Social Media and the Automatic Production of Memory: Classification, Ranking, and Sorting of the Past

by Ben Jacobsen and David Beer, University of California Press, 2021, 168 pp.
Hardcover $19.95, eBook $19.95, ISBN: 978-0-5203-4490-7

Social Media and the Automatic Production of Memory: Classification, Ranking, and Sorting of the Past by Ben Jacobsen and David Beer

In Social Media and the Automatic Production of Memory, Ben Jacobsen and David Beer examine the pervasive power of social media to intervene in one of the most intimate aspects of our lives: how we remember. The book opens with Walter Benjamin’s famous provocation comparing a person who remembers with an archeologist who must dig their own past actively, repeatedly, and methodically to unearth meaningful memories. Benjamin’s framing, of authentic memories as products of meticulous labor, is a point of departure for Jacobsen and Beer to examine the shift towards the automatic production of memory. How are algorithms increasingly deciding for us what is worth reminiscing? What are the processes behind the automation of memory-making? What are the implications when the job of excavating memories is done for us by automated systems? These are the questions that the authors set out to answer in this book.  

The introduction chapter contextualizes and situates the algorithmic facilitation of memory-making at the intersection of two broader, interrelated trajectories: the expansion of metric-based sorting that strives to turn everything, even things as intangible as memories, into numbers; and the rise of targeting that predicts and personalizes content for users. This dual process of metricization and targeting, argue Jacobsen and Beer, judges memories based on their relative worthiness in terms of the commercial imperatives driving social media services to sustain users on their platforms.  

Chapters 2 and 3 turn to the underlying mechanisms that classify and rank intimate digital traces, using Facebook’s throwback feature as a case study. Analyzing Facebook’s own published research on the development of its Memories feature, the authors highlight how its algorithms fundamentally participate in delineating our field of visibility, circumscribing which content users would encounter as memories and what counts as meaningful reminiscence. Chapter 2 starts with the classification process, detailing how Facebook developed what it calls the “taxonomy of memory themes.”  This is achieved by a threefold approach: (1) asking a small number of participants to categorize memory into themes; (2) verification by quantitative surveys; and (3) supplementation with linguistic analysis to understand the kinds of semantic and discursive attributes of the so-called “good memories.” Drawing on Jacques Rancière’s notion of partitioning the sensible, the authors conceptualize Facebook’s classificatory technique as partitioning the memorable, (re)shaping “what can and should be remembered, celebrated and shared (p. 93).”  

Chapter 3 discusses the targeting of digital traces—the way in which past events are swept to the surface through the process of ranking. In this process, automatic filters are developed to winnow memory fragments deemed negative—ones that deviate from the platform’s intertwined logic of positivity and shareability such as painful experiences, broken relationships, or accidents. Here, the “Taxonomy of the themes” proves useful again as it informs Facebook to assign different weights to different kinds of memories, often in the way that valorizes fun and positive life moments while silencing others. Jacobsen and Beer go on to draw on Pierre Bourdieu’s (1984) classic Distinction: A Social Critique of the Judgement of Taste and his notion of “ontological promotion” to examine the effects of ranking on memory practices. Specifically, the authors argue that ranking participates in the ontological promotion of certain data points, leveraging certain past data to “memory status.” As a result, Facebook’s throwback feature has come to serves as the archive of everyday life—the means by which people encounter their past and understand their own selves.  

Chapter 4 shifts the focus from technological processes to human processes, drawing attention to how users receive and negotiate with the automatic production of memories. Drawing on data from original in-depth interviews and focus groups, the authors show that memory practices facilitated by classification, ranking and targeting are not necessarily a straightforward process but fraught with tensions and frustrations. For example, participants have recounted how they felt their attentions have been orchestrated in certain directions; how the resurfaced memories are at times not relevant to their own experiences of the pasts, or simply not what they want to remember; and finally, how their privacies have been invaded on. In this chapter, however, the authors seem to contradict what they have argued in previous chapters. Specifically, despite the routine claims throughout the book about how memory features of social media “reshape” memories and memory practices, the authors now argue “being targeted by memories on social media is not a passive experience.” Instead, the authors suggest that this is an “interactive and iterative process of interpretation and reinterpretation of ‘imaginative reconstruction’ of the past and our relationship to that past (p. 64).” Surely the relationship between technology and human agency is not one of total technological determination, yet the dialectical relationship between the two could have been more clearly addressed.  

Another important gap is the broader impacts of the automatic production of memory, both on individual and collective levels. In the introduction chapter, the authors alluded to Deborah Lupton’s concept of “materializing data,” which does not simply consider data as immaterial abstractions but instead sees them as having material consequences. Although the authors did go to some length in explaining Lupton’s framework, material implications, if any, are only mentioned in passing here and there. And as though to acknowledge this gap, the authors state their findings only “provided an avenue to explore and examine the practical effects of processes of classification and ranking on memory practices (p. 60).”  

Despite these shortcomings, the book offers a simultaneously fascinating and accessible account of the technological process behind and various conceptual lenses through which to understand the role of classification and ranking processes in memory making. The book comes at a crucial moment when automation through algorithmic systems is widespread: it alerts us to the next terrain that these automated systems are likely to colonize: our memory; hence our biography, our sense of selfhood, our pasts, presents, and futures. 

Trang Le, Monash University