Chapter 2 Replatforming

Re:platforming

“Platform thinking” is characteristic of DH, argues Steven E. Jones, stressing the human collaborations behind “layered systems, built and rebuilt by maker-scholars as starting places, platforms on which to build future things” (16, 174). I focus here on such scholar-led infrastructure that emerges from researchers’ practices (Anderson, 10), including three interlinked DH projects, one that stumbled into creating infrastructure and the others funded as such, that I have codeveloped (Orlando Project, “Credits”; Canadian Writing Research Collaboratory, “Credits and Acknowledgements”; Linked Infrastructure for Networked Cultural Scholarship, “Our Team”). The first, the Orlando Project, an experiment since 1995, recently redesigned its original publication platform, but before that had retired its homegrown production backend in favor of the second, the Canadian Writing Research Collaboratory (CWRC, pronounced “quirk”), which launched in 2016 and relaunched in 2025 after a full rebuild. CWRC serves as a virtual research environment for many individual projects built on it. The third, the Linked Infrastructure for Networked Cultural Scholarship (LINCS), provides tools for converting and publishing datasets as linked open data. They are illustrations, accompanied by other projects that I will mention, of the complexities of what Alan Liu has called the diversity stack: “a fused techno-ideological apparatus—a platform in all senses” (133).

The born-digital scholarship of the Orlando Project aimed to platform marginalized voices through detailed treatment of British women’s writing. Orlando now profiles 1,444 writers using 3 million semantic tags embedded in 9 million words. Of course, such counting is not in itself meaningful, and this summary misrepresents the ethos of a project devoted to qualitative data. In evaluating feminist digital archives, Jacqueline Wernimont cautions against reliance on “patriarchal tropes of size, mastery, and comprehensive collection” that reinforce funders’ notions of impact (par. 4, par. 5). Yet these numbers are important as measures of inclusion and a recognition that formation of the literary canon in digital space stems as much from inequitable access to infrastructure as from abstract measures of value. On the early do-it-yourself (DIY) web, Amy Earhart notes, digital literary scholarship “reinserted women, queers, and people of color into the canon” (Traces, 66). However, “the democratization of knowledge made possible by the developing technological infrastructure” nosedived as web publishing became less accessible and much early recovery work was lost, along with the ability to sustain platforms (“Can information?”). Miriam Posner therefore assigns an ethical mandate to those who operate DH platforms: “it is incumbent on all of us (but particularly those of us who have platforms) to push for the inclusion of underrepresented communities in digital humanities work, because it will make all of our work stronger and sounder” (39). Scholarly platforms can fight injustice by amplifying marginalized voices and nonhegemonic perspectives.3

The Orlando Project dedicated its semantic markup to platforming difference while seeking to diversify online content. From what is now called an intersectional feminist approach, Orlando devised a means of encoding differences, refusing, for instance, to let whiteness operate as a silent and unmarked norm against which a smaller number of Black, Indigenous, and people of color (BIPOC) writers were defined, and encoding ethnicity, religion, class, and gender in conjunction with race. It was activist in avowing the constructedness of its markup and the ideological investments of its knowledge representation. Embedded semantics helped to counter some of the problematic implications of platforming individualist authorial profiles—for instance, by enabling materials to be retrieved by the semantic tag for “destruction of work” instead of only by text search, revealing the recurring suppression of women’s writing by male relatives and authorities. The tagging works across individual profiles to illuminate systemic patterns and processes of change. Orlando’s ability to platform difference depended in the early 2000s upon building a multilayered platform comprised of a markup schema or data model, a backend collaborative editing system, and an experimental front-end that leveraged tagging.4 This platform allowed Orlando to register as a significant intervention that shifted how one thinks (Nixon), modeled intersectional knowledge representation (Risam), and changed the study of women’s writing (Bowers).

The CWRC emerged from rebuilding Orlando’s aging infrastructure, aiming to increase the diversity of scholarly web content and empower those with modest technical expertise to adopt DH practices. CWRC hosts 400,000+ objects (e.g., texts, images, videos, and bibliographic records). The thirty-five projects that it houses include Karen Skinazi’s collection of writings by Winnifred Eaton (pen name Onoto Watanna), the first Asian North American novelist; Canadian Jewish Women Writers, directed by Catherine Caulfield; and The People and the Text, a history of Indigenous texts directed by Métis/Cree scholar Deanna Reder. For CWRC as a settler-led project to collaborate with The People and the Text in platforming Indigenous scholarly and literary voices has been a meaningful way to work toward truth and reconciliation.

Indigenous infrastructures lead the way in decolonizing information systems and centering sustainability. For some content, both data and platform sovereignty are essential to Indigenous sovereignty. The OCAP® principles of “ownership, control, access, and possession” assert First Nations control over data collection and use, and with the CARE principles for Indigenous data governance (Collective benefit, Authority to Control, Responsibility, Ethics), complement the FAIR (Findable, Accessible, Interoperable, Reusable) principles for open data (GO FAIR) in recognizing the tension between openness and data sovereignty (Carroll). Assumptions about universal openness as an unquestioned good are challenged when ownership is understood as collective and access as more complicated than all or nothing. Traditional Knowledge (TK) labels, as developed by Jane Anderson, Kim Christen, and colleagues (Anderson and Christen, “‘Chuck a Copyright on It’”) meet the infrastructural need to inscribe protocols developed in consultation with Indigenous communities. Platforms such as Mukurtu operationalize the metadata represented by TK labels to guide who should have access to what content under what circumstances (Christen, “Does Information?”). Such systems reinforce other scholars’ insights that people, communities, and relationality are the heart of infrastructure (Endings; Edwards et al.; Star).

The state’s “power to name” was key to how railways, as communications infrastructure, advanced colonialism (Duarte and Belarde-Lewis, 681, quoting Olson), so Indigenous approaches to situating knowledge through metadata can model responsible platforming. For Wemigwans, a “digital bundle” of data and its cultural protocols provide an infrastructure for “decolonizing the digital” by countering decontextualization (A Digital, 42, 43; “Digital Bundles”).5 Imaginings grounded in Indigenous knowledges and decolonizing methodologies that accommodate “a plurality of knowledge systems” (Duarte and Belarde-Lewis, 678) can help overwrite colonial classifications, but they themselves require platforms. The Respectful Terminology Platform Project, led by Stacy Allison-Cassin (Métis Nation of Ontario) and Camille Callison (Tahltan Nation), will enable the development of a dynamic, multilingual set of terminologies to replace inappropriate metadata and model strategies for countering harmful data legacies.

Parallel concerns emerge when considering how platforms can promote diversity and epistemic justice in collaboration with affected communities (Fricker; Poirier and Costelloe-Kuehn). T. L. Cowan and Jasmine Rault’s Cabaret Commons, a community-based and -sourced project in CWRC initially devoted to ephemera from the queer cabaret scene, highlights the complexity of platforming data involving living persons. Their speculative exploration, continuing through the Digital Research Ethics Collaboratory, produced a closed archive in which data initially uploaded in a pilot to CWRC was kept dark due to privacy and permissions concerns. The collaborative Feminist Data Manifest-No, with Cowan and Rault as coauthors, similarly creates a no-data anti-archive that treats data as fundamentally personal, tactical, and temporary: “We refuse the use of data about people in perpetuity. We commit to embracing agency and working with intentionality, preparing bodies or corpuses of data to be laid to rest when they are not being used in service to the people about whom they were created” (Cifor et al.). The Manifest-No collaborative builds on a long tradition of feminist understanding of the imbrication of processes of knowing with power and social relations (Scott). Focused on the data extraction, dataveillance, and vulnerability of minoritized groups associated with corporate platforms, it stresses the harm inflicted by data on persons treated as research objects rather than coresearchers, and insists upon the right to be forgotten (Cifor et al.; Scott).

Platforming thus evinces online infrastructure’s dynamic presentism. Facing current scholarly and social debates, resurgent and emergent epistemologies, and evolving understandings of responsibility and accountability requires that scholarly platforms be grounded in the relationality of care and repair for people and communities, as well as for machines and software. Like large corporate platforms (Helmond, “Facebook’s Evolution,” 125), but for other reasons, scholarly platforms engage in ongoing updates and iteration as resources permit, resituating content (re-platforming) and renewing systems (replatforming)

Re-platforming

While deplatforming typically limits platform access to mitigate harm (Thomas and Wahedi), the concept of re-platforming raises questions about the potential harm involved in republishing online already public information or migrating online content to a different platform.6 Re-platforming emphasizes the importance of asking whether and how to republish existing data, its hyphen symbolizing the pause that should accompany such decisions, even if the data was previously published online. Deplatforming, as noted previously, came to the fore in connection with social media platforms, and, like platform, the word is etymologically rooted in live speech (Perlman).7 Objections to deplatforming in a social media context typically critique limitations on speech, whereas arguments for it stress the public good and denounce the erosion of the information commons by private interests (Gillespie, “Politics”; Jhaver et al.; Rogers). Debate informed by postcolonial, queer, and feminist critiques addressing the ethics of what I am terming re-platforming has been extensive within the archival community (Agostinho; Dalgleish; Povinelli). It deserves more discussion within DH, particularly since sustainability, preservation, and access considerations are making re-platforming increasingly common activities in the field.

Re-platforming can seem straightforward, especially with respect to historical data. For instance, CWRC re-platformed Thomas B. Vincent’s CD-ROM Index to Pre-1900 Canadian Cultural and Literary Magazines (originally published in 1994) as the Early Canadian Cultural Journals Index.8 This index constitutes a valuable resource for researching early Canadian culture because it indexes many journals found online in the Canadiana collection, which is among the largest collections of digitized Canadian documentary heritage in the world. Canadiana provides the full text of many of these magazines but lacks detailed indexing, whereas Vincent indexed editorial content, provided keywords, and designed the dataset to support future linking (“Opportunities”). Online and downloadable, with a Creative Commons license, Early Canadian Cultural Journals Index is more accessible and supports analysis in ways that go beyond the FAIR principles because it is not only archived for preservation with other CWRC data but also accessible alongside related content through a dynamic interface.

There are problematic angles to data rescue, however. Marika Cifor draws on Sara Ahmed, Lauren Berlant, and Ann Cvetkovich to argue that archives create affective relations with social justice implications, “creating, documenting, maintaining, reconciling and (re)producing such relations—between records and people, ideologies, institutions, systems and worlds.” As noted, Vincent’s original Index recorded much of the content that would later be put online in Canadiana, whose home page warns that “The Canadiana Collections contain content created over five centuries that tell an incomplete, oftentimes distorted and, sometimes harmful, story of Canada. The content, metadata, and resource descriptions in the Canadiana Collections contain language that reflects the biases, norms, and perspectives of the time in which they were created. This includes harmful and offensive wording, cultural references, and stereotypes.”

This warning is relevant to Vincent’s Index. Counting item titles and keywords, the word “Indian” or “Indians” appears in 436 of the Index’s records as the predominant term for Indigenous peoples in the area now known as Canada. Moreover, titles such as “The Indian Problem and the Washakada Indian Home, Elkhorn, Man” refer to residential schools that caused individual and collective traumas. The CWRC home page for the re-platformed Index displays the same warning about problematic historical content. In re-platforming the data again for the updated platform, CWRC is also adopting measures followed by memory institutions, including Canadiana’s host (the Canadian Research Knowledge Network), in revising the most egregious terms in the Index, and it hopes to engage Indigenous experts, scholars of early Canadian writing, and library partners on how best to handle other potentially problematic terms that require deeper knowledge to address respectfully.

Re-platforming existing data, then, even when that data does not represent living persons, can re-present epistemic violence and perpetuate trauma associated with racism, colonialism, and other forms of oppression. Dialogue with evolving standards and values and affected communities is therefore crucial. Jessica Marie Johnson lays out the affective tensions between the historical record and present needs to critique the complicity of digital design and implementation in commodifying Black bodies. Yet despite the horror invoked in her essay, she notes how the Trans-Atlantic Slave Trade Database, which is a central exemplum of her argument, was able to reshape the study of slavery by filling in, albeit with the smallest traces of material existence, the gap left by women and children excluded from the historical record. Johnson calls for a “Black digital practice [that] requires researchers to witness and remark on the marked and unmarked bodies, the ones that defy computation, and finds ways to hold the null values up to the light” (70–71).

Not only does the colonial archive suffer from harmful gaps and silences—untold numbers of unnamed human data points and “missing” datasets—but it also suffers from the reinforcement of the imperial gaze that can occur without careful contextualization of the cultural work done by “digital artifacts of a sensitive and dehumanizing nature” (Odumosu S298). Careful re-platforming can mitigate harm by negotiating the complexity of networked archives in the era of big data (Agostinho; Carter; Hartman; Onuoha). Alternatively, disrupting the algorithmic logic of “search, find, identify, claim, or steal,” which keeps “the colonial episteme intact,” Temi Odumosu argues, may require denying an artifact further reproduction, which is to say, refusing re-platforming (S298).

Re-platforming, then, risks perpetuating the extractive colonial logic that initially populated the collections of memory institutions through the further dissemination of representations of violence, violation, and atrocity, whether via images of enslaved persons, diagrams of slave ships, or representations of sacred Indigenous ceremonies. Decontextualized, often misattributed, or simply reproduced, such materials participate in what Susanne Kappeler describes as the pornography of representation, which reproduces and legitimizes the unequal power relations embedded in the artifact’s creation, against which social justice strives (Anderson and Christen, “Decolonizing”; Duff; Callison et al.; Guiliano and Heitman; Kappeler).9 As with the University of Nebraska’s decision not to publish online a digitized archive of the satirical Awgwan student paper from the early 1900s, which was rife with discriminatory and derogatory content, sometimes outright refusal to re-platform is the answer (Brink et al.) and can serve as a generative stance (Tuck and Yang).

When digital content references living persons, as indicated earlier, complexities multiply. The digitization of On Our Backs, a pornographic lesbian magazine pivotal to the twentieth-century feminist sex wars, is a powerful example of the need for care in platforming formerly published content. Reveal Digital, a company that crowdfunds through libraries to digitize collections and make them open access, re-platformed the journal in 2012 as part of its “Independent Voices” collection. Reveal Digital withdrew On Our Backs in 2016 for multiple reasons, including concern about outing individuals represented in the collections. With respect to the latter, librarian Tara Robertson initiated the ethical debate in a blog post on her diversity, equity, and inclusion consultancy website, highlighting the tension between access and consent: “Consenting to a porn shoot that would be in a queer [print] magazine is a different thing to consenting to have your porn shoot be available online.” Digital preservation and remediation often blur into re-platforming. Indeed, platform limitations (namely, the inability to restrict access by age to comply with pornography laws in some states) were a determining factor in the decision to pull On Our Backs, even though the periodical’s content had been reframed as “alternative press” by its placement within the Independent Voices collection (Reveal Digital). That reframing remediated not only the content, but also “scale, audience, and interaction” through the JSTOR platform (Groeneveld).

These kinds of affordances are included in Jay Bolter and Richard Grusin’s notion of remediation (Remediation). However, considering re-platforming raises the question of who is accountable for the distribution of remediated materials. Major social media platforms deny responsibility for hosted content (Gillespie, “Regulation,” 666; “Politics”), a position contradicted by the practice of deplatforming. The mandate advocated here for scholarly platforms is an ethical one, based on the insight that the politics of digitization are inseparable from platform and cataloging choices that determine access.10 Unpacking the tensions associated with such responsibility is crucial to understanding both the implications of how research platforms work and how we might build them otherwise.

Parallelling Indigenous articulations of data sovereignty are arguments that vulnerable groups need to control their own means of digital reproduction in ways that go beyond conventional understandings of copyright or consent (Boyles). As Elizabeth Groeneveld argues, corporate and academic archives whose “digitization frameworks have not been developed with the specific needs of minoritized communities in mind” operate with different values from “a grassroots community archive that emerges directly out of queer/feminist principles” (78). Exemplifying the magnitude of this challenge is Cowan and Rault’s search for “media and infrastructures of risk and care” that produce a Cabaret Commons that refuses extractive practices and exploitation by transforming “institutional and platform logics” (122).

For Johnson, practices of hacking, remixing, and institutional resistance contribute to “an effort to dismantle the residue of commodification that is slavery’s legacy” and an ongoing source of trauma. Such practices compel designers to “collaborate with the living descendants of the enslaved, who still claim as ancestor and kin those who can only be rendered in databases as ‘1’ or a single pièce d’Inde” (Johnson 66, 71). The Digital Oral Histories for Reconciliation (DOHR) project, a platform for first-person voices of survivors of the Nova Scotia Home for Colored Children, hints at the possibilities. Collaborators Jennifer Roberts-Smith, Kristina Llewellyn, and members of the Victims of Institutional Child Exploitation Society (VOICES) incorporated “relational presence,” which foregrounds difference rather than immersion, into the design of DOHR’s virtual reality experience based on values that include “affective dissonance” (Smith et al.; Roberts-Smith). This platforming emerged from a community-based restorative justice process that supported intense design collaboration and the embedding of digital artifacts within a curricular process supported by trained facilitators—a whole-project rather than solely an infrastructure initiative. Moreover, VR permitted a degree of control over access and context impossible for web platforms. Scaling up and opening up on the web entail significant challenges, many associated with the linking that is the web’s foundational affordance.

In fact, ordinary linking on the web needs considerable supplementation through linked data to overcome such challenges. Linking in the latter sense—that is, the disambiguation of data through reconciliation against shared and persistent vocabularies—is identified by Donald J. Waters as a crucial component of “emerging digital infrastructure for research in the humanities” (87). Re-platforming existing datasets in new forms, such as linked open data, makes data FAIR and yet demands critical consideration. Representing people and other entities unambiguously on the web and interlinking them has immense potential for advancing knowledge. For instance, linking named to unnamed subjects provides one approach to addressing the null values invoked by Johnson, opening up positions for the unnamed within historical networks of relationships. Indeed, the project Enslaved: Peoples of the Historical Slave Trade is re-platforming historical records related to enslaved persons as linked open data for inclusive and reparative ends.

Yet disambiguation can be deeply problematic in the context of queer identities. For instance, reusing unique identifiers to link people from On Our Backs to current identities or to link two discrete personas from different contexts could be ruinous or traumatic for those involved. As Cowan and Rault observe of past queer cabaret performers, “even the research required to find these folks and ask for permission can have the unintentionally harmful consequences of detonating a former name in a search engine, algorithmically linking that name, gender, sex, and/or other marker to a current identity that the person has worked hard to delink” (125).

LINCS recognizes that open is not an indisputable good, and that linking may harm. This suite of tools for remediating and re-platforming datasets aims to support the use of linked data in reciprocal and respectful ways as a bridge to closed or sensitive data. Enabling such use includes developing processes for flagging problematic data and honoring requests to anonymize or deprecate data, notwithstanding the challenges of excising portions of interlinked datasets. What problematic or sensitive data often requires, however, is context, which is greatly determined by platform configurations and affordances. This means that although domain experts bear responsibility for materials that they select or create, scholarly platforms have an obligation to address concerns that arise from re-platforming those materials11 and to support contextualization. LINCS has sought to ensure that its core data structures—the ontologies that govern relationships between entities, determining what it is possible to represent structurally—support contextualization through the adoption of two core ontology structures, the CIDOC-CRM ontology and the Web Annotation Data Model. These two ontologies situate linked data assertions in ways that reduce the risk of decontextualization and help address the ethical concerns associated with re-platforming (Canning et al.; Brown, Canning et al.). LINCS maintains living policy documents to communicate platform values and to work with researchers toward a shared understanding of our mutual responsibilities. For instance, Figure 2.1 shows how the platform’s ontology policies adapt ideas from Catherine D’Ignazio and Lauren Klein’s Data Feminism (218–19) to lay out strategies for countering systemic inequality.

As LINCS grows, it will monitor whether attempts to convey values and critical data awareness through policy actually shape practices. Similarly, the ability of linked data to structure data in ways that support diverse epistemologies can be evaluated only when there is a critical mass of data representing different ways of knowing.

Re-platforming involves the “capacity of data to move on from the original signature and produce new possibilities” (Agostinho et al., 438). Those possibilities are exciting but entail risk. At times, not platforming may be more appropriate than platforming or re-platforming. CWRC has declined to host projects that its Research Board deemed incongruent with the platform’s values, including a case where categories associated with the representation of Indigenous persons were not aligned with current vocabularies. Likewise, LINCS chose not to re-platform a social media dataset involving vulnerable subjects. Platforming becomes complicated when a dataset is no longer active, since domain experts are no longer engaged in caretaking related to sensitive content. It becomes a question of whether a platform has the expertise and resources to undertake remediative labor alongside all the other forms of care and repair required to sustain itself, the final sense of replatforming that I explore next.

Replatforming

There is rising awareness within infrastructure studies of the critical importance of maintenance rather than innovation (Russell and Vinsel), and within DH of the importance for sustainability of factors such as people, communities, and the environment (Neylon; Nowviskie “Capacity,” “Digital Humanities”; Drucker; Sample; Goddard and Seeman). Replatforming in this maintainer’s and caretaker’s sense involves the repetitive processes of operating, repairing, and upgrading scholarly platforms. For Steven J. Jackson, care is an ethical proposition that breaks down subject-object relations between humans and things. These blurred boundaries, coupled with the “gendered infrastructural imaginaries” (Agostinho and Thylstrup, 754) associated with container technologies (Sofia), lead to infrastructure and operations being devalued as support or service in contrast to higher values placed on innovation and production (Anderson; Brown, “Delivery Service”). Not only is the “repair or replacement of broken infrastructure . . . necessary for any form of sociality to extend itself,” Lauren Berlant argues, but the extension of infrastructure can also be sufficiently generative to open up utopian possibilities of a renewed commons (393).

The desire to experiment and renew, of course, pulls against the need to standardize and stabilize for sustainability. Orlando had ambitions well beyond what its initial interface offered, planning affordances for exploring intersectional identities, analyzing markup patterns, and enabling user contributions. Some aspirations were realized, but with limitations; most resources were instead absorbed in fixing, tweaking, and upgrading the initial platform in ways that, while making a difference, were not transformative. After publication in 2006, Orlando’s growth reflected the types of resources available; support in the form of research assistance enabled ongoing content production. Experimental interfaces created for the project stimulated insights (Holland and Elford; Brown, Ruecker et al.), but resources could not stretch to building production versions.12 Sustainability needs thus pull against infrastructural ambitions; yet the paucity of funding for operations, as opposed to innovation, pushes platforms toward overextension rather than improvement of core functionality. In a DH context, this story of possibilities curtailed by sustainability needs is unusual only in that Orlando’s longevity speaks to its phenomenal good fortune in being able to operate for as long as it has and rebuild both its front and back ends.13

Care and repair can be generative and rewarding. However, the term replatforming stresses the Sisyphean aspect of platform maintenance, even when conducted according to best practices within stabilizing institutional frameworks (Smithies et al.). The quotidian task of monitoring and maintaining a platform with scant resources so it is always accessible is shadowed by awareness that systems need major repairs, often suddenly and urgently, and eventual rebuilds. Even conceptualizing systems as distinct from people is wrongheaded (Ciula and Smithies), as becomes glaringly obvious when DH infrastructure fails because a sole sustainer disappears. People critical to the task of sustaining systems include professional technical staff, students, librarians, and/or faculty, whose labor may be divided, soft-funded, ill defined for this purpose (in relation to their job as a whole), or voluntary.

The precarity of scholar-led research platforms can exact high levels of administrative work in the perpetual production of grant applications to keep platforms going and, if the applications are successful, in project tracking and reporting. Such precarity consumes labor and creativity that could otherwise be directed toward generating infrastructural insights for platform renewal. Design thinking is easier at early stages, as is evident even in the case of the innovative Scalar platform for multimodal long-form arguments, which built front-end customizability into its architecture from the start (McPherson, Feminist). The more established an infrastructure, the greater the friction associated with design changes it increases once a tool shifts from being a purely lab-based solution or tactical object for thinking (Galey and Ruecker) to operational use. It is hardly surprising, given his expertise, that Bruno Latour built and then “gracefully degraded” (Nowviskie and Porter) rather than tried to sustain, the bespoke web platform for contributing content to his An Inquiry into Modes of Existence project. The platform, an act of critical making, was revised toward the end of its funding period, prioritizing normative platform criteria (speed, stability, usability, reusability, and durability) and mundane maintenance (“the host of problems such as errors, backtracking, modifications, and bugs that are all inherent to this kind of adventure”) while dispensing with more innovative, labor-consuming features for engaging “co-inquirers” and active contributors (Ricci 50; Latour; Brown, “Taking AIME”). The major outcomes of Latour’s project were disseminated in a monograph report, articles, and a website with reduced functionality.

Platforming with a Difference

I favor verb forms of platform because, as Susan Leigh Star and Karen Ruhleder emphasize, “infrastructure is something that emerges for people in practice, connected to activities and structures” (379). Platforming is processual and performative, played out each time functionality is invoked (Hayles, 274). Replatforming, as explored here, performs knowledge production, elevating voices through iterative acts that embody value. The political stakes of platforms can be elicited in “the ways in which social forms are written into the technological scaffolding of information, and how they reflect and materialize power dynamics, thereby structuring the possibilities for social action” (Agostinho and Thylstrup, 750). In recognizing that platforms are imbricated with power, academics and professionals can work to push scholarly platforms beyond established practices and reshape the intimate space of information exchange (Christen and Anderson, “Slow Archives”).

Researchers are ethically responsible for how they interact with archives (Moravec), and yet how they do so is massively shaped by the platforms with which they interact as consumers, remediators, and producers of knowledge online. Envisioning how an archive of colonial violence might signal “where and how sensitivity is required, not as an optional stance but as a prerequisite for the digital encounter” (Odumosu, S298) suggests how platforms might perform what Anjali Arondekar calls “a counterrecord of that history” of the colonial moment (12). For Odumosu, metadata is a quiet “undercommons reconfiguring the digital thoroughfares (associations, keywords, hyperlinks) that bring a public into encounters with challenging histories” and “capable of narrating in full an object’s life and afterlife, . . . making that known to users with each right-click and download” (S299). As her invocation of affordances offered through hyperlinks and clicks indicates, this vision conjoins metadata with platforming as an active, ethical intervention in embodied knowledge spaces. Johnson similarly envisions a subversive Black digital practice that “engages data promiscuously, across multiple platforms, taking up the nearest tools at hand to defy, dismiss, jeer, and sneer at the presumed legitimacy claimed by institutional structures and categories of analysis generated by the Ivory Tower” (71). The extent to which Indigenous, feminist, Black, queer, and trans values challenge normative institutional logics and epistemologies underscores the impediments to (re)platforming with a difference. As Cowan and Rault warn, it is a long game: “this will be a collective struggle that may take more than a generation to reframe” (136). Yet shifting slightly, iteratively, is still movement.

Jones characterizes DH platforms as “Frankenstein’s monsters that we stitch together ourselves and for which we take responsibility” (174). Thinking through platforming helps identify the messy overlapping areas of collective and individual responsibilities involved in the creation, situation, and dissemination of networked content. The ongoing, iterative processes of platforming materials, re-platforming them in new forms, and renewing platforms in pursuit of better ways of knowing is rife with precarity but also with empowerment and promise. Understanding the relationships and responsibilities bound up in these multifaceted processes is key to pursuing the potential that is glimpsed, but only partially and unevenly realized, in scholarly platforming to date: a path to better situated and more ethical knowledge production and circulation.

Replatforming invokes the inevitable human processes whereby analog or digital materials are resituated in the process of migrating from one platform to another, and whereby scholarly platforms require care, repair, and eventual retirement or renewal. The challenges of replatforming with a difference are vast and varied because even platforms that offer modest alternatives to mainstream platforms defy logics that are fundamental to the economic and ideological conditions of our historical moment. For that reason, replatforming must be understood as a crucial component of ongoing efforts with DH to embody cultural critique and advance justice in the knowledge that we network and the systems that we build.

Notes

1. 1. Notably, after Google’s 2006 acquisition of YouTube consolidated its power, YouTube swapped out the terms website, service, forum, and community in favor of describing itself as a “distribution platform for original content creators and advertisers large and small,” thus aiming to “strike a regulatory sweet spot between legislative protections that benefit them and obligations that do not, and to lay out a cultural imaginary within which their service makes sense” (Gillespie, “Politics,” 348). See Helmond et al. on how Facebook transformed from social media site to platform. See also Bratton (34).

2. 2. Thus “Earth” is the name of the fundamental layer of Benjamin Bratton’s The Stack, and of the opening chapter in Kate Crawford’s The Atlas of AI, which stresses that the cloud is “made of rocks and lithium brine and crude oil” and is voraciously consuming rare “elements that required billions of years to form inside the earth” (31).

3. 3. Earhart stresses the need for infrastructure to support small as well as large, high-profile projects; Moya Bailey highlights the challenges for community members of engaging with such infrastructure (cited in Boyles, 123).

4. 4. The Orlando Project, a generously funded project directed by Patricia Clements (1995–2007), Susan Brown (2008–2015), and Brown, Katherine Binhammer, and Isobel Grundy (2016–present), originally presumed that available solutions would serve for project publication; therefore, nothing was budgeted for programming. The project was fortunate to be able to build both back-end content management and front-end web publication systems with in-kind resources, grant funding from the Canada Foundation for Innovation (secured by project member Susan Hockey), and a loan from the University of Alberta.

5. 5. FourDirectionsTeachings.com, developed by Wemigwans in reciprocity and relationship with Indigenous Elders, “speaks back to dominant colonial systems of knowledge” through multimediality that expresses Indigenous Knowledge “in ways that contribute to the reflection of Indigenous ontology and values on the World Wide Web” (Four Directions Teachings.com; Wemigwans, A Digital Bundle, 2, 42).

6. 6. Other senses of replatforming include shifting an application from one platform to another, typically cloud-based, platform, and Julia Ebner’s use of the term to discuss the parallel online worlds of right-extremist alternative social media.

7. 7. Deplatforming is most strongly associated with attempts to reduce hate speech and harassment associated with groups and individuals on the political right (e.g., Khoo), but affects others as well. For instance, mainstream journalists critical of Elon Musk were removed from Twitter after he acquired that platform and renamed it X; and, in another instance, feminist academics in the United Kingdom were allegedly silenced for their stance on gender self-identification legislation, inspiring the Cambridge Radical Feminist Network’s Spring 2021 “Replatforming Deplatformed Women” series. Lovink doubts that deplatforming will “fundamentally change the ways entertainment and distraction are organized” (Stuck, 144).

8. 8. This massive bibliographical dataset, published in 1993 as a CD-ROM (Vincent, Index; Early), contains 137,943 records of prose, poetry, drama, fiction, and other items from 203 nineteenth-century periodicals published in Canada, many now digitized and available online through the Canadiana collections. It was re-platformed on CWRC with Vincent’s permission.

9. 9. Kappeler’s choice of a photograph of white men displaying the body of a Black man whom they murdered as a paradigm of the representational structures she seeks to reframe anticipates Tonia Sutherland’s analysis of the traumatic impacts for the black community in the United States of the repeated online trafficking of images of Black bodies in the absence of reparative and transitional justice: “While communities of color have long engaged in ritual practices of (re)membering and bearing witness to violent acts as modes of resistance and mourning; in digital spaces these practices have been appropriated to reinforce systems of white supremacist power and racial inequality, re-inscribing structural and systemic racism” (33).

10. 10. Zaagsma’s summary of the politics of digitization nicely articulates these imbrications.

11. 11. Tara McPherson’s handling of complaints about privacy violations in a Scalar project that had harvested social media content illustrates the tension between ensuring that re-platformed content aligns with a platform’s ethical commitments and assigning researchers responsibility for the content they produce (“From Vectors,” 48, 56).

12. 12. Orlando’s platform remained substantially unchanged from 2006 to 2022, a span that attests to the design foresight, talents, and dedication of technical-team lead Jeffery Antoniuk in keeping aging production and publication platforms functioning while juggling other responsibilities.

13. 13. With its rebuilt interface launched in 2022, the team again hopes to extend functionality to do better justice to the data but struggles to meet fundamental production requirements. For a more detailed account in the Canadian funding context of infrastructure challenges (including for Orlando and for CWRC), see Susan Brown, Kim Martin, and Asen Ivanov.

Bibliography