Uncertain Archives: Critical Keywords for Big Data

Uncertain Archives

Uncertain Archives

Critical Keywords for Big Data

Edited by Nanna Bonde Thylstrup, Daniela Agostinho, Annie Ring, Catherine D’Ignazio, and Kristin Veel

The MIT Press

Cambridge, Massachusetts

London, England

© 2021 Massachusetts Institute of Technology

All rights reserved. No part of this book may be reproduced in any form by any electronic or mechanical means (including photocopying, recording, or information storage and retrieval) without permission in writing from the publisher.

This book was set in ITC Stone Serif Std and ITC Stone Sans Std by New Best-set Typesetters Ltd.

Library of Congress Cataloging-in-Publication Data

Names: Thylstrup, Nanna Bonde, editor.

Title: Uncertain archives : critical keywords for big data / edited by Nanna Bonde Thylstrup, Daniela Agostinho, Annie Ring, Catherine D’Ignazio, and Kristin Veel.

Description: Cambridge, Massachusetts : The MIT Press, [2020] | Includes bibliographical references and index.

Identifiers: LCCN 2020004704 | ISBN 9780262539883 (paperback)

Subjects: LCSH: Big data—Social aspects. | Archival resources—Management. | Uncertainty (Information theory)

Classification: LCC CD931 .U53 2020 | DDC 027—dc23

LC record available at https://lccn.loc.gov/2020004704

10 9 8 7 6 5 4 3 2 1

d_r0

Contents

Acknowledgments

Big Data as Uncertain Archives

Nanna Bonde Thylstrup, Daniela Agostinho, Annie Ring, Catherine D’Ignazio, and Kristin Veel

1 Abuse

Sarah T. Roberts

2 Affect

Marika Cifor

3 Aggregation

Sune Lehmann

4 Algorithmic Racism

Alana Lentin

5 Bots

Aristea Fotopoulou and Tanya Kant

6 Care

Daniela Agostinho

7 Complicity

Annie Ring

8 Conversational Agents

Ulrik Ekman

9 Cooling

Nicole Starosielski

10 Copynorms

Minh-Ha T. Pham

11 Database

Tahani Nadim

12 Demo

Orit Halpern

13 Detox

Pepita Hesselberth

14 Digital Assistants

Miriam E. Sweeney

15 Digital Humanities

Roopika Risam

16 DNA

Mél Hogan

17 Drone

Lila Lee-Morrison

18 Error

Nanna Bonde Thylstrup

19 Ethics

Louise Amoore

20 Executing

Critical Software Thing: David Gauthier, Audrey Samson, Eric Snodgrass, Winnie Soon, and Magdalena Tyżlik-Carver

21 Expertise

Caroline Bassett

22 Field

Shannon Mattern

23 Figura

Frederik Tygstrup

24 File

Craig Robertson

25 Flesh

Romi Ron Morrison

26 Glitch

Rebecca Schneider

27 Hashtag Archiving

Tara L. Conley

28 Hauntology

Lisa Blackman

29 Instrumentality

Luciana Parisi

30 Interface

Christian Ulrik Andersen and Søren Bro Pold

31 Intersectionality

Brooklyne Gipson, Frances Corry, and Safiya Umoja Noble

32 Latency

Kristin Veel

33 Metadata

Amelia Acker

34 Migrationmapping

Sumita S. Chakravarty

35 (Mis)gendering

Os Keyes

36 Misreading

Lisa Gitelman

37 Natural

Mimi Onuoha

38 Obfuscation

Mushon Zer-Aviv

39 Organization

Timon Beyes

40 Outlier

Catherine D’Ignazio

41 Performative Measure

Kate Elswit

42 Pornography

Patrick Keilty

43 Prediction

Manu Luksch

44 Proxies

Wendy Hui Kyong Chun, Boaz Levin, and Vera Tollmann

45 Quantification

Jacqueline Wernimont

46 Remains

Tonia Sutherland

47 Reparative

Katrine Dirckinck-Holmfeld

48 Self-Tracking

Natasha Dow Schüll

49 Shifters

Celia Lury

50 Sorting

David Lyon

51 Stand-In

Olga Goriunova

52 Supply Chain

Miriam Posner

53 Technoheritage

Nora Al-Badri

54 Throbber

Kristoffer Ørum

55 Time.now

Geoff Cox and Jacob Lund

56 Unpredictability

Elena Esposito

57 Unthought

N. Katherine Hayles

58 Values

John S. Seberger and Geoffrey C. Bowker

59 Visualization

Johanna Drucker

60 Vulnerability

Birkan Taş

61 Word

Daniel Rosenberg

Contributors

Index

Acknowledgments

The effort of many more institutions and people than can be mentioned here is required to produce a book such as this, and we are grateful for the support we have received throughout. Our deepest appreciation goes to the contributors to this book for trusting us with their inspiring work. We would like to thank the Independent Research Fund Denmark and the Carlsberg Foundation for supporting our work. We express our gratitude to Ece Elbeyi, Naja le Fevre Grundtmann, and Johan Lau Munkholm for their comments and suggestions on the introduction and some of the chapters. We thank Charlotte Johanne Fabricius and Sayuri Nakata Alsman for organizing our many events throughout the years. We thank the former research fellows of the Uncertain Archives collective for expanding the reach of the initial project and for cobuilding the supportive and lively environment that Uncertain Archives has become: Katrine Dirckinck-Holmfeld, Pepita Hesselberth, and Ekaterina Kalinina. We would also like to express gratitude to the scholars who participated in our events and helped us advance our thinking: Ramon Amaro, La Vaughn Belle, Mons Bissenbakker, Nisrine Boukhari, Mathias Danbolt, Anthony Downey, Keller Easterling, Knut Ove Eliassen, Anders Engberg-Pedersen, Anat Fanti, Maria Finn, Rune Gade, Mariam Ghani, Adam Harvey, Ben Kafka, Kara Keeling, Laura Kurgan, Lee Mackinnon, Kevin McSorley, Rabih Mroué, Lene Myong, Emily Rosamond, Antoinette Rouvroy, Evelyn Ruppert, Susan Schuppli, Hito Steyerl, Sarah Tuck, Louise Wolthers, Brian Kuan Wood, and David Murakami Wood. We thank the artists who have collaborated with us for critically expanding our imaginative horizons: Honey Biba Beckerlee, La Vaughn Belle, Katrine Dirckinck-Holmfeld, Stense Andrea Lind-Valdén, and Kristoffer Ørum. For hosting, supporting, and cobuilding the manifold Uncertain Archives events, we thank Timon Beyes, Simon Enni, Christopher Gad, Solveig Gade, Anne Kølbæk Iverson, Natalie Koerner, Lee MacKinnon, Mette Kia Krabbe Meyer, Mace Ojala, Renée Ridgway, Devika Sharma, Henriette Steiner, and Rens van Munster.

We thank the members of the Living Archives research project at Malmö University, Susan Kozel and Temi Odumosu, for fruitful collaboration and exchange. We thank the students whom we had the pleasure to teach and supervise throughout the duration of our project for engaging with our ideas and enriching them with questions. Moreover, we would like to thank the many friends and colleagues who have contributed to the supportive environment where Uncertain Archives has had the privilege to flourish, including Lene Asp, Taina Bucher, Marisa Cohn, Rachel Douglas-Jones, Ulrik Ekman, Mikkel Flyverbom, Rasmus Helles, Marianne Ping Huang, Jacob Lund, Ulrik Schmidt, Jens-Erik Mai, Annette Markham, Kathrin Maurer, Torin Monahan, Helle Porsdam, Brit Ross Winthereik, Mette Sandbye, Laura Skouvig, Sille Obelitz Søe, Karen Louise Grova Søilen, Mette Marie Zacher Sørensen, Stina Teilmann-Lock, Frederik Tygstrup, Bjarki Valtysson, Katarzyna Wac, Tanja Wiehn, and the Digital Culture research group in the Department of Arts and Cultural Studies at the University of Copenhagen. We would also like to thank Merl Storr for language editing and ensuring consistency throughout the book and Julitta Clancy for creating an extremely comprehensive index. Working with the MIT Press has been a pleasure. We are grateful to our editor Doug Sery, who recognized the potential of this book early on and helped us to develop it, and the anonymous reviewers whose helpful comments and constructive suggestions helped us shape this multivocal endeavor. And we owe many thanks to the wonderful team of professionals at the MIT Press, including Noah Springer, and Wendy Lawrence and Helen Wheeler at Westchester Publishing Services, who provided invaluable guidance on the project.

Finally, we owe infinite gratitude for the love and support of our friends, families, and communities. Nanna would like to thank Liv Bonde Graae, Georg Gammeltoft Thylstrup, Thomas Gammeltoft-Hansen, Karen Lise Bonde Thylstrup, Asger Thylstrup, her extended family, and all of her wonderful friends. Daniela would like to thank Amr Hatem, Ana Teresa Maltez, André Alves, Danilo Óscar Fernandes, Pepita Hesselberth, Katrine Dirckinck-Holmfeld, Lucian Leahu, Sébastien Lorion, Pedro Montoya Navarro, Sara Magno, and the Sorte Firkant community. Annie’s thanks go to Anna Bull, Ellen Pilsworth, Gözde Naiboglu, Marie Kolkenbrock, Ina Linge, Kristin Veel, Jens Elze, Stefanie Orphal, Fiona Wright, Adam Jacobs Dean, Philippe Marie, Victoria Camblin, Joey Whitfield, Sarah Mercier, and Leila Mukhida; to Catherine, Tess, Anne, Brendan, Louise, Barbara, and Patrick Ring, with a special mention for Emilia Jasmine Ring; and to her Pink Singers family. Catherine would like to thank David Raymond, Maria Lopez Rodas, and Mimi and Bup Raymond for the time to write and reflect amidst the chaos of children, as well as her dear friend Mushon Zer-Aviv, for brainstorming during early iterations of this project. Kristin would like to thank Rasmus Veel Haahr and Margrethe Veel Haahr for participating in building the archives of the future, as well as her forebears, who live on in archival memory and to whom she owes so much—not least a love of archives.

Working together has been a joyful and enriching experience, and we hope this book will inspire others to think, write, and organize collectively.

Big Data as Uncertain Archives

Nanna Bonde Thylstrup, Daniela Agostinho, Annie Ring, Catherine D’Ignazio, and Kristin Veel

Big data seem to promise humanity a previously unknown sense of certainty. The argument of this book, however, is that this promise of certainty from big data comes along with a whole host of uncertainties equally unknown to humankind. The big of big data refers to the sheer volume of data that can now be gathered from networked societies—data that are too numerous to be processed by human minds, and are therefore subject to analysis, and archiving, by smart machines. The huge bodies of information that big data archives contain thus augment human capacities to the powers of a prosthetic goddess, at the mere click of a button. The mass collection of data by corporations and state agencies promises, meanwhile, to make the world’s populations increasingly traceable and—it is hoped by some—predictable. In this era of big data, as the notion of the archive moves from a regime of knowledge about the past to a regime of future anticipation, big tech tells us that we have (or rather, it has) gained command of everything from trends in culture and thought to potential epidemics, criminal acts, environmental disasters, and terrorist threats. However, we argue that the data storage institutions of the present offer a false sense of security. Recent data ethics and information scandals, including those exposed by Reality Winner, Brittany Kaiser, and Edward Snowden (Agostinho and Thylstrup 2019), have caused experts and observers not only to question the statistical validity of the diagnoses and prognoses promised by big data but also to consider the broad implications of big data’s large-scale determination of knowledge. A consideration of those implications, in terms ranging from art to computation, from ethics to sociology, is the purpose of this book.

It is our argument that big data must be analyzed from a range of different disciplinary vantage points—not least from the perspective of the humanities because big data interact at every level with the human. For instance, the highly metaphorical notions of machine learning and artificial intelligence (AI), which are associated with the automated gathering and management of big data, imply an independence of technology from the human, suggesting in effect that an inherent ethical separation exists between human acts and machine capabilities. Yet the writers in this book demonstrate that concepts of disembodied intelligence are often initiated and sustained by material, embodied events and therefore rely on human factors in every way. When we go digging in the uncertain archives of big data, we find affective and material labor are absolutely central to the programming of algorithms and the selection, storage, and use of big data, in ways that require ethical consideration and are already having a huge impact on how people live all around the world. Hence, we contend in this volume that big data urgently need consideration and analysis from ethically and human-oriented perspectives.

On one level, archival uncertainties in the age of big data challenge assumptions, and indeed industries, with regard to the knowledge promised by data-driven predictions. The suggestion that big data can err—just as humans can—has subversive potential. And yet, on a structural level, big data archives are so sophisticated that they thrive on uncertainty and disruptive moments. Such moments can be viewed as moments of critique in the form of error, glitch, and subversion when the new institutions of power/knowledge built around big data are revealed as not so powerful or knowledgeable after all. Yet the conceptualization of uncertainty as a creative process can also quickly find itself in alliance with political economies in which performances of critique are easily co-opted as ventures: epistemological uncertainties are leveraged for risk management or to capitalize on the addictive incentives of digital media. In the worst cases, uncertainty can even function to deflect state and corporate accountability because uncertainty is endogenous to digital-age capitalism, and its pursuit therefore promotes and strengthens neoliberal economies and practices. From these perspectives even the subversions of uncertainty available to the contemporary archive appear vulnerable to calculated abuse. Therefore, in order to expose, counter, harness, resist, or evade the forces of today’s political and technological regimes of uncertainty, new theoretical vocabularies, methods, and alliances are needed.

This book offers several such vocabularies and alliances, and in the process it highlights the potential for collaboration across fields. It thereby insists on the importance of building communities, beginning with scholars working across the humanities, social sciences, critical data studies, and beyond. The book is rooted in a series of workshops hosted by the Uncertain Archives research group, which was formed by Kristin Veel, Annie Ring, Nanna Bonde Thylstrup, and Anders Søgaard in 2014 and supported by a strategic fund from the Danish Research Council directed at the promotion of a more balanced gender composition in research environments.¹ In its initial formation, the group encompassed literary theory (Kristin Veel), film and cultural theory (Annie Ring), media and cultural theory (Nanna Bonde Thylstrup), and computer science (Anders Søgaard). Soon Daniela Agostinho joined the group, enriching its perspectives with her knowledge of archives, visual culture, and colonial histories. While the group’s starting point was to draw on long-standing theories of the archive from continental philosophy to think about the aesthetics, politics, and ethics of big data, we were soon compelled to mediate and instigate much more interdisciplinary conversations and collective conceptualizations of big data through a broader application of the terms archive and uncertainty. We worked closely and continuously with practice-based scholars, archivists, artists, designers, activists, and computer scientists. Feminist data theorist Catherine D’Ignazio, who attended the first workshop, continued to work with us, offering the project new perspectives on how to democratize data and technology for community empowerment and social justice, as well as new feminist perspectives on data visualization. The workshop format proved to be ideal for these interdisciplinary conversations, thanks to the dialogical opportunities and intimacies it provided. Artists, activists, and academics from a range of different fields who would not normally meet were brought to the same table to share their theoretical and practical interests. In many ways these workshops mirrored this book, and the book in turn retains the impression of our exchanges. Encounters are key to our conception of this volume as a print manifestation of those exchanges—one that is easier to file in the archives of the future than the more fleeting memories of our physical copresence. Not all of the contributors to this book were present at a workshop, and not all of our workshop participants contributed a chapter to the book, but all have had an influence on the book in some way, and all are essential to its existence.

The project has also drawn inspiration from research communities working on machine learning design and human-computer interaction that have been pivotal in developing novel ways of working on and articulating contestatory AI, critical gender frameworks, and other approaches to challenging oppression in technology design processes.² Beyond these intellectual and interdisciplinary aims, the book foregrounds arts-based research as a methodology to address the uncertainties of big data archives and as a generative mode of knowledge production and critique in itself (Dirckinck-Holmfeld 2011). By featuring arts-based contributions, the book carves out an epistemic space for modes of inquiry that are motivated less by providing answers than by posing new questions—thereby exploring uncertainties productively and offering embodied, imaginative, and speculative approaches to the challenges that big data archives pose.

This introduction outlines the scope of the book—first, by situating it among theories of the archive; second, by homing in on the notion of uncertainty as it operates across the book; and third, by setting out the rationale for the book’s format. Finally, we map out thematically how the sixty-one critical keywords flow and intersect across the volume, encouraging the reader to seek further connections between them in order to think about a more ethical future for big data.

The Archives of Big Data

Archival theory has always been preoccupied with scale. Indeed, thinking about archives as a scholarly question was in many ways a response to accelerations in knowledge production due to new technological inventions (such as typewriters) that meant more knowledge could be processed, as well as broader social changes that demanded new forms of knowledge organization. The growing amount of information available in the late 1800s and early 1900s caused the British archivist Hilary Jenkinson (1922) to caution—in a chapter titled A New Problem: The Making of the Archives of the Future—that the 1900s raised at least one new question in Archive Science; one which has been little considered prior to that time: quantity (21). One major source of the accelerating growth of information was the First World War, which (among other firsts) had amassed an unprecedented and impossibly bulky holding of records (21). By 1937, Jenkinson was concerned that the post-War years have only served to emphasize the problem of accumulation in modern archives (21): "There is real danger that the Historian of the future, not to mention the Archivist, may be buried under the mass of his [sic] manuscript authorities; or alternatively that to deal with the accumulations measures may be taken which no Archivist could approve" (138).

While these archives of the twentieth century emerged as information collections on a hitherto unseen scale, the gaps in them also grew. Archives relating to women’s lives, for instance, were considered to fall outside of the need for preservation, and if preserved, the traces of these lives were confined to the w—women index entry (Sachs 2008, 665). The same strategy of omission and reductive classification applied, of course, to minority groups and those living under colonial rule. At the same time, however, certain parts of the population, including Black, immigrant, and refugee populations, were also subject to surveillance–and archiving–to a disproportionate degree, with devastating implications for the individuals concerned (Browne 2015; Gilliland 2017). Rather than effacing these injustices, digitization has in many ways exacerbated them (Thylstrup 2019). It is at the core of our argument in this book that the questions raised by big data archives belong in the longer history of archives under modernity, alongside all the injustices outlined above. We argue that while big data often appear to offer new, shiny, and automated methods that render older archival orders obsolete, big data in fact often repeat—with a difference—the epistemologies, injustices, and anxieties exemplified by previous archival orders (Spade 2015; Spieker 2017).

This book therefore addresses the large digital archives that have come to characterize our time as a very specific—but still only the latest—installment in the much longer history of archival phenomena. Big data archives, we argue, represent a long negotiation between techniques for organizing knowledge and archival subjects, between control and uncertainty, order and chaos, and ultimately between power and knowledge. At the same time, however, these highly networked repositories also challenge traditional understandings of what counts as an archive, insofar as they fail to adhere to the same logics and procedures of appraisal, preservation, and classification. As shown by YouTube’s ongoing takedowns of user-generated content documenting human rights violations in Syria, Yemen, and Myanmar, such media-sharing platforms cannot be relied on as stable archives since they do not preserve contents in an equal or reliable fashion. Instead, these new data archives are highly curated by human-machine processes that determine preservation and erasure. In the case of human rights violations, these—often deliberately—opaque processes tend to alter original content by stripping metadata away, and corporate interests tend to prioritize profit and low liability over the documentation of human rights abuses and other crimes (Roberts 2018, 2019; Saber, forthcoming). At the same time, other violent content, such as user-generated content depicting violence against people of color, is kept online, contributing to its traumatic but highly profitable virality (Sutherland 2017b; Wade 2017); the same goes for misogynistic, homophobic, transphobic, and fascist content (Breslow 2018; Chun and Friedland 2015; Gillespie 2018; Lewis 2018; McPherson 2019; Nakamura 2019; Roberts 2019; Shah 2015; Waseem, Chun, et al. 2017; Waseem, Davidson, et al. 2017). As several chapters in this volume attest, the big data archives into which information is now gathered by means of surveillance and automation display continuities with earlier archival regimes, but they also bear witness to shifts that require critical attention.

Poststructuralist Theories of the Archive

Because of the issues raised above and the certainty with which corporate big data discourses assert themselves, we respond in this book to critiques of archival reason articulated both by mid-twentieth-century poststructuralist thought and by feminist, queer, postcolonial, and critical race theories and critical archival studies up to the present day. In different ways these theoretical approaches and political movements have challenged the authority of archives as reliable repositories and questioned the capacity of archives to produce truth, offer evidence, and categorize human identities. In poststructuralist and cultural-theoretical terms, archives have always been regarded as dynamic and ultimately generative of knowledge. Rather than neutrally storing knowledge, archives produce what can be known and what becomes forgotten through what Annie Ring (2014, 390) calls their hermeneutic operations of selection, preservation, and modes of permitting (or denying) access, all of which add to the dynamism and knowledge-creating nature of archives.

That dynamism was at the heart of how theorists Jacques Derrida, Michel Foucault, and Michel de Certeau viewed archives as the places from which a crucial ordering of knowledge had proceeded since ancient times. In his foundational text on the psychic life of archives, Archive Fever, Derrida (1998) traced the etymology of the term archive to arkhe, the Greek noun signifying beginning and commandment, and drew attention to the related noun arkheion, designating the homes where ancient magistrates (archons) stored the documents of the law (2). Given this history of archives as locations intimately bound up with the issuing of laws, archives might be viewed in cultural-theoretical approaches as only authoritative: as the origins "from which order is given (Derrida 1998, 1; italics in the original), as a totalizing assemblage (50), and even as institutions that lay down the law of what can be said at all (Foucault [1969] 2007, 4). In Certeau’s (1988) account, the writing of history was the central mechanism by which modernity replaced the myths of the past with more systematized, archival knowledge systems aimed at replacing the obscurity of the lived body with the expression of a ‘will to know’ or a ‘will to dominate’ the body (6). However, even given the totalizing, ordering will of archives, and despite their historiographic logic that aims to dominate and overcome human obscurity, archives are by no means static, stable institutions invulnerable to transformation. Ring (2014) argues that Foucault’s writings present an especially dynamic vision of the archive, stemming from what he saw as an an-archic energy that pushes out from the lives that [the archive] ostensibly governed (388). Foucault ([1969] 2007) writes: Nor is the archive that which collects the dust of statements that have become inert (146). Take, for instance, the archive of informant letters dating from 1660 to 1760, an anthology of petitions to the king of France submitted by ordinary citizens seeking punishment for their errant acquaintances and family members. Along with Arlette Farge (2013)—an eminent theorist of the allure of archival research—Foucault consulted these letters in the archives of the Bastille (see Ring 2015, 133–134). Farge and Foucault’s (1982) subsequent joint text took the view that the archives were unstable based in the localized, human disorder—Foucault called this the miniscule commotion" (Foucault [1969] 2002, 167)—that emanated from the lives they were intended to make more orderly.

Even before his joint project with Farge in the archives of the Bastille—and presciently, given the technological transitions that were still nascent when he wrote The Archaeology of Knowledge in the 1960s—Foucault ([1969] 2007) identified archives as making up a web of which they [the holders of the archive] are not the masters (143). Certeau (1988), too, found potential for instability in the blank page he saw as undermining any attempt at historiography (6). Even in Derrida’s (1998) more feverishly orderly archives, there is an aggression and destruction drive (19) such that, as Ring (2014) writes, the violent patriarchive is rendered less authoritative by the haunting impossibility of its own totalizing desire (398). The critique of the violent patriarchive has also been articulated by French feminist thinkers such as Hélène Cixous and Luce Irigaray, who famously argue that systems of language, discourse, and logic, rather than being universal and natural, rest on an implied and veiled masculinity. Both Irigaray and Cixous extend Derrida’s critique of logocentrism to scrutinize and undo phallogocentrism, a combined expression of patriarchy and the representative determination of language and discourse that excludes women as thinking and speaking subjects.

The exclusionary logic of archives is often explained by French feminist thinkers through the pervasive patriarchal systems of language, discourse, and social action. In her important text The Mechanics of Fluids (in This Sex Which Is Not One), Irigaray (1985) argues that science is also symptomatic of this phallogocentrism, biased as it is toward categories typically personified as masculine (such as solids, as opposed to fluids). Irigaray’s discussion remains relevant today, when subjects speaking outside of the masculine norm keep being rendered invisible and discredited across fields, affecting the ability of their expertise, and their speaking of truth to power, to be recognized (Agostinho and Thylstrup 2019). Cixous’s critique of phallogocentrism is at the heart of her examination of archival power and her expressed desire to unravel and transform it. Akin to Derrida’s conception of immanent drives that undermine the archive’s totalizing desire, for Cixous a number of internal attributes subvert archival order from within.

In the novel Manhattan: Letters from Prehistory, written as she negotiated the donation of her own archive to the National Library of France, Cixous ([2002] 2007) refers to archival orders through the notion of omnipotence-others. The novel is populated by a number of other figures for the archive, such as the tomb, mausoleum, necropolis, hospital, trace, scar, and wound, which stand for these omnipotent qualities. But it is also from these figures that the subversive internal attributes materialize, as archival scholar Verne Harris points out. First, any corpus is bound to remain vaster than the storehouse supposed to hold it: an archive is always incomplete, defined by its exclusions and absences and by a disposition to forgetting. Second, the meanings at play in the archive’s holdings are indeterminable, subject to ever-shifting recording, imagining, narrativizing, and fictionalizing. And third, the archive has a way of subverting the conscious desires of those who create and maintain it.

As we see from these leading poststructuralist and feminist accounts of archives, one of the archive’s most important powers is that of omission, and archives have long tended to neglect and ignore groups of people and the perspectives they represent. Many archival scholars and archive theorists have subsequently pointed out that the exclusion of women from historical archives, and the difficulty of locating their subjectivities in archival sources, coexists uncomfortably with the feminized labor that has come to predominate in and sustain the archival profession (Chaudhuri et al. 2010; Mulroney 2019). Indeed, as historian Antoinette Burton (2008) notes, women’s appearance in archives is often obscured by figures considered more relevant, by large-scale events deemed more significant than those that frame their lives, and by grand narratives that may touch on contexts of significance to them but that effectively brush by them, in part because of the comparative lack of archival trace to secure them in the sightlines of history (vii). Yet at the same time, as Jessica M. Lapp (2019) notes, the feminization of the labor done in archives also holds subversive potential, whereby the handmaidens of archival work, once considered mechanical, servile, and invisible, have become powerful and disruptive, seizing opportunities for political intervention and social change.

In their work on critical feminist archive theory and practice, Marika Cifor and Stacy Wood (2017) offer us ways to unleash this subversive potential, activating the insights of critical and intersectional feminist theory to transform the potential of feminism for archives and the potential of archives for the dismantling of heteronormative, racist, and capitalist patriarchy. Such interventions must also be supported, however, by a political will to acknowledge the historical and relational links between archival material, archival knowledge, and archival labor. As archivist Kellee E. Warren (2016) notes, the scarcity of archival materials collected, owned, and governed by marginalized groups correlates with today’s archival profession, which remains largely populated by white and middle-class agents. It is therefore not enough to subject archival content to intersectional critique; we must also analyze the composition of that content’s labor and how these two factors intersect. And yet, as Cifor and Wood (2017) caution, a feminist praxis needs to be developed that aims to do more than merely attain better representation of women and minorities in particular archives. What is at stake is a praxis that fully challenges and uproots the oppressive systems that underpin archival reason and archival practices in general.

We argue that these lessons, learned from poststructuralist and more recent critical archival theory, can be productively harnessed in the field of big data studies to look at the new archives in which the crucial methods of appraisal, storage, and classification are once again being performed by a small group that exercises white patriarchal power over the rest of the world, with disproportionate impact. As Safiya Noble (2019) states, Political struggles over the classification of knowledge have been with us since human beings have been documenting history. The politics of knowledge plays out in whose knowledge is captured, in what context and how it gets deployed. Current practices of data production, collection, distribution, and consumption both build upon and draw from the history of theorizing the archive, even as they raise pertinent new questions that exceed the horizon of physical archives. To think about the politics of knowledge and of archiving in this way allows us to recognize the historical roots of current practices of data gathering, hoarding, storing, leaking, and wasting while also remembering that today’s seemingly streamlined interaction between human beings and our digital files and folders is every bit as messy, porous, and generative as archival encounters have always been.

Archival Turns and Returns

The archival operations associated with big data raise important political and epistemological questions that the archival turn in the humanities has addressed in relation to analog archives (Stoler 2002): questions about access, selection, exclusion, authority, lacunae, and silences. Both of the gestures that archives carry out—that is, selection and interpretation—gain new epistemological and political implications when we look at them in relation to the collection and use of big data. These central archival gestures have been scrutinized by scholars across different fields who have examined the limitations and possibilities of the archive. Within performance studies, thinkers such as Diana Taylor (2003) and Rebecca Schneider (2011) have questioned archival logic and its exclusion of (or failure to integrate) repertoires of embodied knowledge formed by gestures, voices, movement, flesh, and bone. At the same time, these theorizations have reconceptualized archives to propose that embodied practice offers alternative perspectives to those derived from conventional archival inscription. Such critiques are taken further by feminist and queer theories of the archive, which have pointed out that archival reason has overlooked the experiences of women and queer people and that these histories are often obscured within existing sources or discarded altogether (Stone and Cantrell 2016).

Literary theorist Ann Cvetkovich (2003) has famously argued for an archive of feelings to preserve everyday queer experiences that are difficult to chronicle through the materials of a conventional archive. In response to this claim, historian Sara Edenheim (2013) has argued that the traditional archive is often a queer body of knowledge in and of itself, a disorderly, contingently organized place rather than a site of systematic order wherein information would give itself up for easy retrieval. For feminist and queer scholars, the archive often emerges as a place for the recovery of suppressed or marginalized histories, in recuperative projects of moving from silence to more inclusive discourses.

If archive theories have questioned the failure of archives to integrate embodied experience, it has also become clear that big data archives increasingly record and sort such embodiments for the purposes of surveillance and profit. Such all-encompassing archives have never sounded creepier or more intimidating, as the subjects of large-scale surveillance increasingly recognize that being archived and mapped comes with new and unforeseeable risks. This uncertainty gives rise to new data anxieties that call for new strategies to manage both data and uncertainty (Pink, Lanzeni, and Horst 2018). At a time when powerful companies are developing technologies that will purportedly detect a subject’s emotions—and even her state of health—from the sound of her voice, it is crucial to revisit such discussions to make sense of the repertoires of embodied experience that big data archives now track on a massive scale.

Related debates have taken place within African American, Caribbean, transatlantic, and postcolonial studies in relation to the archives of slavery and colonialism. Here, too, scholars have questioned both the capture and the exclusion of people of color in and from archives and the kind of knowledge that can be gleaned from the archives of the ruling classes—archives that dehumanize those under colonial rule (Fuentes 2016; Hartman 2008; Helton et al. 2015). At the same time, these studies have reminded us that the archives of slavery and colonialism continue to inform who counts as a human subject today (Browne 2015; Gikandi 2015) while also urging us to actively mobilize archives as instruments for social and restorative justice. Finally, these studies have called forth new ways of both mining and undermining the evidence of the archive (Arondekar, quoted by Nadim in this volume) in order to restore the subjectivity denied to those accounted and unaccounted for by these archives (Fuentes 2016; Hartman 2008; Kazanjian 2016). Such contributions remain relevant to understanding and challenging the processes through which digitization and datafication subject already vulnerable individuals and communities to new harms and exclusions, disproportionate visibility, and unequal life chances (Benjamin 2019a, 2019b; Browne 2015; Gates 2011; Nakamura 2009; Noble 2018; Wang 2017). By drawing attention to the need to uproot oppressive systems and imagine new modes of redress and freedom, these reflections also prompt us to forge new imaginaries of contestation under datafied conditions.

Alongside these approaches, the field of critical archival studies has questioned the metaphorical use of the concept of the archive in humanities scholarship, where the archive emerges as an abstract, depopulated space, untouched by human labor and laborers (Whearty 2018). Critical archival scholars have urged humanities researchers to consider actually existing archives, as well as to acknowledge the intellectual contribution of archival science scholars, in order to advance critical work on archival reason (Caswell 2016). We are inspired by the call for exchange between these disciplines, and we want this book to take that work further by bringing critical archival perspectives into dialogue with fields such as critical data studies. Critical archival studies have been pivotal in questioning archival praxis from the perspectives of feminist, queer, postcolonial, and decolonial studies, drawing attention to affective responsibilities in archival practice, the often invisible and gendered labor of archivists, the materiality of digital archives, the ethical challenges of archiving sensitive material, the need to advocate for and with marginalized and vulnerable communities, and the relevance of archives for human rights and social justice (Caswell and Cifor 2016; Caswell, Punzalan, and Sangwand 2017; Cifor 2015, 2016; Cifor and Wood 2017; Ghaddar 2016; Sutherland 2017a; Williams and Drake 2017). Here, we make the case for mobilizing such critiques of archival reason and practice for a critical analysis of big data repositories. In doing so, we show that the piecing together of information in big data archives is not a neutral pursuit; that both capture and exclusion have important ethical consequences; and that archives always have been and remain, into the age of the digital, contested sites of power, knowledge, risk, and possibility.

Uncertainty as a Contemporary Condition

As the above outline of archive scholarship shows, uncertainty is inherent to archival practices: the archive as a site of knowledge is fraught with unknowns, errors, and vulnerabilities that remain equally present in, and are indeed amplified by the sheer, constitutive scale of big data archives. In our view, the uncertainty endemic to archives is enhanced by the emergence of datafication, with its complicity in systems of neoliberal global governance, authoritarian regimes, and the massive dispossession that has been wrought by wars and climate change—a global context wherein uncertainty has become a function of disruption complicit with, rather than resistant to, power. This book is therefore intended as a response to this situation; moreover, through its sixty-one entries it charts multiple routes into conceptualizing and speaking about uncertainty in the contemporary moment.

Uncertainty, and the question of what is perceived as uncertainty, has a long and complex history closely interlinked with questions of power. Uncertainty is integral to many of the scholarly strands that underlie data science, including probability and statistics. If the 1990s and early 2000s were dominated by a discourse of risk, with references to the risk economy, risk society, securitization, and a host of related terms, recent years have instead seen researchers in different disciplines beginning to call for a finer, more nuanced differentiation between uncertainty and risk (Amoore and Raley 2017; Ellison 2016; Keeling 2019; Schüll 2014). These researchers do much to demonstrate that this cultural condition gives rise to new forms of governance that work productively with uncertainty, harnessing its affective potential. Uncertainty has emerged in this way as an engine of creativity and innovation, bearing the positive potential to change how information is held and employed (Esposito 2012; Parisi 2013).

Our present times are thus imbued with a spirit of uncertainty, a mood or backdrop pertaining to social life as a whole (Appadurai 2012, 7). Uncertainty is the crux of neoliberal governance’s dilemma of security and freedom: if modern governments demanded scientific predictability, universality, and rationality, today’s globalized economies now demand a future that is open to risk-taking and not entirely calculable yet still somehow subject to a degree of control and predictability (O’Malley 2009). Nowhere is this tension between risk and uncertainty, and the presence of conflicting desires for freedom and security, more visible than in the way big data are used and stored (Amoore and Raley 2017). On one hand, private corporations and governments across the globe promote big data as effective solution to deal with informational uncertainty, risks, and unknowns. Big data’s promise of accurate calculations, precise predictions, and preemptions speaks to contemporary concerns about the taming of social, economic, financial, environmental, and political risks. On the other hand, as N. Katherine Hayles (2017) has demonstrated in her work on high-frequency trading, the very same companies and governments also exploit big data as drivers of creativity and high-gain opportunity. Uncertainty and control are therefore embraced by technocapitalism on an equal basis.

If we situate our rethinking of the archive within this regime of controlling uncertainty—which is simultaneously political, technological, and cultural—we can understand big data archives not simply as rational apparatuses but also as reflections of a political and social reality in which uncertainty is profoundly feared and yet simultaneously embraced as potentially disruptive and even desirable (Thylstrup and Veel 2020). Moreover, as Louise Amoore (2019) shows, we can also understand big data archives as raising questions about our very definitions of uncertainty and doubt in relation to certainty and truth. Uncertainties in the age of big data can feel very frightening indeed. For instance, there are clearly political and financial gains to be made from the uncertainty attached to new data regimes. The dataveillance of social media users by big data firms—such as Cambridge Analytica, which the Trump and Brexit campaigns employed for personalized political advertising via Facebook—was made possible by the unregulated, cross-border transmission of Internet user data; legal protections were only brought in post hoc, in the form of the European Union’s General Data Protection Regulation, implemented in 2018 and only partially applicable in the rest of the world. At the same time, it is questionable whether the psychographic targeting employed by Cambridge Analytica and many other big data companies was the sole reason for the referendum and election results in the US and Britain: to blame data harvesting and Russian data farms for these results risks diverting attention from the systemic racism, economic inequality, and rising fascism that corrode both countries. Meanwhile, deep fakes—hyperrealistic AI-generated videos that mimic people’s most fine-grained expressions, portraying them as doing or saying things they never did or said—introduce a frightening uncertainty that contrasts with the certainty supposedly heralded by policing regimes of prediction through facial recognition. Increasingly, however, we also see moments of resistance when individuals and collectives turn toward uncertainty, despite its co-optation by risk regimes, as a mode of contestation that can undermine archival authority. If uncertainty is taken as a value of indeterminacy that can challenge dominant regimes of control, it has the potential, given the right circumstances, to change the way information is held and employed and even to restore to marginalized individuals and groups the ability to contest dominant organizations of information.

Beyond the Glossary

The idea of creating a glossary of central concepts in contemporary data regimes was suggested twice during workshops held by the Uncertain Archives group: Catherine D’Ignazio and Mushon Zer-Aviv presented the idea of a dictionary of uncertainty as an early art/design intervention for the project at the first workshop; later, in 2017, Orit Halpern independently suggested the idea of a dictionary of uncertainties for big data at a postworkshop lunch. We combined these forces and ideas, and we subsequently consulted and were inspired by other projects that use keywords as a format for knowledge production across disciplines—for instance, Speculation Now, edited by Vyjayanthi Venuturupalli Rao, Prem Krishnamurthy, and Carin Kuoni; Posthuman Glossary, edited by Rosi Braidotti and Maria Hlavajova; Cultural Anthropology’s Lexicon for an Anthropocene Yet Unseen, edited by Cymene Howe and Anand Pandian; The Infrastructure Toolbox, edited by Hannah Appel, Nikhil Anand, and Akhil Gupta; Environmental Humanities’ Living Lexicon for the Environmental Humanities; Software Studies: A Lexicon, edited by Matthew Fuller; and Transgender Studies Quarterly’s Keywords. We also realize that this form of knowledge production has a longer lineage in single-author endeavors such as Raymond Williams’s classic Keywords and in multiauthor projects such as the cultural-theoretical lexicon provided by Theory, Culture and Society with the apt title Problematizing Global Knowledge, edited by Mike Featherstone, Couze Venn, Ryan Bishop, and John Phillips with Pal Ahluwalia, Roy Boyne, Chua Beng Huat, John Hutnyk, Scott Lash, Maria Esther Maciel, George Marcus, Aihwa Ong, Roland Robertson, Bryan Turner, Shiv Visvanathan, and Shunya Yoshimi. What this book has in common with these projects is a desire to problematize, stake out, and contribute new knowledges and perspectives, as well as to develop new forms of knowledge production.

Uncertain Archives: Critical Keywords for Big Data is a glossary with essay-length entries that set out, each in turn, to interrogate the meanings of a particular term. Each entry contains its own mini bibliography of works cited that may be taken as a starting point for further reading. Form and content are thus intertwined, in ways that make the book in its multivocal and dialogical nature a performative enactment of the uncertainty of archives. This performative dimension is perhaps best captured by reading the glossary as a heteroglossia, a term stemming from Russian literary theorist and linguist Mikhail Bakhtin’s translation of raznorechie, meaning different speechness. Heteroglossia is a broader concept than polyphony since it describes the interaction not only of coexisting but also of conflicting voices, linked to different cultural meanings, ideologies, and materialities within the same linguistic space. We regard this book as such a centrifugal force, spinning out and thereby performing a diversification of thinking about datafication by circling the multiplicity of ways in which the uncertainty of big data archives can be identified and conceptualized.

The book goes beyond the glossary format by offering interdisciplinary and transdisciplinary dialogues between scholars, activists, and artists from a multitude of disciplines. All of them make contributions that will be valuable to an interdisciplinary audience working on datafication and big data archives, and they refer to each other’s chapters to build their arguments throughout the book. In this way, the book takes forward a conversation that we deem acutely necessary to address our age of big data, the uncertainties big data bring, and their political, ethical, and social implications, as well as reflections of those implications in cultural expressions of all kinds. Therefore, the book aims to make a significant intervention in the field of critical data studies by reconceptualizing and infusing new meaning into established terms while also proposing new concepts to make sense of the epistemological, political, and ethical dimensions of big data.

Many terms included in this glossary have different meanings depending on the disciplines that use them. For instance, terms such as aggregation (Lehmann), executing (Critical Software Thing), latency (Veel), proxies (Chun, Levin, and Tollmann), reparative (Dirckinck-Holmfeld), values (Seberger and Bowker) have various, sometimes disparate, meanings in different disciplinary contexts. While acknowledging that these terms are complex in their polysemy, we believe that apparently inconsonant frameworks can generate especially fruitful interdisciplinary dialogues precisely by virtue of the uncertainty of their differing terminologies. Therefore, the glossary seeks to bring together differing meanings in order to shed light on the multidimensionality of big data and to stress the necessity for an interdisciplinary approach when one engages with big data phenomena and their reverberations. Furthermore, by expanding the study of big data to fields such as literature and art history, which have only recently begun to inquire into these phenomena, the book posits that the study of big data has significance beyond conventional disciplinary interests, as the gathering and use of data increasingly shape our everyday lives. The book’s recognition of the inherent transdisciplinarity of the terms it includes thus challenges the limits of available vocabularies; it also offers valuable counterweights to the centripetal forces of data informatics as they aim to integrate, tame, and optimize everything. Here, we offer a nonexhaustive mapping of some of the flows and constellations that emerge from the book as a whole. We invite our readers to seek out even more connections and clashes.

Thematic Flows and Constellations

The book’s guiding assumption is that any issue related to big data is interdisciplinary at heart, such that the topic of big data requires a response rooted in collective and collaborative forms of knowledge production. As a result, the book’s approach is collective: it adopts the shared frameworks of intersecting critical languages to inject a productive uncertainty into the vocabulary of digital studies. It does so by pointing to terms that escape the disciplinary fixities and modes of engagement that tend to separate otherwise intersecting processes of knowledge production and political issues into distinct fields. For instance, we argue that computational errors are at once material phenomena, epistemological concepts, and cognitive processes and that we need to regard them as such in order to understand their complexity.

These new systems of complexity not only span a spectrum from pure mechanical operations to human thought processes but also entangle these in complex new ways. Such systems are invariably called AI, machine learning, and more, but as N. Katherine Hayles (unthought) and Luciana Parisi (instrumentality) show in their chapters, they are infinitely more complex than the often reductive discourses about AI as merely smart. These systems not only force us to rethink what we mean by intelligence and cognitive processes but also—as Caroline Bassett (expertise), Celia Lury (shifters), and Ulrik Ekman (conversational agents) show—fundamentally reorder our systems of expertise, agency, and interaction.

Moreover, they give rise to new questions about representation and perception, as demonstrated in the entries by Johanna Drucker (visualization), Daniel Rosenberg (word), Frederik Tygstrup (figura), Kristoffer Ørum (throbber), and Christian Ulrik Andersen and Søren Bro Pold (interface). These shifts raise questions not only of epistemology and ontology but also, fundamentally, of power. What does it mean, for instance, to misread data? Lisa Gitelman muses on this question in her contribution on misreading, showing how this seemingly mundane term in fact raises questions central to big data studies in terms of both precision and power. As Gitelman notes, transposing two numerals is misreading as fact, while failing to understand an author’s statement can be misreading in much less certain terms. Moreover, the outcome of such misreading depends on who is in power since misreading as misunderstanding lies in the eye of the beholder: it depends, as Gitelman notes, on who calls the shots. And as Amelia Acker points out (metadata), we must ask not only what it means to classify and misclassify something but also who gets to determine and uphold structures of knowledge.

Such questions about knowledge organization and the disciplinary and political structures that uphold it are not only of practical importance; as outlined by Miriam Posner (supply chain), Nanna Bonde Thylstrup (error) and Timon Beyes (organization), they are also a matter of imaginaries and mythologies. These imaginaries often open up to the softer power dynamics of seduction and complicity. Indeed, as Wendy Hui Kyong Chun has previously shown, the cyberspace imaginary was predicated on seductive and sexualized imaginaries from its inception (Chun 2011). In his entry on pornography, Patrick Keilty discusses PornHub as a supreme example of the intersection of big data archives and desirous imaginaries, making explicit big data industries’ vast surveillance powers as well as their reliance on the public fascination with data’s promise to reveal something about us to ourselves. On one hand, such surveillance regimes rely on not only meeting data subjects’ desires and imaginaries but also preempting them. As Manu Luksch shows in her contribution on prediction, data capitalism thus relies on harnessing the imaginary powers of predictive technologies, often with devastating effects. Yet, on the other hand, as Luksch also reminds us, the subversive power of the critical imagination can never be fully mapped, archived, or controlled. Marika Cifor describes how these twin logics of empowerment and exploitation coexist in her analysis of affect in relation to the AIDS Memorial. In the entry on executing, the Critical Software Thing collective offers us one meditation on how such resistance appears, showing how formal systems and affective bodies expose their own indeterminacy in computational systems. Such moments of evasion and refusal can be strategically engineered as obfuscation (Mushon Zer-Aviv), but they are also a fact of life itself, full as it is of unpredictability, unknowns, and unknowables (Elena Esposito).

In addition to cultural imaginaries, datafication processes also rely on the material. In her entry on performative measure, Kate Elswit shows how data get visceral to approximate, measure, and capture even the most basic, sustaining processes of human life respiration. Information converges and clashes in big data archives not only on the level of political ideologies but also across different materials: papers and shelves intermingle with wires, fingers with touch pads, sweat with blood, and dust with DNA. As Mél Hogan points out, the speculative imaginaries of information storage and transmission are transforming, from the materiality of metals to the body itself as an archive from which value can be extracted. Her contribution on DNA explores current attempts to use DNA as an archive, as well as how the speculative work of these scientists reflects a much longer-lived archival desire to conserve and theorize the vast traces of thought and feeling that forever evade the archive—the secret, lost, destroyed, and dismembered. In contrast to the microscopic scale of DNA, Shannon Mattern’s entry on field shows how archives emerge at the level of the Anthropocene, materially compressing multiple temporalities and bringing together ancient and cutting-edge forms of preservation. In this respect, archives emerge against a backdrop that combines the long-term temporal spans of evolution with the short-term span of digital circuits. These material archives, as Nicole Starosielski shows in her entry on cooling, are uncertain in not only informational but also material terms, and they are dependent on the uncertainties of climate change even as they map those very uncertainties. Contributions by Orit Halpern (demo) and Lila Lee-Morrison (drone) demonstrate how technological cultures work to uphold regimes of material violence and military engagement.

At the heart of these entanglements of new and old power structures, imaginaries, and materialities lies the question of how temporality as (sometimes imagined) past actions sticks to and coproduces imagined and lived-out futures. As Geoff Cox and Jacob Lund argue in their chapter on time.now, new machinic systems inaugurate a new temporal order through their material operations. But what happens when the archive is over? Lisa Blackman outlines the concept of hauntology to help us understand the ghostly agency that data retain long after they have fulfilled the instrumentalist purpose of their collection. And as Kristin Veel shows in her entry on latency, such temporalities are a question not only of materiality but also of mental topographies. New sociomaterial temporalities have given rise to new ethical dilemmas and problematics that pertain not least to negotiations of power.

Combining these folds and transformations of archival temporalities, materialities, and imaginaries gives rise to new mechanisms of oppression and empowerment that sometimes clash and sometimes intersect with older structural forms of oppression. One pervasive structural force is the persistence of coloniality, alluded to in different ways by Tahani Nadim’s discussion of database imaginaries and Roopika Risam’s notion of the digital cultural record within digital humanities. As Minh-Ha T. Pham shows in her chapter on copynorms, the seemingly technical question of copyright—the legal governance infrastructure that subtends much of the digital economy—is as much a normative framework shaped by colonialism and racism as it is a juridical figure upholding globalization’s technological infrastructures. Similarly, by reading mapping alongside migration, Sumita S. Chakravarty illuminates how the mapping mechanisms that are such a pervasive feature of our everyday computational systems are underpinned by both technological affordances and colonial ambitions, raising questions not only about how things and people can be mapped but also to what end, by whom, and to what effect (migrationmapping). In her contribution on technoheritage, artist and researcher Nora Al-Badri, cocreator of the Nefertiti Hack, shows how cultural heritage institutions and their digitization practices extend colonial legacies of looting, dispossession, and gatekeeping and asks how technology can be radically mobilized toward decolonization and restitution.

Critical race theorist Alana Lentin also claims that the assumed neutrality of the algorithm serves to obscure the underside of modernity—its racialized subjects and colonized others. In her discussion of algorithmic racism, Lentin makes the crucial point that networked communications do not only contain racial bias within them; in fact, racism is integral to the way technology is operationalized, just as it is integral to liberal societies. Extending earlier studies by Lisa Nakamura (2008), Wendy Hui Kyong Chun (2012), and Safiya Noble (2018) on race, technology, and the Internet, Lentin also points to how the deployment of algorithms supports the proliferation of hate speech online. The widespread belief in algorithms’ better ability to manage outcomes—because they are supposedly immune to racial and other forms of bias—is consistent with the notions of free speech and the neutrality of platforms and with the view that all ideas deserve an airing, leading to a constant stream of racist, sexist, homophobic, and transphobic content that does not need to be regulated. As Tonia Sutherland explores in her entry on remains, racism also structures the logics of digital and datafied archives in the ways in which Black people and communities continue to be rendered as commodities and spectacle—even after their deaths, when their remains are resurrected for profit through digital afterlives. In her entry on natural, artist Mimi Onuoha notes that data-driven stories concerning Black subjects too often begin from an assumption of disenfranchisement and perpetuate deficit narratives about brutality and suffering, a tendency exacerbated by advances in corporate and state surveillance applications of machine learning and automated decision-making systems. Rather than calling for a redress of such narratives in the form of archival counternarratives, Onuoha seeks to carve out space for herself, within digital space, where she can devise her own mode of existence.

Colonialism’s and racism’s pervasive archival infrastructures of power are always accompanied by gendered dynamics of oppression and discrimination. These come in the form of making-invisible, sexualization, infantilization, and misrecognition in several layers of technologies and logics, dealt with in contributions by Catherine D’Ignazio (outlier), Miriam E. Sweeney (digital assistants), and Aristea Fotopoulou and Tanya Kant (bots). As Craig Robertson’s media-historical contribution on file shows, the logic of disembodiment in information handling is shaped by a historical trajectory in which women were once visible as infantilized yet capable information handlers, only to disappear from view as laboring bodies receded into the background of computer systems, which today take center stage. Rebecca Schneider, meanwhile, uses the concept of the glitch to describe the aesthetics of error employed by feminist performance artist Carolee Schneeman to rupture social norms and routine operations.

Media history shines a light on the making invisible of women in the field of information-handling and computational systems, but gendered lines of oppression also cross through the logic of quantification that structures big data archives and datafication practices. As Os Keyes points out in their essay on (mis)gendering, these lines of oppression remain lodged in the binary imaginary of data science, which at once excludes trans experience from its binary organization of information and at the same time keeps trans people looped in through static gender narratives drawn from archival material from their pretransition lives. Therefore, as Jacqueline Wernimont points out, the logic and practice of quantification is never merely descriptive but is always already engaged in the processes by which bodies and people have become and are becoming visible to themselves, to others, to nation-states, and to globalized governance regimes. Indeed, as Olga Goriunova shows in her entry on stand-in, such standardization processes produce the imaginary of an average that never aligns with the richness of empirical reality and that indeed looms as a threat to those who find themselves replaced by standardized models.

The visibilization afforded by quantification is often given in terms of imperfectly counted steps, pounds, and heartbeats, and the value of data is often calculated through the lenses of late capitalist paradigms. Yet, as Wernimont notes, this is sometimes preferable to the condition of those excluded from even this basic form of recognition. Thus, as Wernimont notes with reference to Diane Nelson, quantification is both essential and insufficient, dehumanizing and reparative, necessary and complicated (Nelson 2015, xi). Contributions by David Lyon (sorting) and Sarah T. Roberts (abuse) explore the ways in which subjects are made vulnerable differentially through their interactions with, and subjection to, big data regimes. Yet, as Birkan Taş points out, there also lies great power in reclaiming and embracing such vulnerability. Taş shows that by integrating disability as a critical category of analysis, we might broaden our understanding of big data and, more importantly, open up modes of engagement and care that have potential for a politics of noncompliance.

Increased concern about the harmful effects of the predictive and preemptive algorithmic extraction and analysis of data has brought ethics to the center of public discourse on big data. Over the last couple of years, several research projects and institutes devoted to data ethics have started to flourish, national and supranational councils on data ethics have been founded, and corporations have begun to integrate data ethics into their vocabularies and policies. Researchers and critics regard these developments with suspicion, worrying that the corporatization and legislation of ethics has resulted in an impoverished understanding of ethics that centers individual responsibility and institutional liability rather than confronting—and working to redress—structural discrimination and power differentials. Concepts such as bias, fairness, accountability, and transparency increasingly come under scrutiny as locating the source of discrimination in individual behavior or