Digital companion species and eating data: Implications for theorising digital data–human assemblages

Abstract

This commentary is an attempt to begin to identify and think through some of the ways in which sociocultural theory may contribute to understandings of the relationship between humans and digital data. I develop an argument that rests largely on the work of two scholars in the field of science and technology studies: Donna Haraway and Annemarie Mol. Both authors emphasised materiality and multiple ontologies in their writing. I argue that these concepts have much to offer critical data studies. I employ the tropes of companion species, drawn from Haraway, and eating data, from Mol, and demonstrate how these may be employed to theorise digital data–human assemblages.

Introduction

While an intense interest in digital data in popular and research cultures is now evident, we still do not know enough about how people interact with, make sense of and use the digital data that they generate. A developing body of literature in critical data studies has begun to focus on elements of these practices. Sociologists and anthropologists have pointed to the highly contextual aspects that shape human-data encounters, including their specificities and contingencies (for example, Nafus, 2014; Taylor et al., 2014). Other researchers have identified a growing concern among citizens about data privacy and security (Madden and Rainie, 2015; Pew Research Center, 2014). Everyday data practices remain under-theorised, however: particularly in relation to how people conceptualise, gain purchase on and perform the data assemblages that are generated by and for them. In this commentary, in attempting to identify and think through some of the ways in which sociocultural theory may contribute to understandings of data practices, I develop an argument that rests largely on the work of two scholars in the field of science and technology studies: Donna Haraway and Annemarie Mol.

Haraway: Digital assemblages as companion species

I begin with Haraway, as the legacy of her writing underpins some of Mol’s work. Haraway’s work has often attempted ‘to find descriptive language that names emergent ontologies’ (Haraway, 2015: np). I use her ideas here in the spirit of developing new terms and concepts to describe people’s encounters with digital data. Haraway emphasises that humans cannot be separated from nonhumans conceptually, as we are constantly interacting with other animals and material objects as we go about our daily lives. Her writings on the figure of the cyborg (for example, Bhavnani and Haraway, 1994; Haraway, 1991, 1995) have been influential in conceptualising human and computer technological encounters. In this work, Haraway draws attention to the idea that human ontologies must be understood as multiple and dynamic rather than fixed and essential, as blurring boundaries between nature and culture, human and nonhuman, self and other. She contends that actors, whether human or nonhuman, are never pre-established; rather they emerge through relational encounters (Bhavnani and Haraway, 1994). The cyborg metaphor encapsulates this idea, not solely in relation to human–technology assemblages, but to any interaction of humans with nonhumans.

This perspective already provides a basis for thinking through the emergent ontologies of digital data assemblages. Haraway’s musings on human and nonhuman animal interactions (Haraway, 2003, 2008, 2015) also have resonance for how we might understand these assemblages. Haraway uses the term ‘companion species’ to describe the relationships that the human species has not only with other animal species but also with technologies. Humans are companion species with the nonhumans alongside which they live and engage, each species learning from and influencing the other, co-evolving. Haraway (2015) refers to companion species as ‘post-cyborg entities’, acknowledging the development of her thinking since her original cyborg exegesis.

This trope of companion species may be taken up to think about the ways in which humans generate, materialise and engage with digital data. Thrift (2014) has described the new ‘hybrid beings’ that are comprised of digital data and human flesh. Adopting Haraway’s companion species trope allows for the extension of this idea by acknowledging the liveliness of digital data and the relational nature of our interactions with these data. Haraway has commented in a lecture that she has learned

through my own inhabiting of the figure of the cyborg about the non-anthropomorphic agency and the liveliness of artifacts. The kind of sociality that joins humans and machines is a sociality that constitutes both, so if there is some kind of liveliness going on here it is both human and non-human. Who humans are ontologically is constituted out of that relationality. (Haraway, 2015: np)

This observation goes to the heart of how we might begin to theorise the liveliness of digital data in the context of our own vitality, highlighting the relationality and sociality that connect them.

These aspects of digital data assemblages are perhaps becoming even more pronounced as the Internet of Things develops and humans become just one node in a network of software, digital data repositories and smart objects that configure and exchange digital data with each other. Humans move around in data-saturated environments and can wear personalised data-generating devices on their bodies; including not only their smartphones but objects such as sensor-embedded wristbands, clothing or watches. The devices that we carry with us literally are our companions: in the case of smartphones regularly touched, fiddled with and looked at throughout the day. In distinction from previous technological prostheses, mobile and wearable devices are also invested with and send out continuous flows of personal information. They have become the repositories of users’ communications with others, geolocation information, personal images, biometric information and more. These devices also leak data outwards, transmitting them to computing cloud servers.

All this is happening in real-time and continuously, and in some cases without the users’ knowledge or consent. This raises important questions about the security and privacy of the very intimate information that these devices generate, transmit and archive (Seneviratne et al., 2015; Tene and Polonetsky, 2013). Political issues concerning data ownership are also emerging. Critical data scholars have identified the asymmetries in the access of citizens to digital datasets (including their personal data) and that of government and commercial entities (Andrejevic, 2013; Crawford and Schultz, 2014).

Like companion species and their humans, digital data are lively combinations of nature/culture. Digital data are lively in several ways. They are about life itself (details about humans and other living species); they are constantly generated and regenerated as well as purposed and repurposed as they enter into the digital knowledge economy; they have potential impacts on humans’ and other species’ lives; and their commercial and research value contributes to livelihoods (Lupton, 2016). Just as digital data assemblages are comprised of specific information points about people’s lives, and thus learn from people as algorithmic processes manipulate this personal information, people in turn learn from the assemblages of which they are a part. The book or potential dating partner choices that Amazon and OKCupid offer, the ads that are delivered to users on Facebook or Twitter, the returns that are listed from search engine queries or browsing histories, the information that fitness trackers provide about users’ heart rate or calories burnt each day are all customised to the users’ digitised behaviours and preferences. Perusing these data can provide people with insights about themselves and may structure their future behaviour.

The companion species trope recognises the inevitability of our relationship with our digital data assemblages and the importance of learning to live together and to learn from each other. It suggests both the vitality of these assemblages and also the possibility of developing a productive relationship, recognising our mutual dependency. We may begin to think about our digital data assemblages as companion species that have a life of their own that is beyond our complete control.

These proliferating digital data companion species, as they are ceaselessly configured and reconfigured, emerge beyond our bodies/selves and into the wild of digital data economies and circulations. They are purposed and repurposed by second and third parties and even more actors beyond our reckoning as they are assembled and reassembled. In effect, they are digital data-human assemblages. Even as our digital data companion species engage in their own lives, they are still part of us and we remain part of them. We may interact with them or not; we may be allowed access to them or not; we may be totally unaware of them or we may engage in purposeful collection and use of them. They have implications for our lives in a rapidly growing array of contexts, from the international travel we are allowed to undertake to the insurance premiums, job offers, discounts or credit we are offered (Crawford and Schultz, 2014; Robinson et al., 2014).

Mol: Eating digital data

The work of Annemarie Mol offers an additional conceptual framework with which to understand digital data practices at a more detailed level while still retaining the companion species perspective. Mol advocates for the use of ontologies as a structuring concept in developing understanding of the relationship between humans, technologies and knowledge practices. Her focus is on the performative nature of knowledge, or how these knowledge assemblages are enacted in practices and configure different versions of objects, hence creating different ontologies of these objects.

If digital data–human assemblages are positioned as objects, Mol’s emphasis on multiple and shifting ontologies is a basis for further theoretical work on lively data. Mol has developed a framework that incorporates elements of enquiry that can be mapped onto the topic of digital data practices. These include the following: understanding language/discourse and its context and effects; tracing the development and use of objects of knowledge as they become objects-in-practice; acknowledging the dynamic nature of processes and the ‘endless tinkering’ that is involved in processes; incorporating awareness of the topologies or sites and spaces in which phenomena are generated and used; and finally, directing attention to the lived experiences or engagements in which practices and objects are understood and employed (see, for example, Mol, 2015; 2002; Mol and Law, 2004).

If this approach is applied to digital data practices and their configurations, then focusing attention on the language that is employed to describe digital data, viewing digital data as objects that have both discursive and material effects and that are constantly changing, recognising the processes of tinkering (experimenting, adapting) that occur in relation to digital data and the spaces in which these processes take place are all important to developing an understanding of the ontology of digital data and our relationship with them.

Mol’s (2002) concept of ‘the body multiple’ has resonances with the Haraway’s cyborg ontology. This concept recognises that the human body is comprised of many different practices, sites and knowledges. While the body itself is not fragmented or multiple, the phenomena that make sense of it and represent it do so in many different ways so that the body is lived and experienced in different modes. Her research on a particular medical condition, atherosclerosis, demonstrated that multiple ways of seeing, monitoring and treating a medical condition (or the different actor networks that cohere around and in response to the condition) generate different versions and lived experiences of this condition (Mol, 2002). So too, the digital data–human assemblages that are configured by human users’ interactions with digital technologies are different versions of people’s identities and bodies that have material effects on their ways of living and conceptualising themselves. Part of the work of people’s data practices is negotiating the multiple bodies and selves that these assemblages represent and configure.

Mol’s writings on human subjectivity also have implications for understanding data practices and interpretations. In her essay entitled ‘I eat an apple’, Mol points out that once a foodstuff has been swallowed, the human subject loses control over what happens to the content of the food in her body as the processes of digestion take place. As she notes, the body is busily responding to the food, but the individual herself has no control over this: ‘Her actorship is distributed and her boundaries are neither firm nor fixed’ (Mol, 2008: 40). The eating subject is able to choose what food she decides to eat, but after this point, her body decides how to deal with the components of the food, selecting certain elements and discarding others.

This raises questions about human agency and subjectivity. In the statement ‘I eat an apple’, comments Mol, is the agency in the ‘I’ or in the apple? Humans may grow, harvest and eat apples, but without foodstuffs such as apples, humans would not exist. Furthermore, once the apple is chewed and swallowed, it then becomes part of and absorbed into the eater’s body. It is impossible to determine what is human and what is apple (Mol, 2008: 30) The eating subject, therefore, is semi-permeable, neither completely closed off nor completely open to the world. Mol then goes on to query at what stage the apple becomes part of her, and whether the category of the human subject might recognise the apple as ‘yet another me, a subject in its own right’ (Mol, 2008: 40). Apples themselves have been shaped by years of cultivation by humans into the forms in which they now exist. In fact, they may be viewed as a form of Haraway’s companion species. How then do we draw boundaries around the body/self and the apple? How is the human subject to be defined?

To extend Mol’s analogy, the human subject may be conceptualised as both data-ingesting and data-emitting in an endless cycle of generating data, bringing the data into the self, generating yet more data. Data are absorbed into the body/self and then become new data that flow out of the body/self into the digital data economy. The data-eating/emitting subject, therefore, is not closed off but is open to taking in and letting out digital data. These data become part of the human subject. Data assemblages also represent the individual in multiple ways that have different meanings based on their contexts and uses. Just as eating an apple has many meanings, depending on the social, cultural, political, historical and geographical contexts in which this act takes place, generating and responding to digital data about oneself are highly contingent acts.

Further questions

Drawing on Haraway and Mol, a new set of questions are generated when thinking about digital data–human entanglements. If we adopt Haraway’s companion species trope, we might ask the following: What are our affective responses to our digital data companion species? Do we love or hate them, or simply feel indifferent to them? What are the contexts for these responses? How do we live with our digital data companion species? How do they live with us? How do our lives intersect with them? What do they learn from us, and what do we learn from them? What is the nature of their own lives as they move around the digital data economy? How are we influenced by them? How much can we domesticate or discipline them? How do they domesticate or discipline us? How does each species co-evolve?

Alternatively, using Mol’s concepts of the eating subject, we might wonder: If digital data are never ‘raw’ but rather are always ‘cooked’ (that is, always understood and experienced via social and cultural processes) (Gitelman and Jackson, 2013), and may indeed be ‘rotted’ or spoilt in some way (Boellstorff, 2013), can we also understand them as ‘eaten’ and ‘digested’? What happens when we ingest/absorb digital data about ourselves? Do we recognise some data as ‘food’ (appropriate for such ingestion) and others as ‘non-food’ (not appropriate in some way for our use)? Are some data simply indigestible (our bodies/selves do not recognise them as us and cannot incorporate them)? How are the flavours and tastes of digital data experienced, and what differentiates these flavours and tastes? What role does human agency play in relation to nonhuman agency in the configuration of the digital data–human assemblage?

Taken together, these questions have implications for research focusing on the nature of digital data–human assemblages and practices. Both Haraway and Mol emphasised the political implications of their work. Haraway’s cyborg theorising was developed to explain her socialist feminist principles. In all of her work she emphasises the importance of paying attention as critical scholars to the exacerbation of socioeconomic disadvantage and inequalities that may be outcomes of these relationships. Mol similarly notes the political nature of technologies. In her ‘I eat an apple’ essay, for example, she comments about her distaste for Granny Smith apples, once imported from Chile and therefore associated in her mind with repressive political regimes. As she notes, while she may eat this type of apple and while it may nourish her body as other apples do, she is unable to gain sensory pleasure from it.

Data science writings on Big Data often fail to acknowledge the political dimensions of digital data. They do not see how data are always already ‘cooked’, or how their flavour or digestibility are influenced by their context. Just as ‘eating apples is variously situated’ (Mol, 2008: 29) in history, geography, culture, social relations and politics, resulting in different flavours and pleasures, so too eating data is contextual. Like Haraway’s cyborg figuration (see her interview with Gane, 2006), the digital data–human assemblage may be viewed both as a product of global enterprise and capitalism and as representing possibilities for radical creative and political possibilities.

Perhaps critical digital data scholars could adopt an approach from ‘multispecies ethnographies’ (Kirksey and Helmreich, 2010) in attempting to research the entanglements of humans and their digital companion species, by investigating the nature of co-humanity and the co-evolution of these species, their symbiotic interminglings and becomings and their ‘mutual ecologies and coproduced niches’ (Kirksey and Helmreich, 2010: 546). Such an approach may serve to bring the political nature of these entanglements to the fore, by asking such questions as: What are the mutual ecologies and coproduced niches of digital data-human assemblages? Who wields power over how these ecologies and niches are configured and experienced and who are allowed entrée to them? How are digital data practices and technologies enacted as political modes of engagement within these ecologies and niches?