, productive, and equitable interactions. In the interest of refining the process and expounding on its utility for collaborative editors in the Web 2.0 environment, the ETCL-based members of the team used a combination of methods to gather data on the social edition building process. We conducted qualitative interviews with advisory group members, none of whom had extensive experience editing in Wikimedia, to gather their perspectives on the content of our evolving and fixed editions, as well as on issues of credit, peer review, and collaborative editing. We also enumerated interaction in Wikibooks. Furthermore, we invited feedback via Iter’s social media space, Twitter and guest blog posts.

As with print facsimile editions, the accurate transcription of the source is at the heart of the social edition. The transcription of the Devonshire Manuscript predates public crowdsourced transcription projects such as Transcribe Bentham. Transcription in an online space would have allowed interested readers to follow, or even to contribute, to the project and such transcriptions would have been in keeping with the ethos of the project. The closed transcription process does not necessarily produce a feminist organization of knowledge and or ethos of collaboration. At the time of transcription, however, the team did not have permission to post or circulate the manuscript facsimile, so they had to produce the transcriptions without public input. Two team members worked from paper copies of the Devonshire Manuscript and produced independent transcriptions. In general, their transcriptions were in accord with one another. To settle any conflicting transcriptions, Ray Siemens returned to the British Library in order to compare the transcriptions to the manuscript itself.

In order to ensure an encoding that would be useful to scholars outside of the project, the team then encoded the text in TEI, the mark-up language of the Text Encoding Initiative. In order to keep the editorial and encoding process transparent, the Wikibook edition includes links to the baseline xml-encoded transcription. Thus, in addition to being able to use the xml for their own projects, readers can see the editorial choices the TEI allowed encoders to preserve. Other digital humanists may continue working with the TEI-encoding document, allowing the project to evolve in ways that could not be anticipated at the outset. With the firm foundation of documented encoding, all those working with the document can refer to, build on, or adapt the project’s foundation.

Text encoding, like other digital editorial interventions, merits explicit feminist reflection especially where it may shape what edition users can, through search or programmatic retrieval and counting, learn about the text. While wikicode, which underpins Wikimedia projects, has no gendered hierarchy built into its standard tag set, at the time of encoding the TEI certainly did. TEI, as Laura Mandell and Melissa Terras have argued, with its reliance on ISO standards, relegates women to second standing [Mandell 2013] [Terras 2013]. ISO standards allow four values for sex: 1 - male, 2 - female, 0 - unknown, and 9 - not applicable. The TEI changed the standard in 2013 to allow for locally defined values (Simone de Beauvoir need no longer have her work literalized, by being represented as a member of <sex value=“2”>, or so the joke goes). The encoding of the manuscript in TEI P4, which was converted to TEI P5 in 2007, predates the change in the values for @sex and the <sex> element. The germ of the Wikibooks edition was created by transforming the TEI into wikicode, which has no sexed markers, via XSLT transformation, therefore any (now outdated) use of @sex or <sex> does not appear in the code of the Wikibook’s edition of the Devonshire Manuscript. “XML and SGML,” Wernimont reminds us, can be read “as political rather than neutral tools”  [Wernimont 2013, 11]. That said, does wikicode’s avoidance of the ISO standard refuse to recognize the value of sexed difference or simply erase evidence of women’s participation? In the final analysis the information organized for the human reader announces the value of women’s participation in the production of the original manuscript, but does not let the programmatic reader algorithmically determine, for example, how many items in the manuscript were entered by women, although it does let that reader, armed with an algorithm, list, for example, all the poems entered in Mary Shelton’s hand, and, furthermore, which of those poems are her original creation. Another layer of encoding could be added to make the sex of the Devonshire Manuscript compilers clear in Wikibooks, but it is not clear that focusing on the sex of the contributors, whether they be historical or contemporary, is a central, or even requisite criterion of a feminist editorial methodology in Wikibooks.

Feminist social edition building does not rely solely on thoughtful encoding of sex and gender; it relies on the behavior of contributors, readers, and critics in and around the edition. It is this behavior, rather than simply the publication of material by female compilers, that makes an edition a feminist work. The formation of the social edition’s advisory group, in particular, provided a unique opportunity to invite potential critics to shape the process and the products associated with the social edition, and to negotiate and even encode the changes that they would like to see as the edition evolved rather than have them critique a fixed and final edition. If a feminist edition is one that guards against the personal and institutional sexism that may (and indeed does [Bury 2010]) filter into the online world from the offline world, the members of the DMSEG working in the Electronic Textual Cultures Lab (ETCL) at the University of Victoria had to work out which tools and procedures for collaboration, credit, and social engagement would let us iteratively test the feminist affordances of the social edition.

The edition-building process situated our text at the intersection of academic and wiki culture. As we traversed this new multidisciplinary ground, we sought advice and responses from a variety of sources. Procedurally, before moving the edition into Wikibooks, the ETCL team prepared a static digital edition of the manuscript. This static edition served as a base text, to which our international advisory group of early modern and Renaissance scholars could compare the Wikibooks as it evolved. Finally, the ETCL team moved the wikicode version of the manuscript into Wikibooks where any member of the DMSEG could edit it. We received feedback from the advisory group via Skype interviews and Iter, from the scholar and citizen community via Twitter and blogs, and from the Wikibooks community via Wikibooks itself. In addition to informing and instructing the ETCL team on the early modern content of the edition, the advisory group also offered their opinions on our method.

At worst, we expected a clash of interpretation between the advisors and the public.[9] What transpired was not a clash, however, but rather an instance of trolling that is in keeping with reports of the sexism and drive-by trolling of open Web resources and comment threads. In light of our experience of the iterative production of the social edition, we argue for the importance of incorporating various social platforms and venues that enable conversation across previously divergent lines of knowledge production in order to mitigate the trolling that tends to repel female contributors, thus decreasing the likelihood that those resources will contain quality information of interest to women and feminine people.

The advisory group was unanimous in their assertion that the Wikibooks platform challenges our traditional means of assessing an edition’s authority while facilitating the type of conversation that peer review is meant to embody. An advisor remarked, “the main advantage [of an evolving edition] is the openness to further corrections and improvements — of the introduction, the texts, and the commentary.” Not all of the advisors, however, were interested in opening the edition to annotations by graduate students or members of the public, no matter how equitable our goals. As an advisor asserted, “there are very few people qualified to read this manuscript and say anything I would want to read.” This comment points to the divide between some academics and the public: the scholarly community has produced authoritative editions and other resources, such as the Oxford Dictionary of National Biography and Oxford English Dictionary, however, the cost of accessing those authoritative resources has pushed members of the public to platforms that let them share and circulate knowledge for free. The content in Wikimedia platforms only costs time to compose and consume. The divide between two spheres, academic and public, who make and share knowledge within their own spheres, but who do not collaborate with one another in Wikimedia, is wide. Furthermore, both groups may not act in ways that support equity. The key is to push back against the definition of qualification, as we will see, in both communities to find out where institutional prejudice may adhere from the offline world. Within the scope of our study scholars did not stop or even comment on issues of trolling because they did not have the time to gather the expertise and wiki credibility that it takes to be a wiki editor. Feminist scholars would be empowered to contribute to Wikimedia projects by receiving credit for that work. If scholars did contribute they would be less likely to dismiss Wikimedia projects, however, as it stands report of trolling and issues of credit like the ones listed here, might simply prevent scholars from contributing to Wikimedia projects. Indeed, female scholars are subject to the same brevity of free time that is endemic for most women outside of academia, so, as long as scholars do not get credit for Wikimedia work and have to do it outside of their regular work in the knowledge sector, the less likely we are to close the credibility gaps that exist between Wikimedians’ editorial skill and scholars’ faith in Wikimedia projects’ accuracy and authority.

The advisors’ Skype conversations created a rapport, with accompanying civility, between the advisory and ETCL-based members of the team. The next step, however, was to get the advisors to trust one another, to make it clear that there were some people behind the edition that were “qualified to read the manuscript” and contribute with authority, thereby closing the credibility gap that dissenting members of the advisory group brought to the process. Following the Skype consultations, the advisory group continued the conversation in Iter’s social media space — a Drupal installation that shares many features with Facebook. Users each have their own profile pages, and may join groups, send email, and blog. The asynchronous nature of the posts made it difficult to sustain a conversation, but in response to our group’s review Iter is redesigning how it notifies users of new comments. The group was collegial — there was no trace of the personal or institutional sexism found in the Wikimedia space.

The ETCL-based DMSEG’s use of Twitter to find out who might be interested in joining the editorial venture was more fraught. We furthered our social media interactions via Twitter, where twice weekly we tweeted out poems from the Devonshire Manuscript Wikibook.[10] We received feedback, from the ardent support of Tudor avatars to the more critical responses of academics. To our surprise, we found a thriving community of Henrician avatars on Twitter, including members of the public to tweet as Anne Boleyn, Thomas Wyatt, Mary Shelton, and Margaret Douglas. We thought that they might be interested in reading and contributing to the Wikibooks edition. It turned out, however, that the people behind the avatars were more interested in the roleplay and social interaction that Twitter makes possible than they were in editing. The ETCL team, via the Devonshire Manuscript Twitter account was invited to join Twitter picnics in which the Henrician avatars met at an appointed time to tweet to one another. In this case the affordances of Twitter were to our advantage. The Wikibook space would not necessarily have been a place that supported roleplay and socializing, but Twitter offered a space for that type of interaction, leaving the Wikibook comment threads and talk pages free from the pretense of people professing to be the original contributors to the Devonshire Manuscript. We received more constructive feedback from scholars. Where Lady Madge Shelton — a Twitter avatar of Devonshire Manuscript contributor Mary Shelton — may write “@Devonshire_MS You know you have my heart, love. Thou art my life’s ambition. Xoxo #FOLLOWFOLLOW”  [Shelton 2012], more helpful questions were raised with tweets like Andy Fleck’s question, “Quoting Wyatt. In what context?”  [Fleck 2012], or William Boyle’s comment on regional dialect [Boyle 2013], or Chris Shirley’s curiosity about technical and legal matters: “By the way, would be very interested in the IP issues involved in publishing the edition online. Any notes?”  [Shirley 2012] (see figure 2).

Perhaps predictably, academic and Wikimedia culture do not easily align. In the current academic environment, job promotion and security rely on tangible records of service. The inability to receive credit for editing in Wikibooks may deter even the most interested feminist scholars from contributing to Wikimedia projects. As one advisor noted,

perhaps some day, probably in the next generation, people … won’t be as worried about [credit] as we are. If it becomes a question that, if it’s tenure-related, you have to prove authorship of X amount of work … if that suddenly is adjusted so that tenure, peer evaluation, and peer review, becomes something that it isn’t right now, something more reflective of the way we’re doing our research work, then it may be that people are less concerned about who gets credit for what, or how you approach the idea of collaborative research.

The more fixed structure of the academic credit system is at odds with the evolving frameworks of projects like the Social Edition of the Devonshire Manuscript. As one advisor remarked, “I think there’s way too much focus on the end product, there’s less attention paid to preserving and sharing the process that leads up to the end product. I would like to see that done a bit more often.” In the ongoing nature of most digital projects inheres a dilemma: how do we assign and receive credit for work that may never be completed in a traditional sense?

Frequent editors gain recognition for conscientious Wikimedia editing by being granted more administrative power within the system, power that their fellow editors vote to assign them. We would like to see scholars engage in this process to make their knowledge more readily available to the public. Members of the scholarly community, especially graduate students and untenured faculty, however, need to be able to account for their contributions. Female faculty are more likely than their male counterparts to engage in service work to the detriment of their academic careers [Misra 2012, 318]. It is important, therefore, to frame social editing as scholarship, rather than service, even if it does have a significant community outreach (and therefore, some might argue, service) component. Furthermore, we need to provide a mechanism that preserved contributors’ anonymity within Wikimedia, but let them point to the amount of work that they had done for the purpose of tenure and promotion. The Magic Circle (a visualization tool for assigning credit by showing contributions to every page in pie chart form) included contributions for the discussion pages (a space appended to each Wikibook page where users can discuss pending changes or revisions) (see figure 3). In an effort to extend methods for assigning credit, colleagues at the University of Alberta’s Canadian Writing Research Collaboratory and School of Business donated the code that underpins their Magic Circle to the social edition project.

The Magic Circle visualizes the nature and extent of collaborative editorial contributions in a wiki environment. Drawing on expertise in business, design, and computer science, the Magic Circle’s core development team articulates the need for this type of visualization: “there are various ways in which editors can make contributions to a wiki page, e.g., one editor may add new content, another editor may reorganize the text, and a third may remove redundant text to make the page flow better. We believe that a method for estimating editors’ contributions should capture these various authorship categories”  [Arazy et al. 2010, 1167]. The Magic Circle helps make visible our conviction that starting a new page or originating material are not the sole useful contributions to the editorial process.

The Magic Circle gives us the opportunity to assign credit for important editorial work that extends beyond the creation of original content. Discussion and feedback are central to scholarly revision. A print edition, however, often only acknowledges these forms of labor with a line or two on the acknowledgments page. Our initial plan was to visualize user-defined major and minor edits, as well as contributions to the discussion pages. For example, the Magic Circle lets us include the tips from our discussion pages offered by Wikimedia editor Jomegat. Jomegat joined in the discussion but did not edit any of our pages directly. S/He did, however, offer advice on the finer points of importing content from Wikipedia pages [Jomegat 2011]. In the final analysis, we decided not to import content — Jomegat’s suggestions helped us to refine our own thinking, and including Jomegat’s contributions in the Magic Circle adds to the record of our decision-making process. Ideally, this record will help anyone who considers importing content into the edition to see how we have addressed the issue in the past. Jomegat’s intervention and input on this topic was very helpful and, we argue, deserves credit.

The Magic Circle also lets Wiki editors point to the work that they do to help moderate the tone of the Wikibooks editorial discussion and who make the subtle changes that cumulatively fight institutional sexism in the organization of information. Controversy and debate are key to the maturity of a Wikimedia project [Currie 2012, 244], and yet moderating and contributing to debate, which are not so integral to production of a print edition until the peer-review stage, are not usually assigned special credit. In the production of a social edition, however, debate, mediation, and synthesis are so important that they deserve special credit. Contribution to feminist edit-a-thons which, like the ones organized by THATCamp Feminisms unconferences and Brown University’s edit-a-thon to improve articles about women in science in commemoration of Ada Lovelace’s birthday, should not just happen as the additional labor of activism, but rather work in the interest of public which deserves credit [Koh 2013b] [Winston 2013]. The Magic Circle gives edit-a-thons participants visualizations of their labor that they can point to when seeking credit in the academic workplace.

Addressing issues of credit alone is not enough to make the Wikimedia suite a feminist editorial space. Sexist and racist trolling are persistent problems in the Wikimedia suite of projects, to which the members of the DMSEG were as susceptible as other editors. For example, in December 2011, one of the ETCL team members, editing under the user name Cultures4, was subject to sudden abuse on her personal discussion page. The trolling user, Tyrone Jones 2, made sexist comments against the member of our team, with an aggressive and racist tone, on both her page and on others’ talk pages. Nevertheless, although discussions in Wikibooks are occasionally fractious, the Wikibooks community remains dedicated to the site’s integrity: within a day of the attack on our team member, Jomegat had deleted the offending user’s Wikibooks account (although further investigation has shown that this user is a recurring menace in Wikibooks).

Like a private wiki community, Wikibooks has its own social conventions, which may be at odds with the behavior that would encourage behavior culturally coded as feminine. As one advisor stated, in a wiki

you don’t necessarily want to go in and intrude without permission on somebody’s entry proper. You want to actually be able to work through it in the Talk section, and then from there … you introduce yourself into the environment, you offer suggestions, you point out where things may or may not gel with what you think… from that point you engage with the actual editing on the page.

The other advisors offered similar sentiments: they wanted to discuss before they revised.

We discussed the community’s talk page norms with Panic, another editor who has taken an interest in our edition. He told us that Wikimedia editors do not use talk pages in the way that our advisors wished. Panic claims, “[p]eople will only use (write) into talk pages to express discontent about something, clear some controversial contribution or gather support for some major change”  [Panic 2012]. Furthermore, he says, in a sidelong critique of Wikimedia’s hostile environment “[o]ne thing that I always have in mind is that we are all volunteers so I try to balance criticism with praise for work well done”  [Panic 2012]. In short, the Wikibooks discussion pages are comprised of more personal commentary than editorial suggestions. The Wikibooks discussion pages are predominantly venues for editors to offer one another personal support rather than to discuss Wikibook content. The potential for abuse, however, is high, and the hostile environment created by the type of trolling the DMSEG experienced may deter otherwise interested editors from contributing.

Thus, rather than relying on the discussion pages for editorial decisions, we made the most substantive changes in Wikibooks based on our Skype and Iter interactions with our advisory group. Although our hope had been to have the advisors edit directly in Wikibooks, some of our advisors found the technological threshold for contributing to Wikibooks too high and the environment too hostile. We found that it was more practical to have the ETCL team make the proposed changes in the Wikibook. We responded to the advisors’ recommendations in near-real time, adding navigation menus and images that the advisors suggested through our ongoing consultation. It happened that we needed many avenues for editorial conversation in order to foster the sense of a community that, as one of our advisors noted, is “virtually there, as if everyone is crowded around a page, putting their two cents in on matters great and small.” Multiple social media platforms facilitate social editing, whereas relying on one single communication platform (such as Wikibooks alone) may impede the success of an evolving social edition with feminist aims.

We started the Wikibook initiative driven by curiosity about the new knowledge communities that have sprung up around social media, with the expectation that a process-driven approach could keep feminist methods at the fore. Process is key, and we certainly had to refine ours to meet multiple communities’ needs. Our short-term goal was to spark conversation around the Devonshire Manuscript, but our long-term goal is to work toward a model for preserving and disseminating our cultural heritage where it will be seen, taken up, and expanded by both academic and citizen scholars, in ways that increased the content culturally coded as of interest to women and feminine people.

As previously noted, Shyong Lam has identified both the systemic sexism leveled against Wikimedia contributors who identify themselves as women and the community’s ambivalence about topics that are culturally coded as feminine. There are, of course, many women who edit in the Wikimedia suite of projects who do not identify their sex. Non-identification might initially seem like the solution: if women do not disclose their sex, they will be assumed to be male, and will benefit from the greater deference accorded to men online. But just as the nineteenth-century novelist who hid behind a male nom de plume did not directly challenge the assumption that women could not produce great literature, non-gendered interaction does nothing to improve the poor impression of the validity of knowledge that women share online. We propose a tiered solution: for the time being, women ought not feel pressure to reveal their sex online, ought to be provided with mechanisms for receiving credit, should have the training and status to combat sexist trolling (via reverting content, enforcing rules, or even deleting editor accounts) and, in order to foster the peer-review-as-conversation model, edition builders ought to use multiple online social spaces in order to build the trust and collegiality required to produce investment in a knowledge creation environment.

The poem “All women have vertues noble & excellent” comes from a collaborative, evolving production space that allowed for various interpretations and amendments to authoritative text, and we must reflect this process in our contemporary modes of knowledge conveyance and edition building. Issues of authority, credit, or technological threshold are not the only reasons scholars resist contributing to Wikimedia projects. While it was easy to ensure civility (if not outright concern for gender equity) in Iter’s social space, it was very challenging to confront the systemic and direct sexism in Wikibooks. The wiki format itself does not discourage gender equity per se — we interpret our experience of sexism in Wikibooks as a reflection of women’s continued status in the culture at large. Regardless, simply avoiding engagement with Wikimedia is not a suitable or effective response to this issue. The Wikimedia suite of projects remains a key information resource for the general public; therefore, it behooves us as scholars and feminists to be certain that as a knowledge community Wikimedia is free of trolling, engages in the non-patriarchal organization of knowledge, and offers appropriate credit in order to increase gender equity, rather than marginalize women and the topics of interest to them.