Uncertain Archives: Critical Keywords for Big Data

edited by Nanna Bonde Thylstrup, Daniela Agostihno, Annie Ring, Catherine D’Ignazio and Kristin Veel, The MIT Press, 2021, 640 pp. 
Paperback $55.00, ISBN: 978-0-2625-3988-3.

Uncertain Archives

The narrative around big data has been largely shaped by a quest for certainty, built upon expectations of algorithmic efficiency, perfect predictive models, and strategic decision-making. However difficult to attain, these cyber-utopian visions of objectivity, order, and control have directed countless techno-cultural experiments with the unprecedented mass of data, information, and content found online. This book, by contrast, emphasizes the value of uncertainty, using a feminist postmodern perspective to read unknowns, flaws, errors, and instabilities as generative moments of learning and critique. It does so by centering the debate on big data around a rich tradition of archival research and practice, with its dynamic, instrumental role in selecting, preserving, enabling, and sometimes even denying access to knowledge. 

Originating from a series of workshops led by the Uncertain Archives research group (2015-2019, University of Copenhagen), the book brings together 73 scholars from a variety of disciplines across North American and European institutions to construct a shared vocabulary for digital archives and critical data studies. The result is an interdisciplinary, collaborative, generative work that benefits from the collective effort of five authors— Nanna Bonde Thylstrup, Daniela Agostinho, Annie Ring, Catherine D'Ignazio and Kristin Veel— who edited and organized the volume into 61 short chapters about big data, archival uncertainties, and new forms of knowledge production. The book is arranged into a bipartite structure, with an introductory section containing the overview, context, and scope for the publication, followed by a list of keywords presented in alphabetical order and defined by different writing styles and academic approaches.  

The introduction, entitled Big Data as Uncertain Archives, situates the volume at the intersection of post-structural archival theories and the social, economic, and political conditions of uncertainty in big data environments. It then discusses the idea of a glossary or dictionary, first as envisioned by Catherine D’Ignazio and Mushon Zer-Aviv, and later in a contribution by Orit Halpern. The introduction also serves as a conceptual guide to the thematic flows and possible readings that emerge when connecting keywords and tracing overlapping discourses throughout the book. The authors provide a panoramic overview of the book’s main themes, before presenting a list of stand-alone keywords. Although fragmented, this work is not just a semantic exercise for facilitating interpretations, but a more complex semiotic operation that accounts for a network of signs, messages, and references in a data-driven, algorithmic culture. 

Among the innovative narratives around big data offered in this volume, the authors identify terms that acquire new meanings in connection with artificial intelligence and machine learning technologies (unthought, instrumentality, conversational agents, expertise, shifters), along with entries that reflect on the issues of representation, perception, interpretation, and categorization associated with big data (visualization, figura, throbber, interface, misreading, metadata). On a less abstract level, the authors also outline keywords that reason on forms of knowledge organization and their societal impact, such as supply chain, error, organization, pornography, prediction, executing, obfuscation, and unpredictability. Uncertain, and yet still so material, big data archives are physical, spatial, and temporal, always profoundly intertwined with definitions of agency, ethical dilemmas, and power dynamics (cf. performative measure, field, cooling, demo, drone, time.now, hauntology, latency).  

Even more, these keywords show how digital archives are directed by mechanisms of technological affordances and embodied cognition that affect our imaginaries (database, digital cultural record, copynorms, migrationmapping, technoheritage) and often contribute to systemic forms of oppression (algorithmic racism, remains, natural, outlier, digital assistant, bots, file, glitch, [mis]gendering, quantification, stand-in, sorting, abuse, flesh). Each of these themes draws upon the need for defining an ethics of big data, while also discussing new methods and approaches to the massive amount of data (hashtag archiving, self-tracking, detox, complicity). This co-curated lexicon is meant to repurpose, resignify, and recontextualize concepts, rather than to repeat them. In this list of words there is a system of meaning that is different from the mainstream terminology of the digital age.  

The lack of a conclusion gives the reader a sense that this is a work in progress, focusing on questions rather than definitive answers, and making uncertainty not only a mode of inquiry but also the book’s main feature. These fragments of discourse are left open to reassemble as separate anthological pieces that can be grouped or ungrouped, read in both linear or non-linear fashions. The alphabetic order provides an arbitrary organization for the chapters; however, many paths of interconnected concepts can be traced by following an irregular order of chapters. In this conceptual patchwork, the epistemologies of pre-digital archival practices blend with the contemporary organizational structures of big data repositories. Most thematic strands are concerned with the impact of big data at the individual and social level, or the role of algorithms in defining the personal and political sphere. These passages alternate with chapters dedicated to feminist, queer, and postcolonial methodological explorations (affect, care, intersectionality, reparative, vulnerability). This inquiry in archival theory and methods ultimately responds to the need for addressing uncertainty not only as an epistemological state, but also a core ontological feature of big data. 

In the complex operation of mapping digital culture while advancing ethical questions, Uncertain Archives creates an archive of its own, made of critical terms, concepts, and definitions. It is a textual archive of big data archives’ fears, hopes, failures, and discomforts. By embracing the form of a list, this book marks the beginning of a process of acknowledgement of our own biases, one that actively generates knowledge through a vocabulary open for expansion, rewriting, and redefinition. Rehabilitating uncertainty as a framework for learning, Uncertain Archives is a vademecum not only for academic researchers and media practitioners engaged in critical data or archival studies, but also for anyone who is willing to explore the complexities of big data archives at the crossroad between humans and machines. 

Giulia Taurino, Northeastern University