Strategies for Reflexive Ethnography in the Smart Home: Autoethnography of Silence and Emotion

An Autoethnographic Preamble

An elderly woman – my mum – is living alone with dementia, supported by an array of carers and by us, her three adult children: two living close by and a third who lives 200 miles away. Carers visit in the week, we two siblings who live locally make alternate visits at the weekend, and periodic phone calls from our other sibling to check everything is OK never quite hear the whole story. We two local siblings are emotionally close, but have an increasingly fractured relationship with the third, exacerbated by distance. We all share worries that things can’t go on as they are much longer. The house, owned by my parents since the 1950s, lacks basic amenities expected of the modern age, like warmth. It’s cold in winter, and turning a gas fire on, or remembering not to turn heaters controlled by timers off, is becoming increasingly difficult for mum to manage but still she insists that nothing need change. She always hated change, but now that she has lost the capacity to learn, change is terrifying and always resisted. Without warning the rest of us, one sibling swoops in for a week-long stay, sends mum to stay with her sister, and has central heating installed. The intention might have been good but the emotional fallout is huge. Angry emails are exchanged. With this unilateral act lines have been crossed – the line between the fiction that we’re all in this together and the brutal fact that we siblings are not in any way a team – the line between mum being a decision-maker and a decided-about. But whatever we feel, however bad the fallout, we’re still tied together. Nobody in this set up can walk off on a decisive huff. Life goes on in the regular routine – dementia smooths Mum’s anger at the disruption to her house and within weeks she’s not quite sure any more what the issue was.

In a short time we’re back to the same messy, dysfunctional support team set up. But we’re not the same. The temperature of the house has become a new feature in the family relationships, thanks to the heating system. The new boiler hangs in the kitchen. Just next door high on the wall in the hallway, a smart thermostatic controller is mounted, controlled by an app on the smartphone of the sibling 200 miles away. Before, when I visited and the house temperature felt wrong I could do something about it – open a window, turn on a gas fire, but now I can’t. The controller sits there mute, a blank shiny black plaque. The house feels incredibly hot when I visit. For a woman who has never lived in a centrally heated house and won’t change her usual practice of wearing five layers at all times, this is quite a change to cope with. Mum grumbles. I drag a chair across the hall and climb onto it, standing on tiptoe. I poke at the shiny black plaque and the display lights up. 26°. Smaller numbers on either side, coloured lights indicating… I don’t know what. I poke the display and 26 changes to 25.5, 25, 24.5… have I actually changed the setting? I come back for a look later in the visit and it’s back to 26°. Mum hasn’t mentioned the heat again so I leave it. Next visit, there’s no doubt the house strikes chill as soon as you step inside. A hand on the radiators confirms, no heating at all. Mum is unperturbed. I repeat the process of drag chair, climb, poke. Again, the display lights up − 26°. Surely not. I phone my sibling. The house is cold. They check their phone – it’s set at 26°, that’s right. No, really, I insist, it is freezing cold. No heating at all. After they give me a speech about how I must expect that the heating will go on and off throughout the day because that’s how thermostats work, I insist again, it is cold. They recheck – actually the controller is offline. I crawl under the settee in the sitting room and restart the router. The boiler fires into action and the radiators begin to warm up.

Introduction

Smart homes ¹ are increasingly becoming a mundane reality, as Internet-connected meters, thermostats and switches have become commonly available and smartphones and voice-connected devices proliferate. The autoethnographic fragment above captures several features of contemporary smart home technologies enabled by the Internet of Things ² that are deserving of closer investigation. In particular, the fragment highlights significant silences within the simplistic story of empowerment and control that tends to permeate public discourse around smart home technologies. The official story, as captured in the Hive smart home platform’s ‘Control your home from your phone’ advertising jingle, presumes a single person unproblematically acting on behalf of a household with coherent wishes and a transparently functioning and faultless device. However, the experiences related above suggest that the smart technology does not simply enable and it does not so much reflect the wishes of a pre-existing household as bring that construct into being. In this autoethnographic fragment we are not a cohesive household; we are a messy, fractured, just-about-getting-by, involuntary collective. The smart technology disrupts relationships and occasions questioning of things we took for granted about who did what and where, shifting us into a different form of collective. The flow of data entails also a moral shift in the balance of responsibilities across the collective and it influences our judgement. We clash over the relative worth of the sensory evidence of the person on the spot versus the digital display, disputing who can sense ‘too hot’ and ‘too cold’ and determine the appropriate action. However, only when the technology fails does much of this become apparent. The re-ordering of relationships that the technology occasions is largely unspoken, most of the time. Also silent in my account, now I reflect on it, is any sense of flows of data. Whatever data might be flowing between and beyond the plaque on the wall and my sibling’s smartphone, I thought nothing about it in the moment, and only started to consider when I sat down to write and reflect for this article.

The Internet of Things and the proliferation of smart technologies in the home undoubtedly have significant repercussions for the experience of domestic life and in turn raise the kind of profound question about re-ordering of social relations that sociology habitually addresses. Sociologists are well placed to consider what this way of living entails for those who are living it, and what new forms of emotion, relationship and identity might be emerging, as they have done for previous generations of digital technologies (see, for example, Bakardjieva, 2005; Silverstone and Hirsch, 1992). Generating the kind of rich and detailed insights necessary to fuel such an informed sociological analysis does, however, pose some interesting methodological challenges. Many aspects of these technologies remain opaque in operation and ‘users’ (and indeed sociologists) can often only guess at what decision-making processes and data flows are going on behind the scenes. In fact, the term ‘user’ becomes somewhat problematic, since those affected by a technology such as a smart home may largely forget on an everyday basis that their environment is being mediated in this way. Some members of the household may have made an active choice to welcome the technology, while others are more passive or even unconscious co-habitants with it. The technology participates in the parcelling out of responsibilities – the moral economy of the household (Silverstone et al., 1992) – in insidious, often unspoken fashion.

Smart technologies operate through sending and receiving data, much of it travelling without the conscious intervention of the owners of the technology. There have been numerous calls for ethnographic attention to the experience of smart technologies focused particularly on the cultures of big data, datafication and dataveillance that they inhabit. ³ As boyd and Crawford (2012) argued, it is important not to take claims for the significance of big data at face value, simply by extrapolation from purported qualities of the technology alone. While digital technologies may seem to have transformed a broad array of cultural spheres, an inherent schism with a pre-big data age should not be assumed (Boellstorff, 2013). For example, much of the experience of self-tracking and the resulting digital data is very mundane and deserves to be examined for what it is (Pink et al., 2017) rather than immediately assuming we should view it as dataveillance. In similar vein, Couldry and Powell (2014: 1) argue that the ‘emerging cultures of data collection deserve to be examined in a way that foregrounds the agency and reflexivity of individual actors as well as the variable ways in which power and participation are constructed and enacted’. Various forms of agency in response to perceived computational surveillance are enacted (Knapp, 2016), and these deserve to be investigated ethnographically rather than beginning, and remaining, with an assumption that the qualities of the technology are such that dataveillance, and loss of agency, are inevitable consequences. As researchers interested in sociological dimensions of smart technologies, we may be pre-armed with sensitivity to data flows and dataveillance and conscious of new forms of data capitalism emerging under the guise of personalization and consumer control. We may, however, need to accept that such issues may not be foregrounded at all in the meaning-making processes that surround smart technologies as they become embedded in our everyday lives.

Armed with this intention to take the technology as it emerges in context, an ethnographer focusing on smart technologies as they become embedded in everyday life will encounter intransigent silences and diverse understandings. This article discusses ethnographic approaches to exploring smart technologies in situ, suggesting that to take seriously the experience of smart technologies in action requires an array of methodological approaches, many of which may deviate markedly from a forensic understanding of ‘actual’ data flows. While the pervasive and largely invisible yet highly agentic nature of smart technologies is a new and remarkable situation in some sense, it also maps onto a methodological dilemma common to all ethnographers: how to put into words what is largely silent in daily life and how to bring into the foreground that which is the unspoken background of the everyday. This article therefore focuses on unpacking some aspects of the role of the ethnographer in the specific case of smart technologies in the home by drawing on wider ethnographic literature relating to silence, technology and the role of reflexivity and personal narrative.

The next section of the article maps out an existing set of approaches to exploration of smart home technologies, highlighting key methodological challenges for research that aim to capture the lived experience of these technologies. The article then goes on to explore methodological strategies to address these challenges, focusing on the approach known as the infrastructural inversion for researching taken-for-granted technologies and considering how this approach treats silence as a productive focus. The potential of autoethnographic approaches to enable close examination of otherwise unspoken aspects of experience is then examined. While not all studies may be autoethnographic, for very sound methodological reasons, a reflection on the difference between the autoethnographer’s insights and the kind of accounts generated by methods such as retrospective interviews or by the data logs recorded by smart technologies themselves, gives a strong indication of the need for ethnographers to be reflexive about what counts, for their purposes, as data. In conclusion, the article argues for a stress on reflexivity in the conduct of ethnographic studies of the smart home. Such reflexivity entails continually examining assumptions about the nature of the phenomenon and remaining open both to the significance of the unspoken and to the ambiguous, shifting nature of the field site itself.

Ethnography and Smart Homes

Ethnography of smart homes has been for a long time envisaged as important to provide the kind of rich qualitative description that might inform the design of these technologies (see, for example, Lee et al., 2006) and also to enable interrogation of the processes of social shaping that lead technologies to be used in ways other than those the designers intended (Harper, 2003). Such ethnography often entails recruiting households for a series of interviews, tours of the home, video recording of members taking part in everyday activities, use of cultural probes and shadowing sessions in which the ethnographer spends time in the household: generating data in which members of the household reflect verbally on their way of life and the potential or actual role that smart technologies might play there. Thus, many of the design-focused studies badged as ethnographic explorations of the smart home focus on what people in the household can say about the technologies.

Asking people to talk about these technologies can be a very productive way to understand differences in orientation to and experience of the technology. Mennicken and Huang (2012), for example, carried out qualitative interviews with people who were in the process of making their homes ‘smart’. The interviews were able to highlight varied motivations and understandings of the technology and identify a sequence of phases on the way to achieving a temporary sense of stability in which the technology was deemed to be working as participants wished. Such interviews may be very useful in highlighting difference and in identifying the work involved in becoming and living ‘smart’, but they are not ideally purposed for delivering the kind of nuanced insights suitable for a sociological analysis of a way of living because they focus so directly on one technology taken in isolation as a topic of discussion.

Questioning this standard approach to a design-focused ethnography, Tolmie et al. (2003) highlight the merits of considering the unremarkable in domestic ethnography and interrogating closely the ways in which technologies may become so thoroughly embedded as to become ‘invisible in use’ and yet highly consequential for the routines of domestic life. Extending this reasoning, Crabtree and Tolmie (2016) set out to document the mundane interactions with objects in their own homes over the course of an ordinary day. Such documentation turns out to be extremely burdensome on household members, but instructive in revealing the ‘methodical assemblage of things’ (2016: 1738) that people draw on in unspoken fashion to get everyday life done. The authors use this insight to question the tendency in other forms of design-focused research to isolate out individual technologies as a focus of attention. Moving away from design-focused approaches to understanding smart homes, we might expect a sociologically oriented ethnography to focus even less on relations with the technology in itself and to be more open to encountering the technology as and when it manifests itself, both within interactions with other people in the home and in conjunction with assemblages of things used there. In order to move in this direction, we need to think more about what it means, in ethnography, to ask people to talk about the phenomenon we are interested in knowing about, and what to do when such talk does not naturally arise in everyday interactions.

Ethnography and Silence

A key starting point for an ethnographer of technology who is interested in silence is the notion of the infrastructural inversion (Bowker, 1994; Bowker and Star, 2000). As an analytic stance, infrastructural inversion seeks to counter the tendency of infrastructural technologies to become taken-for-granted and instead draws the infrastructure into the analytic foreground of the study. Because, by their nature, infrastructures are designed not to be noticed, the ethnographer must make special efforts to uncover their silences and to highlight the forms of visible and invisible work that bring them into being and sustain them in operation (Star and Strauss, 1999). Foregrounding the infrastructure, analytically speaking, allows the examination of otherwise unnoticed or naturalized forms of marginalization, exclusion and inequality (Star, 1991). One stance to take on ethnography of the smart home, would therefore be to operate the infrastructural inversion. By bringing the technology into the foreground of the study and exploring the sets of social relations that operate around and through it, we may be able to expose its otherwise unexamined consequences. As an analytic stance the infrastructural inversion offers a powerful resource to orient our attention. Taken as a recipe for the actual conduct of ethnographic fieldwork, however, it would offer some significant drawbacks and potentially lead us into the same set of limitations caused by artificially isolating particular technologies for our attention as outlined in the design-focused ethnographies of smart homes discussed above. If ethnographic fieldwork aims, as Olivier de Sardan (2016: 22) outlines, to produce ‘in situ, contextualised, transversal knowledge, accounting for the “actors’ point of view”, everyday representations and practices, and their indigenous significance’, then the infrastructural inversion as a methodological strategy would run in the opposite direction precisely by attempting to overturn the tendency from the actors’ viewpoint for the infrastructure to disappear. Foregrounding the infrastructure analytically may be a powerful device, but that should not be at the expense of attending to the actors’ experiences as they both attend to and ignore these technologies and combine them with others in assembling the routine business of their lives.

For some insight on this concern about a potential lack of fit between the analytic foregrounding operated by the infrastructural inversion and the ethnographer’s interest in the everyday experience from the actor’s perspective, it is useful to turn to a literature focused specifically on ethnography and silence, to explore the epistemological ramifications of studying silence in its own right as opposed to making attempts to breach silence and encourage actors to verbalize the otherwise unspoken. Hirschauer (2006) argues that ethnographers are always putting into words (or encouraging others to put into words) that which is unspoken, in that ‘the social’ is inherently silent and only brought into words through the ethnographer’s actions. Breaching silence, in this sense, is what ethnographers do. There are, however, more targeted approaches to silence that look specifically at the tendency of actors in the setting not to speak directly to the issues that concern the ethnographer. The actors’ agenda is not our agenda, and they will not necessarily tend of their own volition to speak about the topic that we wish to explore. For this reason, as Olivier de Sardan (2016) argues, fieldwork combines the immersive and observational with the active solicitation of discourses in the interview. For Olivier de Sardan, the interview is an artful process of transforming the questions on the ethnographer’s mind into questions that are meaningful in the terms of the interviewee.

Ethnographers do, therefore, breach silences in some purposive ways and this will inevitably be a part of an exploration of the infrastructural qualities of smart home technologies. There are also, however, ways to appreciate silence in its own right. Scott (2018) argues that there are many different forms of silence, and non-voicing in itself can be a meaningful social action. Pagis (2010) explains that it has become commonplace to think of silence as oppressive and to think of the experience of those being silenced, but there is also an approach to sociology of silence that sees silence as potentially constitutive and meaningful without necessarily being negative. Through ethnography of a meditation room – a setting where silence is a norm and yet a level of intersubjective understanding is nonetheless achieved – an insight into the constitutive nature of silence is arrived at by Pagis (2010).

In a similar vein, the notion of silence not as a negative or lack, but in itself a feature of the setting to be taken seriously played a key role in the ethnography of the operating theatre that Gardezi et al. (2009) conducted. This ethnography was characterized by use of an immersive familiarity with the setting to identify and closely examine instances of silence where there seems to be an ‘unresolved or unarticulated’ (2009: 1392) issue rather than silences as instances of comfortably getting on with the job. These were then further sub-divided into three categories:

(1) absence of communication, made evident by prior actions or communication; (2) lack of response to a direct address by another, or responding with silence to another’s question or directive; and (3) aspects of delivery that blur the lines between speech and silence, such as speaking quietly, timidly or hesitantly. (2009: 1393)

Through this approach the ethnography developed a nuanced sense of the role of silence in getting work done and maintaining interprofessional distinctions. Silence, here, is more than an absence.

Taking this stance into the smart home suggests that as much as the ethnographer may wish to operate an analytic version of the infrastructural inversion and bring the technology onto the agenda for discussion, this does not necessarily translate directly into a methodological strategy. It may be useful methodologically speaking to take account of the silences that surround smart home technologies and to develop a nuanced understanding, through immersion in the setting, of the meaning of these silences, akin to the role of silence in the operating theatre observed by Gardezi et al. (2009). Extrapolating to the smart home, the silences of the technology itself also need to be taken into account, in addition to the silences on the subject of technology among the people in the home – the behaviours of the technology become as much subjects of interest as the interpretations of participants. The ethnographer in this setting needs to become attuned to both the presence and the absence of commentary from participants (and devices) on smart actions and to be prepared for lack of commentary to have multiple meanings and manifestations, just as commentary might. Both participants and devices may choose to draw attention to actions of the device or remain silent about the action. Participants and devices develop an intersubjective awareness of one another, we might say, through various forms of action and silence. Ethnographically, it is of interest to learn the etiquette of interaction on both sides.

Yet another form of silence operates in the shift between actual domestic environments and the advertising rhetoric that surrounds smart home technologies. There is a tendency for the advertising for smart home products to suggest that the inhabitants of a household act in harmony as if one individual. The technology offers that individual ‘control’. However, myriad opportunities for disappointment and breakdown are offered up by the mismatch between the technology’s assumption of a single-minded user and the complex social interactions of a household. Here, it is the complexity of domestic social interactions that is potentially silenced. In an ordinary domestic setting, technologies participate in the performance of identities (Silverstone et al., 1992) through such mundane acts as who has possession of the remote control, who chooses channels, sets the thermostat, who wants the doors and windows open or shut. A domestic hierarchy plays out through the hardware of the home. To an extent this is immediately observable – sit down to watch television with a family, and you see that dynamic playing out before your eyes. Smart technologies may tend to obscure the decision making and render the hierarchy less immediately observable, by separating out the human actions of setting up technologies and choosing parameters from the technologically mediated consequence. To return to Gardezi et al.’s (2009) article on silence in the operating theatre in which they outline that silence may be expressive and strategic – silence is not simply an absence of sound, but instead a different way of achieving action when viewed against the backdrop of norms of interaction that build expectations of what might be said and what is usually said.

Far from being a lively domain of constant debate and negotiation, domestic life may in fact be characterized by silences, as family members find much of their daily routine unremarkable and simply get on with doing it. The spoken parts of domestic life are often precisely about the remarkable. Observing domestic life is, therefore, a case of looking at the unspoken as Tolmie et al. (2003) argue, for it is here that practices become embedded. Even while silent, members of a household are often interactionally available to one another as people who are visibly engaged in various kinds of activity, present or absent, here or there. An ethnographer of domestic life will struggle to work out through mere observation the ways in which members of a household are meaningful to one another. Conversations about who knows what about whom and the interpretations placed on these knowledges will help to unpack the complex layers of these puzzles, just as Geertz’s (1973) imagined ethnographer struggled to unpack the meaning of the movement of an eyelid. As smart technologies become part of the existence of the home, we are then challenged to extend our observations to the actions of the smart technology and the extent to which these become interactionally meaningful to members of the household (and vice versa) and where the silences of the smart technologies become meaningful in themselves. Observing changing temperatures, opening and closing curtains and levels of lighting and observing the extent to which these are matters of comment for participants themselves, become part of the ethnographer’s puzzle.

Silences are contextual. A silence in one place may not preclude speaking about the same thing elsewhere. If this is true of members of a household who offload about their experiences at home in other spaces, then even more so it applies to smart technologies that may be silent in the home about their activities yet speaking elsewhere through flows of data. The boundaries of the home are permeable – information flows in and out, and the household is both a site of communication and communicated about. This flow of data may, however, be more apparent to some members of the household than others, depending on who mediates with the companies that provide the devices and who pays the bills. While the advertising rhetoric puts ‘you’ in control, the one being addressed is largely a single individual who takes the decisions about which technology to have and where. Even to this person, the notion of data flowing beyond the home may not be foregrounded much, if at all. When recently negotiating the details of installation of a new smart meter in my own home, I confidently navigated various practical questions until stopped in my tracks by a question about how often I wanted my smart meter to send data to the power company: hourly, daily or monthly. The question momentarily made visible an aspect of its functioning beyond what I had glimpsed for the technology. Without a background to what happens to the data and what it means to my relationship with the power company I struggled meaningfully to engage with the question.

In the autoethnographic fragment with which I began this piece, there were a number of forms of silence. The device itself built in some silences, in its lack of transparency around how to use and set it, what algorithmic decision making was built into it and who had control of its settings. It is literally silent, in that it makes no noise as it does its work. Its situation high on a wall, and the tendency for its display to remain blank and black unless awoken by a poke exacerbated a tendency to be ignored – a helpful tendency where a forgetful resident needs to be discouraged from fiddling with it, but a less helpful quality for a carer concerned about the temperature. Silence is to an extent designed into the device, but other aspects of its silence were qualities of the fractured relationship of siblings unable to discuss its installation and of the particular set of geographical relations and responsibilities produced around who lived where and how roles were consequently carved out. Silence, in this context, could mean not being told, being protected, being in control, being excluded or being marginalized. Silence could mean a smoothing of friction, or a symptom of a relationship in trouble. The classic methodological response to the silences surrounding infrastructural technologies is the infrastructural inversion. Crucially, this does not necessarily have to entail asking participants what is actually happening or uncovering actual flows of data and algorithmic processes. We need to bracket the a priori assumption that smart home technologies are about dataveillance, or control and develop ethnographic approaches that remain open to whatever participants make of them, even when they do not talk about them.

Autoethnography and Reflexivity

As argued above, a sensitive attention to the nature of silence and to the interactional qualities of unspoken relations among household members and between household members and their smart technologies may be key for the development of a sociologically oriented ethnography of the smart home. Such access to the nuance of everyday life lived within a home is certainly possible for an ethnographer, but difficult to develop over long stretches of time in any other than one’s own home. It seems apt, then, to deploy autoethnography as a potential strategy within the repertoire of the sociologically oriented ethnographer of the smart home. Autoethnography is style of writing that connects the author’s reflections on personal experience to broader social, cultural and political themes, as articulated notably by Anderson (2006), Ellis (2004), Ellis et al. (2010) and Reed-Danahay (1997). As a method, as described by Chang (2016), autoethnography builds a particular kind of self-narrative that both provides an evocative account of experience and interrogates that experience for what it tells us about culture. Particularly notable is the access that autoethnographic reflection gives us to the affective dimensions of daily life, as the autoethnographer produces a narrative based not just on events but also on how those events feel emotionally speaking. Autoethnography is by no means without methodological problems and epistemological shortcomings as Sparkes (2002) highlights, prone to accusations of self-indulgence and restricted as it is to a very singular personal perspective. As a component of a methodological strategy, however, it is arguably particularly well suited to the interrogation of the interactional silences of the smart home and of their affective qualities. It is also flexible as to the boundaries of the study – rather than relying on a pre-designated ‘household’, it can respond to the fluctuating patterns of relationships that exist within and extend beyond a house. It can also remain temporally flexible, allowing for reflection on memory and biographical incidents from the past as well as expectations for the future, as they play a role in present experience.

Arguably, autoethnography is particularly suited to smart technologies. Digital life is insidious, personalized and not readily observable from the outside. Understanding digital technologies, particularly such silently embedded technologies as the Internet of Things, takes close examination and deep reflection and is not something we can achieve by scraping and mining, processing and visualizing digital data en masse. Autoethnography has the capacity to be attentive to the nuances of living a (partially) digital life, looking at the emotions and affects, the silences and absences and enabling us to be more evocative (Gergen and Gergen, 2018) in our understanding of digital technologies. In that vein, a reflexive and evocative autoethnography of smart technologies offers a powerful complement to the existing repertoire of methods to interrogate these technologies. As Hine (2015) argues, autoethnography can usefully pay a close reflexive attention to the varied textures of lived experience with digital technologies.

Thinking about how this autoethnographic insight plays out within the experiences of my mother’s heating controller takes me back to a time that was not happy. The entire experience was fraught with tension, as we tried to manage the complex set of demands around keeping our mother safe and content, working out what she could and could not do and failing to agree as siblings on how to go about it. People living with dementia are often not compliant and grateful with those trying to help them and mum was no exception. Adding into the mix a smart heating controller, albeit one intended to remove the worry of manually controlled heating, was not a source of joy. In those early days of the heating controller, disappointment and frustration were key emotions, with a stomach lurching sense of stress each time as I stepped through the front door on a visit to find the temperature wrong again. I particularly strongly recall being overwhelmed by the heat in the house one day and opening the back door for a waft of fresh air, only to realize that the smart heating controller interpreted this as a threat to its control and was fighting back by turning up the heating. I raged at it that day. Another day, I attempted a manual reset to turn down the temperature directly on the controller and was infuriated as the temperature reverted as soon as I turned away. Had I done it wrong, was there an inbuilt over-ride or was my sibling 200 miles away simply resetting the temperature remotely? I never knew.

This experience is not readily interpreted in terms of dataveillance: data flows were occurring, but the experience cannot be fully encompassed by thinking of data alone. Autoethnography is able to catch and reflect upon the moments of rage and frustration, and to interpret them in the light of the relationships that they exemplify and sustain, the identities that they build and the power structures that they embody. My frustration and rage were borne of a sense of impotence, of my helplessness in the face of the assemblage created through my sibling’s actions, and absence and the presence of this implacable ‘smart’ technology. My emotions were also suffused with the intransigent ongoing frustrations and griefs of my mother’s deteriorating condition and the structures of social care that left us, her children, struggling to cope as we also attempted to fulfil commitments as parents and spouses ourselves.

Writing autoethnographically does not necessarily involve a coherent authorial narrative delivered in a single voice. Just as experience is often fragmented, contradictory and confusing, so might the autoethnographic voice be. We are not looking to the autoethnographer to give us the definitive version of what goes on inside their head, any more than any other methodology can give us direct access to the inner world of participants. My sibling’s account of the incidents that I described to begin this article would, unsurprisingly, be very different from my own. Told by them, different silences, different emotions and different notions of control would emerge. We can look on autoethnography as an opportunity to suspend the search for a singular authoritative account, dwell on the complexity of situations as they span location and time and consider their emotional dimensions. Thinking autoethnographically about a ‘smart technology’ enabled home, we think not about what actual data flows are now occurring but about how this mode of potentially connected living feels, how we experience the world around us, our family and our inter-woven activities. Does this feel like control? Do we feel smarter? Do we feel like the same ‘we’ as before? An autoethnographic approach rejects a pre-ordered approach to data collection and does not expect to know in advance what is significant. Movements through the day, and individual moments of connection become significant in relation to the phenomenon we are interested in as we reflect and dwell on them. Because we lived them, we do not need to capture them in the moment – they are as valid as learning experiences whether we document them minutely at the time or realize only in retrospect that they are shaping meaning making down the line. Taking seriously Olivier de Sardan’s (2016: 128) warning that the ‘hero of the narrative’ must remain at all times the research topic, and not the researcher, an autoethnographic perspective has a lot to offer our understanding of the lived experience of smart technologies, particularly in so far as it allows us closely to examine our uncertainties and lack of knowledge regarding the digital technologies that surround us and the emotions that accompany them.

Conclusion

While it is undoubtedly important for some research to take a revelatory and even campaigning stance in relation to data practices, in this article I argue that it is also useful to maintain some space for an ethnography of meaning making that does not assume that we already know what these technologies do and instead discovers their qualities alongside participants. Taking this stance involves stepping aside from the notion that there is a ‘real’ story to uncover about smart technologies in the home and accepting that some kinds of meaning making around these technologies may not, as such, focus on data. The ethnographer of the smart home needs to be critically reflexive about what it might mean to be smart and about the lived reality of dataveillance. It is questionable whether unveiling the ‘actual’ operations of the technology is ever part of the ethnographer’s role. This kind of reflexivity entails sustaining a speculative and provisional interest in what smart homes might turn out to be in any given circumstances, in the face of any apparent certainty about what this technology in fact does that might be encountered from designers, marketers, critics of dataveillance and participants themselves.

Autoethnography allows an analysis of the silences and actions of daily life. Reflecting on the process of arrival of smart technologies in the home gives an insight into the involvement of various household members, both explicitly and in their absences of attention. The autoethnographer is well placed to reflect on the affective dimensions of the lived experience of smart technologies, exploring the emotional response to the technologies themselves and of the relationships that they mediate with other household members present and absent. The autoethnographer also develops a nuanced understanding of what silence means as people go about everyday lives suffused with smart technologies. Studying these experiences in conjunction with the advertising materials and user manuals published by device providers gives cause for reflection on silences on their part. Ethnography might well involve drawing on logs of data produced by sensors and also reflecting on these with participants, but, crucially, the data logs would not be taken as transparent depictions of what people did but treated as resources for negotiating understandings of the situation by participants and ethnographers alike. The ethnography of lived experience of smart home technologies can aim to avoid separating out data flows, algorithms and material devices as separate things, but focus instead on presences and silences across the digital and material. This entails never settling on a single story of what the technology does, but rather developing shifting narratives that look backwards and forwards and shape themselves in the context of whatever the present moment offers.

Ethnography of technology ‘users’ is somewhat problematic in the face of these forms of ubiquitous computing that were designed as a technology that disappears (Weiser, 1991). ‘Data subject’ likewise is too clinical, assumes too much, is too much framed in the vocabulary of the data collectors. As Goulden et al. (2018) argue, what is often termed ‘personal data’ tracked by sensors in fact becomes interpersonal data as people puzzle over its meaning, negotiate how to use it appropriately and use it to reflect on one another’s actions. The many and varied ways of being sensed and sensing presence and the varied forms of co-presence with technologies that do not necessarily involve a conscious usage, need to be examined afresh without assuming that we have privileged access to what it is that the technology does. Just as researchers become conscious of the consequences of datafication, so they have to deal with its disappearance into mundanity as simply part of the way life is lived, paralleling Deuze’s (2011: 137) observation about the media that ‘the key challenge of communication and media studies in the 21st century is, or will be, the disappearance of media’. The disappearance of data in everyday consciousness is both a methodological challenge and a site of anxiety in the face of commentary on the insidious powers of governments and monolithic corporations.

Ethnography seeks to take seriously the actor’s perspective, and autoethnography builds on this through offering a detail of insight into the messiness and contingencies of that ‘perspective’ and its emotional dimensions that is rarely offered by other forms of ethnographic data. For study of smart home technologies, autoethnography acts as a site for examining a flitting back and forth between data consciousness and data ignorance, a seeking to know and ignoring of data flows in the context of more immediate relationships and preoccupations. The approach to autoethnography of smart home technologies that is described here aligns with the stance of taking a ‘non-digital-centric approach to the digital’ (Pink et al., 2015: 7) that does not tackle these technologies head-on, but explores them as they arise among and are embedded within everyday practices and accepts them as the actors there understand them to be. Such an approach is squarely in line with the vibrant tradition of ethnography as respectful of the beliefs of others as we find them. This may provoke, however, a certain anxiety that we may be somehow letting the perpetrators of dataveillance off the hook by failing to focus our attention squarely on data flows and power imbalances between the surveilled and the surveillors. The challenge is to couple a critically oriented infrastructural inversion on an analytic level with a non-infrastructure-centric ethnography of infrastructure at a methodological level.