Ethical Soundings in Collaborative Digital Humanities Research Projects: Critical Scenarios from The SpokenWeb

This series of papers draws upon recent and ongoing experiences from the interdisciplinary SpokenWeb <www.spokenweb.ca> research programme to consider the range of ethical scenarios that inform this collaborative network’s engagement with audio archives of literary and humanities-oriented sound recordings, and the communities of practice that generated them. We will consider ethical scenarios through specific projects of the SpokenWeb partnership, a research network that aims to develop coordinated and collaborative approaches to literary historical study, digital development, and critical and pedagogical engagement with diverse collections of literary sound recordings from across Canada and beyond. The goals and projects of the SpokenWeb partnership that will serve to focus our discussion of ethical scenarios in large-scale collaborative digital humanities networks include:1. rights and access management of digital research data and metadata;2. the ethics of archival listening as they pertain to the development of new forms of historicaland critical scholarly methods of engagement with the contents of documentary audio Archives; 3. automated techniques and tools for searching, visualizing, analyzing and enhancing criticalengagement (for features relevant to humanities research and pedagogy);4. pedagogy, training, mentorship and student labor (the organization of roles and relationsacross the research network);5. innovative ways of mobilizing digitized spoken and literary recordings within performative andpublic contexts;6. project management and governance.[Figure 1. Mainpage of SpokenWeb Research Site]To date, the proposed SpokenWeb archive contains approximately 5000 hours, or 12Tb, of audio data. The SpokenWeb program begins with the preservation and description of sonic artifacts that have captured literary events of the past, and quickly moves into a wide range of approaches and activities activate these artifacts in the present. For the purpose of this set of presentations that will use SpokenWeb projects as case studies for the consideration of ethical scenarios in digital humanities research programs, the authors intend a forward-looking focus that emerges from the fact that research and teaching with literary audio on this scale promises the possibility of paradigmPanelProposal for DH2020 shifting engagement with audio objects that have been, by and large, beyond the consideration of methodologies central to the discipline of literary studies and beyond the primary concerns of research librarians and archivists. An exciting dimension of SpokenWeb is the collaborative enterprise of digital humanists, cultural scholars, literary scholars, archivists, librarians, and sound technicians coming together to generate questions and solve problems beyond the capacity of any one of those fields of expertise. Such an undertaking invites the revisioning of processes of technical development, preservation and access strategies, metadata and systems coordination, rights management, pedagogical methods, student research and training, events curation and community outreach, and overarching concerns of project management and governance. With such revisioning comes the opportunity and responsibility to consider, or reconsider, the ethical dimensions of each set of project activities.Our proposal to explore ethical case study scenarios in relation to SpokenWeb research activities, approached as a multi-authored work with representatives from different constituencies of our research partnership, stems from the governance structure of the network. The governance structure is designed to provide representation of the key research areas of the SpokenWeb network, namely, literary methods, data and metadata, computational analysis, pedagogy and training, community outreach and events, and reserves additional positions for student representation. We will make use of this structure to organize the papers comprising our presentation, and will conclude with some framing observations about the important function of project management and governance in relation to the ethical scenarios discussed.Paper 1“Ethical Considerations of Rights, Access and Metadata Management of Digital Audio Assets”Annie Murray (University of Calgary) and Jason Camlot (Concordia University)Determining the rights holder for an archival audio recording can present challenges to archivists and researchers alike. In many cases, collections of audio recordings may have scant labelling on them, and in many archival collections, descriptive information may be at a basic level. Normally, physical examination of the cassette can yield some information about the possible rights holder; however, listening to the digitized recordings can bring about a fuller awareness of the specific copyright information needed to pursue permissions. The individual or organization who made a given recording is in most cases the rights holder. That is to say that the speaker or reader on the recording is in many cases not the rights holder. To what extent might the speaker or performer be informed of or involved in the permissions process for a recording? How would an institutional archive approach rights management compared to an artist-centred community? Rights management must transcend legal documents and procedures, and take into account relationships amongst archival, library, researcher, and literary stakeholders. Since our project intends eventually to provide open public access to this entire body of recordings, rights management activities need to take into account affective and ethical considerations that arise in order to ethically share archival recordings. Rights and permissions management of time-based media assets that document historical performances and speech acts of individuals before and within communities of the past demands engagement with the implications of replaying such events within communities of the present.In addition to such considerations surrounding the determination of permissions and access gradations for documentary audio recordings, we have been interested in reflecting upon the ethical implications of the nature and use of the descriptive metadata that we have produced and will continue to produce for the audio assets of the collections we are developing. During the first year of the SpokenWeb research program, a Metadata Task Force worked to develop a unique (yet linked) metadata schema for the description of documentary literary and humanities-oriented sound recordings held in both institutional and community-based collections. The schema is used by all team members to describe collections across the country. Interesting ethical questions have arisen both in relation to the establishment of specific metadata categories, and for the larger question concerning the degree to which the metadata we are producing should be made publicly accessible. Fields such as Genre, Content, and Statement of Responsibility have raised interesting questions about the impact of disciplinary presuppositions (in the case of genre), transcription (in the case of content) and definitions of “authorship” and communities of production (in the case of responsibility). We will identify and propose parameters of discussion for these three examples of metadata production and the ethical issues they raise. Finally, we will describe some pointed ethical scenarios pertaining to the use of metadata once it has been collected and made ready to share. While our first position has been (and remains, in principle) to make all SpokenWeb metadata open, we feel it is worth exploring possibilities and scenarios in which the descriptive metadata we are producing about documentary sound recordings may itself raise questions of ethical disclosure. While the argument that the open sharing of the descriptive metadata we produce should, legally, be allowable because such information consists of descriptive and reported facts about artifacts that are linked to, yet separate in substance from, the artifacts that may be subject to more complicated rights requirements, ethical questions remain concerning whether the parties that are captured in documentary audio collections wish their past activities, associations and locations to be known and be made publicly accessible through openly shared data that is designed to be linked and discoverable. For example, description of events that make intensive use of VIAF and Wikidata to establish association links between scenes and individuals, without details concerning the nature of past associations, may raise issues about what Erving Goffman once referred to as “Stigma management” in the present. Paper 2“Closer Listening: or, beyond the ‘anonymity of a murmur’”Deanna Fong (Concordia University), Michael O’Driscoll (University of Alberta)“We can easily imagine a culture where discourse would circulate without any need for an author. Discourses, whatever their status, form or value, and regardless of our manner of handling them, would unfold in the anonymity of a murmur.”Foucault, “What is an Author?”When Michel Foucault conceived of an archaeology of knowledge, predicated on the emergence of “statements” that, as Giorgio Agamben famously challenges in the archival ethics of his Remnants of Auschwitz, are events void of subjectivity, or when Foucault simultaneously contended for the death of the author as a mere subject-function unnecessary to the circulation of literature, he could hardly have imagined a vast digital archive of poetic discourse comprised entirely of the recorded voices of successive generations of creative writers. Amongst its many multi-disciplinary implications, the SpokenWeb Project promises, in moving from page to stage, a paradigm-shifting invitation to consider the digital audiotext as a fundamental medium of literary-historical analysis. In turning our attention from the printed, written text to the voiced performance, the SpokenWeb archive not only prompts new methodological questions for literary analysis, it also insists upon a revived relationship to the human being occupying the reanimated subject-position of “author.” Indeed, the audio objects at the centre of this collective enterprise—with their haunting disjointedness of the voice’s spectral presence—pose pressing questions of relation—of ethical relation, obligation, inheritance, and responsibility—between speaker and listener.This paper moves toward, or contends for, an ethics that is predicated on listening as intentional and affirmative of the speaker's embodied subjectivity—how we might listen to and for the grain of voice, the breath, the murmur, the semiotic rather than symbolic articulations of body—the openness of which runs counter to the hermeneutics of suspicion that governs conventional practices of textual analysis. As Mladen Dolar contends in A Voice and Nothing More: “The voice can be located at the juncture of the subject and the Other, just as it was before, in a different register, placed at the intersection of body and language, circumscribing a lack in both.” The speaker’s embodied subjectivity, foregrounded in the audio artifact, also points us to a set of material questions that have to do with the medium’s “counter-affordances” as a register of public speech. What can and can’t be said on the record, and by whom? Here, we argue that while audio recordings open us up to the stochastic noise of live speech and performance that these are also disciplined by the act of recording, and that this discipline unevenly applied to different subjects based on their relation to dominant power. In this manner, an ethics of archiving digital audio must also consider the material gaps and silences that exist within individual artifacts (in the form of cuts and edits, self-censorship, and muted or inaudible voices) as much as within larger collections (representation, institutional prioritization, preservation practices, etc.) “Automated Annotation, Metadata Production, and Scenarios of Difficult Content”Tanya Clement (University of Texas at Austin), Liz Fischer (University of Texas at Austin)Audio Visual collections, even when digitized, are difficult to search and analyze. Often, AV comes with limited metadata. Consequently, information professionals have to listen to items in real time and add information for indexing. Since enhancing searching, visualizing, and analyzing audio collections is often reliant on some level of metadata, an item that has not been vetted will often remain “hidden”. Libraries and archives, after all, are not want to make cultural heritage items that might potentially have privacy or security issues openly available online. In such cases, libraries have begun to consider using machine learning and other means of automatically generating metadata. Automatically generating metadata introduces ethical concerns, however. Jen Guiliano and CarolynHeitman use the term “difficult heritage” to explore “the aggregation of cultural records related to colonialism and its impacts on Native peoples” in their desire to “reconcile concerns of documenting the past with the need to confront the genocidal practices from which most of these records result”. In the context of SpokenWeb, such concerns are also relevant, though differently oriented. Many of the recordings include well-known poets such as Allen Ginsberg, Robert Creeley, and Muriel Rukeyser, but these poets also appear in recordings of panels or in the context of classroom encounters where unknown or unnamed voices also linger. Further, speakers may use language that is flippant or unconcerned at best or offensive at worst about topics such as race or gender in ways that are not reflective of more current perspectives and conversations. These incongruities are not new; they have always been part of our difficult heritage, but this discussion will articulate which ethical concerns have come to the fore when we use machines to ingest, sort, analyze, and represent sensitive materials in audio collections. A substantial proportion (~67%) of SpokenWeb SSHRC Partnership funding is dedicated to graduate and undergraduate students, whose training and research contributions to the SpokenWeb partnership form a core part of our mandate. The SpokenWeb Pedagogy Task Force has viewed this as an opportunity to develop mentorship and supervision best practices, building upon those outlined in “The Student Collaborators Bill of Rights” and with attention to findings and recommendations of “Student Labour and Training in the Digital Humanities” (Anderson et al). The integration of students into a large-scale digital humanities research network raises some key questions regarding mentorship, training, and graduate vs undergraduate student labour. For instance: what counts as “work” on SpokenWeb? How do we describe work the students are doing? Is it possible to ensure practices around student labour are addressed with consistency across the SpokenWeb network of Canadian and US American institutions? What kinds of iterative conversations are necessary to ensure work is meaningful to students? Murphy and Shearer consider the organization of roles and relations across the research network, with particular attention to the role of undergraduate trainees. What kinds of ethical issues emerge when involving undergraduate trainees in research roles?Anderson, Katrina and Lindsey Bannister, Janey Dodd, Deanna Fong, Michelle Levy, Lindsey Seatter. “Student Labour and Training in the Digital Humanities.” Digital Humanities Quarterly 10.1http://www.digitalhumanities.org/dhq/vol/10/1/000233/000233.htmlStone, Sandy. "Refuse closure; Insist on situation; Seek multiplicity.” http://www.deeplab.net/aboutusThe Student Collaborators Bill of Rights (UCLA). https://humtech.ucla.edu/news/a studentcollaboratorsbill-of-rightsPaper 5“The Ethics of Literary Curation, Archival and Actual”Jason Camlot (Concordia University), Klara du Plessis (Concordia University)Literary event curators adopt a range of curatorial approaches, varying from extremes such as Sarah Longair’s notion of “curatorial authority” (“Cultures of Curating”)—an intimate knowledge of work presented and a predefined context for its public representation—to Hans Ulrich Obrist’s organic model of providing a space in which experience is generated according to the individuals presenting or interacting with works (Ways of Curating). Positioned as mediator between authors and audience, however, the literary curator accepts, by default, a radical responsibility towards both the authors they represent and towards the audience they expose to literary performance. Both as historical or archival study and as continued praxis for contemporary literary events, poetry reading curation hinges on an ethical discussion of relational dynamics as manifested in literary dissemination. This paper takes two pairs of scenarios--each comprised of one archival and one recent experiential example--to frame the ethical concerns surrounding the curation of literary events.As a first case study, du Plessis follows the digital audio archives of Véhicule Art Inc--a bilingual, interdisciplinary artists’ run centre and exhibition space that hosted a weekly poetry reading series and quarterly open mic marathon events in Montreal, 1972-1983--as well as experimental, contemporary Deep Curation readings--an approach which places poets’ work in deliberate dialogue with each other and heightens the curator’s agency toward the poetic product. As a curatorial strategy, the former prioritizes a self-conscious preference for literary novelty and unrestrained openness, offering all agency to the authors as to what they present to the audience. Arguably, this (along with a wide range of cultural and sociological factors inherent to the 1970s in North America) frequently results in an unfiltered performance of sexist, misogynist, rhetoric; this work often signals sexually violent content, especially, when mapped onto a contemporary listening audience. As a corrective, a Deep Curational approach scripts exactly what the authors will present, creating a thematic and theoretical framing for the reading, while also retaining curatorial agency for the curator. Although authors consent to this process, dynamics of agency between curator and authors undergo a radical shift as the curator now enacts an authorial role.As a second case study, Camlot presents brief examples from SpokenWeb’s first “Performing the Archive” series, developed by staging public readings in which poets who read in Montreal as part of the Sir George Williams Poetry Series (1965-1974) return to read alongside their past archival selves. The illustrative examples in this case will include events involving Daphne Marlatt, Diane Wakoski, George Bowering and David McFadden. Mixing live reading with readings from the archive results in an affective performance of the relationship between the modes and meanings of performance in a past cultural and sociological context and the present one. Further as the live reader stands silently on stage while the archival clips are played, it also renders the act of listening a mode of visible performance for the audience. Such performances function as a lo-fi versions of what Steven Benford and Gabriella Giannachi have described as acts of “performing mixed reality.” Staging listening to the archive as a form of live performance can also be understood to open a temporal portal (to use a term from Wolfgang Ernst) between two ethico-cultural moments, and incites reflection upon the meaning of literary performance in different historical contexts. Each “Performing the Archive” event is followed by a period of discussion that explores the meaning of the temporal encounter that such a curated performance involving archival audio, entails, and, inevitably, raises interesting questions about the ethics of activating archives in distinctive, performative ways, and the different aesthetic and ethical meaning of literary performance in changing historical contexts.Beneford, Steve, and Gabriella Giannachi. Performing Mixed Reality. Cambridge, MA: The MIT Press, 2011.Ernst, Wolfgang. Sonic Time Machines: Explicit Sound, Sirenic Voices, and Implicit Sonicity. Amsterdam: Amsterdam University Press, 2016.Longair, Sarah. “Cultures of Curating: The Limits of Authority.” Museum History Journal (2015) 8:1, 1-7.Obrist, Hans Ulrich and Asad Raza. Ways of Curating. New York: Faber and Faber, 2014.Paper 6“Project Management, Governance as Causes and Mitigators”Yuliya Kondratenko (Concordia University) and Jason Camlot (Concordia University)This paper presents scenarios concerning ethical considerations surrounding the management of large scale, interdisciplinary DH projects, and how governance structures and processes may help mitigate conflicts across a research network.The Project Management Institute’s (PMI) Code of Ethics and Professional Conduct states that when project management practitioners are attempting to uphold principles of fairness, they should aspire to “…constantly reexamine [their] impartiality and objectivity, taking corrective action as appropriate.” One of four foundational values of PMI Code (the other three are responsibility, respect, and honesty), this principle translates into the need for a consistent practice that aims to “...recognize when we (project management practitioners) have conflicted loyalties, and to identify when we are inadvertently placing ourselves or others in situations of conflict-of-interest.”How might this principle of fairness be enacted and applied in the context of a large and diverse research network with a wide range of stakeholders, a governance structure that is composed of humanities researchers, librarians, and community partners, all with competing interests. Take librarians, for example. On the one hand, they have a duty of loyalty to their own institution when performing their functional role. On the other hand, when serving in their role as a member of a research project (as a member of the Governing Board or a project-oriented Task Force), they may be asked to reconsider, reexamine or reshape policies and procedures that they have been hired to uphold in their institutional role. The project manager in such a scenario ends up enabling and perhaps even inciting a conflict-of-interest situation, which represents a conflict-of-interest in itself, for the project manager, as it undermines a key ethical principle of the PMI’s Code of Ethics and Professional Conduct.[Figure 2. SpokenWeb Network and Governance Diagram]The establishment of a well designed governance structure can help to mitigate such scenarios of conflict of interest (and other forms of conflict). By identifying core features of the constituencies involved in the research program, ensuring representation of those constituencies on the Governing Board, and Task Forces, and especially by collaboratively establishing guideline and process documents that may be referred to for decision making in areas that represent regular points of tension between the goals of a network project, and the goals of a participating institutional or community partner, scenarios of conflict can be managed in such a way that they do not undermine the advancement of project aims, and to the satisfaction of the interests of the partner. Importantly, creating scenarios for collective reflection upon the complexity of our loyalties and relationships is a productive approach. Staging discussions about agreed upon methods of collaboration -- for example, the use of cross institutional Task Forces designed to advance specific goals -- is a useful and neutral way to reflect upon the effects of institutional loyalties in relation to shared goals of the research program. This paper presents scenarios of such conflict and resolution for consideration and discussion.