Software and the Self

The self. Software. One is an abstraction that has provided food for thought for philosophers through the centuries; the other did not exist as a word until three decades ago. Yet they have become increasingly intertwined since then. Software has become a metaphor for the self; it serves to mediate it, and perhaps we can even think of software as having a self in some cases. More concretely, the Ego-Media project over the past few years has sought to map how the self is presented, documented, and written up in an online context, a process that inevitably involves software.

In this essay, we discuss a number of different ways one may think of the self in combination with software. The primary interest here is the self as seen through software, which is the perspective taken by many of the studies within the Ego-Media project. However, this can be contrasted with other interpretations of the pair, such as the self as software or the self of software, which is also briefly discussed to more clearly position the work that was done as well as suggest some possibilities for future research within this frame that could build upon the results of this project.

After a brief exploration of the history of the relationship between software and the self – looking at identity, the use and critiques of software as a metaphor for the mind – this essay looks at how selves might be, and are, expressed through software. How, we ask, is the self mediated by technology? What do we need to understand about the affordances of both software and platforms? What do we need to understand about how these evolve? About how users work – and innovate – within them?

This ever-shifting combination of software and self allows for research along a range of possibilities from, say, critical code studies, through platform studies, to ethnographies. We outline some possibilities before, finally, exploring the ideas of the self after software – and a self of software that takes us beyond chatbots.

The combination of software and the self has intrigued thinkers for as long as software, or computers in general, have been in our lives and perhaps even longer. As early as 1637, René Descartes mused whether an automaton could be considered to have an autonomous identity in his Discourse on the Method (he thought not).¹ Some centuries later, when automatons and programmable computers had become relatively commonplace, the question of whether such a computer could be the substrate of an autonomous identity became more practical, as did the question of whether perhaps our brains in themselves could be modeled as hardware on which our consciousness runs. Thought experiments like Alan Turing’s Imitation Game² have as an underlying principle that at least theoretically, software could present a convincing facsimile of a fully actualized individual, with an identity that is expressed through software interaction. In other words, in this frame, the mind and the self emerging from it are (indistinguishable from) a computer.

This is an intuitively attractive position for anyone with mild transhumanist inclinations, or even anyone who owns a smartphone; if our mind is a computer, and increasingly much of what we do in everyday life happens inside a computer, the two are fundamentally compatible, and all that is needed to exponentially increase the efficiency (and thereby, joy) of life is a better interface between these two computer systems. It additionally, and perhaps conveniently, elevates those with the right programming skills – the “virtual class”³ – to potential masters of the mind. From this computational fetishism a “hackers-as-magicians”⁴ frame emerges that calls to mind John Perry Barlow–style early-1990s cyberutopianism.⁵

However, while it is exciting, this computational theory of mind has been rejected as naive in scholarly discussion. As Wendy Chun notes in Programmed Visions, this idea of software as a stand-in for any other machine, however complex, is rather dubious.⁶ Even if one for the moment sets aside the problem that human (or animal, for that matter) consciousness is not at all understood well enough to be able to credibly model it as a machine, this “vaguest understanding” of software⁷ naively assumes that any machine (or more accurately, what the machine does) may be represented as computer code, without actually considering what that would look like in practice and how one would go about modeling the mind as software.

Philosophers of mind also have a bone to pick with the computational frame. John Searle argues that the notion that the mind could be modeled as a computer program “obviously” needs to be rejected as minds contain “intrinsic mental content” that is lacking from a computer program which is pure code;^8,9 in this view, code is at best a small subset of a mind, a description of how some of its pathways might work but lacking the substance that makes the mind into an entity from which a self emerges. The “computationalist” tendency to try and fit human minds into a software or computer frame has also been criticized by others. Selmer Bringsjord, in support of Searle, writes that even if a computer program were produced that could succesfully complete the Turing Test, it would in comparison to a bona fide human amount to no more than a “zombie” which would exhibit a superficial similarity to human consciousness but lack more advanced modes of self-actualization;¹⁰ here again such a zombie would perhaps cover part of what a mind is capable of but far from all of it. Even if a computer program, or software, could (if for now only theoretically) cover all of the capabilities of a real human – as Daniel Dennett¹¹ has argued – the issue that we barely understand how the mind works remains, and with it the immense difficulty of doing anything practical with the notion that the mind, and the self, can be viewed as software. Thus, such a discussion remains – for the time being – the territory of philosophers of mind like Searle, Bringsjord, and Dennett.

However, all is not lost, as the pair of software and the self may be thought of in other combinations. Next to the self as software, there is the self through software; one’s identity expressed through the medium of a software platform such as a social network site or internet platform, or more prosaically, an avatar in an online game. The self here becomes mediated by software; even if it is perhaps not computational itself, it is (in many cases) perceived through a computational system. It is still expressed by a (flesh-and-blood) person, but increasingly often that expression can only be observed or taken in through a software interface – an app or a website. Of course, mediation of the self is not a novel concept; computers are hardly the only medium for self-presentations. If we think about life writing as a self-presentational practice – and what is it if not that – a rich history of mediated self-presentations exists in the form of, for example, autobiographies. But arguably, software presents a new dimension to this practice by virtue of being dynamic; it can be reconfigured, changed, and iterated to produce new possibilities and modes of presentation and expression with relative ease.

This notion is more practically applicable, and popular social media platforms are in fact very much aware of this themselves, which one can verify by a simple look at their slogans: Facebook asks for people to “share everything in their lives”; Twitter used to ask “what are you doing?” (it now is more generically interested in “what's happening?”); in the case of YouTube, a simple look at the platform’s name suffices. Share your life, tell us what you are doing, show others what you find important, they ask us, with the implicit addendum that any of this needs to happen on their platform, in their app, or through their API.

Of course, this makes the benign-sounding invitation to “share” or “post” a decidedly nonneutral action. José van Dijck and Thomas Poell point at the “double-edged logic” of platforms’ use of such metaphors that on the one hand promise all kinds of benefits one will get access to by “sharing” content on an “open” and “transparent” platform, while simultaneously having their own for-profit interests with these data that is left ambiguous through such language.¹² This then further emphasizes that in the process of expressing or presenting the self through software, there are two major agents at play: the person expressing themselves and the entity designing or controlling the software this expression happens through. This, then, raises questions about what role the software and its interface play in that process of self-expression; do they perhaps affect how we express ourselves (and our selves) and if so, can we study and map that effect? And what other factors are at play here, besides the nudges the software and its owners give us; how do we square those with our own motivations and ideas about our selves?

In a broader sense, all this is a question of how the self is mediated by technology, which of course is by no means limited to computer software only. The invention of the photo camera invited ruminations on how one’s essence is captured on film or even how the self may be expressed through the pictures one makes (see, e.g., Cathcart and Gumpert’s I Am a Camera¹³). This attention to the medium of course brings to mind McLuhan’s adage that this medium is, in fact, the message¹⁴; in this context one could read this as a call for attention to the specificities of the software in mediating one’s self as it is expressed in an online context.

But in a modern setting, the question about how the self is mediated through software is in practical terms usually a question about platforms. Platforms like Facebook, YouTube, Instagram, Twitter, WeChat, TikTok, MySpace, and SnapChat are where one goes to self-present; where people can document their lives by “sharing” updates about their lives and curate their self as perceived by others using that platform. These are platforms because they are never simply one app; they are part of a larger network of apps, websites, and services. Personal data are added to this network by using the platform’s apps, or other apps and websites embedding it. The platform thus serves as an “empty structure”¹⁵ in which users may craft their own self-presentation by choosing what data to provide within it and under what conditions and to whom they may be shown.

The notion of the empty structure here can be problematized in two ways. One, it is often very difficult to actually keep the structure empty; many platforms constantly nudge those using it to post, share, and react; and as soon as one does so, their activity is used to more accurately and effectively nudge them to post, share, and react. And second, and perhaps most pressingly, while the word structure perhaps implies a neutral, minimal scaffolding, the structure offered by any platform has a major impact on what can be done within that structure. Niels Brügger notes that the empty structure “control[s] users as well as offering them opportunities.”¹⁶ By including certain features while simultaneously withholding others, platforms then have significant influence on the shape taken by anything published on the platform and thereby the emerging self-presentation.

If we are to understand how the self is presented and performed through software, then, we are to understand these forms and affordances, something that Rebecca Roach’s work on Talking Interfaces and Rob Gallagher’s on Voice address very explicitly. Of course, much work has been done on software affordances, also in the context of online social platforms; Anne Helmond and Taina Bucher, in a review of existing literature on the topic,¹⁷ found enough work to distinguish five distinct usages of “affordances” in studies of social media platforms. Acknowledging that these affordances exist and are operative in our interaction with the software is nevertheless only part of the work here. A further question is where these affordances come from, why they exist, and thus why software shapes our self-presentation the way it does. In this context it should further be noted that affordances do not determine what will happen outright; they nudge, they prod, they suggest, but within the hard constraints set by a platform (such as character limits or picture sizes) people still have room to fill the empty structure of the platform to their own wishes, even if some of those wishes are easier to actualize than others.

This is important; one should not reduce platforms to a particular type or genre of self-presentation, ostensibly enabled by their affordances. Rachael Kent, in her work within the Ego-Media project on self-tracking of health on Instagram, found radically different motivations and use patterns that could be linked to factors as diverse as users’ own health history, desires to maintain a positive image, and other external factors. To be sure, another factor found was what kind of picture works well on Instagram, i.e., how one can get the most likes and comments; these are then linked to the platform’s recommendation algorithms, interface, etc. But overall, there is room here within the platform’s structure for innovation and finding one’s own way.

The breadth of that possibility space was the focus of Stijn Peeters’s work, which sought to map the history of a number of “particularly enlightening”¹⁸ genres of use of the platform, investigating to what extent they can be explained by the platform's affordances. Focusing on Twitter and IRC, an analysis of two different platforms can then shed light on how particular ways of expressing oneself have developed within and in negotiation with the specific materialities of that platform, and to what extent such material factors are crucial to the development of particular genres of expression. In the course of this work, it became apparent that while there are clear links between some types of expression found on these platforms and their affordances, people also exhibit considerable inventiveness in thinking outside the box and bending a platform to their needs and wishes, contrary to the goals of the platforms’ owners and developers. For example, while Twitter allows people to indicate their appreciation of a tweet through a like button, no such feature is present in the 1990s-era IRC; but people have set up external websites to archive noteworthy bits of chat, complete with rating systems, that may then be linked to within IRC through the use of chatbots set up for that purpose. Of course, the platform’s affordances are still important here; if IRC did not make it as easy as it does to make chatbots for the platform, a different solution would have to be found. But such examples do show that platforms are still very much a structure within which people can find their own way; the shape of that structure however is worth investigating.

It is this combination of software and the self, of the self as mediated by and presented through software, that has been most central to the Ego-Media project. Some examples of this have been given above. These and other subprojects have looked at self-presentational practices through this lens, discussing along the way the many shapes such a mediated self can take. An attention to affordances and features, the shape of the structure of a platform, suggests a focus on the materiality of software, the way the characteristics of software inevitably shape whatever is expressed through it and thereby any self-presentation. Materiality then means the particularities of the software, in the interface people have to negotiate when using it; the constraints and limitations inherent to the software design; and also the embedding of the software in the physical or virtual environment that through its own materiality and affordances shapes the software.

More practically this allows for research along a range of possibilities. Given the two major factors in the equation – the platform and people using it – a gradient may be envisioned where one or the other can be more in focus, or alternatively a balanced approach is also possible. On the far end of the scale one could place approaches like critical code studies¹⁹, which engage with the code of the software itself as the primary research object. Moving toward a less software-oriented and more self-oriented part of the scale (and skipping some possibilities for the sake of brevity), one may find the platform study,²⁰ where the focus of an analysis is the platform on which whatever is being analyzed happens; indeed the platform itself becomes central to the analysis. While the platform – and its software – are primary here as well, they are explicitly positioned as objects to be studied through what is produced by their means, which in this case would be a self-presentation (or many self-presentations). In this view, the self enters the frame as an object of study, as a lens through which the materiality of the platform can be studied; conversely, this then yields insight about the (material) platform as a lens through which the self is presented and viewed. In the Ego Media context, this is the approach taken by Peeters’s research, in which the materiality of IRC and Twitter was studied through a comparative analysis of these two platform studies to try to tease out self-presentational affordances, how these are developed over time, and what factors affect this development.

This approach then foregrounds the software on which the self is expressed and performed. The fact that the self is mediated, written, or performed into being on an empty but demarcated slate is not only an axiom but also something to zoom in on and investigate further, to map this process and the exact demarcation of that slate, and the possibility space existing within it. As discussed elsewhere, these “starting conditions” have significant effects on what modes of expression become popular or even possible on a platform. This underlines the relevance of studying the software of the self, which in practice means studying online social platforms.

Of course, platforms have been studied for far longer than the label of “platform studies” has existed, which is since 2009 when Ian Bogost and Nick Montfort coined the term with their study of the Atari VCS platform. Earlier examples of platform-focused research abound; as early as 1991 Elizabeth Reid published a much-cited study of IRC and its conventions;²¹ Nick and Michael Hauben’s 1997 Netizens²² presents an account of the history of Usenet; and Howard Rheingold’s 1993 The Virtual Community²³ is a classic study of the W.E.L.L., a contemporary internet forum. All of these focus (primarily) on a single platform. In some aspects these works (and others like them) are then proto-platform studies; had Bogost and Montfort’s work been available at the time, it would surely have been cited as foundational. What however does set these earlier texts apart from more recent work on the subject is a less explicit and less central focus on the platform’s materiality. Instead, the approach can perhaps be characterized as more ethnographic in nature; primary to these analyses are the practices and culture of the people using the platform, and while in the course of discussing this the platform’s features and affordances figure, these are illustrative rather than the objects of research in themselves.

As ethnograpic studies, with a focus on culture and community rather than software, such work is yet further along the aforementioned gradient between a focus on software and a focus on the self. But even within ethnographic approaches some are more interested in the software, or technology, while others put the focus on the self. There are, of course, instances of classic ethnography, where a researcher employs participant observation and regular interviews to map a culture’s practices and values. Bonni Nardi’s My Life as a Night-Elf Priest,²⁴ which investigates identity play in a World of Warcraft community, is a brilliant example of this. But since the late 1990s, ethnography on digital platforms has been developed in multiple directions, sometimes shifting the technology back into focus. Closer to the platform study, for instance, is Taina Bucher’s “technography” approach²⁵, where one conducts an ethnographic study not of a community or group of people, but of a particular technology or technological system. This would be a particularly software-oriented ethnographic approach.

Two Ego-Media projects can be positioned in roughly this spot on the software-self gradient. Mikka Lene Pers’s work on the subculture around mommy vloggers, investigates young mothers who make YouTube videos documenting their lives. While the data are very much platform-based – YouTube for the primary vlogs and comments on them, platforms like Facebook and private forums for the subculture that has emerged from those comment sections – it is again the cultural practices that emerge from the posting of and reactions to mummy vlogs that are of primary interest. Furthermore, Rachael Kent investigates the quantified self on Instagram; here again the study is concerned with one specific platform, but the affordances of that platform are not so much the primary object of research, but rather serve as a backdrop for a careful investigation of the attitudes and motivations of people who document their health on the platform. The self-presentation, in other words. Here the people presenting their selves on the platforms are invited to reflect on their practice and keep a diary, allowing for a closer look at their motivations and preoccupations with a level of nuance that can never be teased out from their posts alone. As in Alice Marwick’s famous study of microcelebrity on Instagram,²⁶ a small sample of people are asked to reflect on their self-presentational practices, in Kent’s case supplementing such interviews with a diary of their posting behavior. As detailed earlier, this allows for charting factors external to the software, thus providing a necessary addition to work that takes a more platform-centric approach.

Alexandra Georgakopoulou’s research into the addition of Story features to Snapchat and Instagram likewise strikes a balance between addressing the design (and redesign) of platforms on the one hand and considering the practices of particular users on the other. Clare Brant’s study of emoji as vehicles for imaginative agency, meanwhile, considers how expressive possibilities emerge from the interplay between software, hardware, and users; while Unicode decides which proposals for new emoji will be accepted or rejected, different platforms and devices render the same emoji very differently and may change their designs in response to users’ habits and feedback (witness the addition of an extra pair of legs to the hitherto anatomically aberrant lobster emoji). While these parties might be thought of as determining what it is possible to say using emoji, users nonetheless concoct their own in-jokes and forms of innuendo, finding new semiotic affordances in familiar pictograms as they attempt to express themselves.

The perspective of the self as mediated through software then allows for a multitude of investigative focuses and methodological approaches. The benefit of a wider project within which this topic is studied is that numerous versions of these can and have been covered. Some detailed examples are given above; together, these studies, while they cannot offer an exhaustive account of how the self is mediated through software and what role the software plays in that process, nevertheless provide a broad overview of the diversity of possibilities.

One could envision further combinations of the self and software beyond the-self-as-software and the-self-through-software. For example, bearing in mind the work done on mapping online self-presentational practices within the Ego Media context, it may make sense to think about the self after software next. Have ideas about the self changed since we started actualizing them through software? Since Erving Goffman’s work on self-presentation in the 1960s it has been well accepted that the self is performative, that one presents different “fronts” depending on the situation. What has perhaps changed is that this is now done quite explicitly in some cases; as shown in Kent’s work, people are very aware of how (in that case) their Instagram account presents a particular impression and how they can craft it accordingly. This is also supported by other research on Instagram, for example that investigating the “Finsta” phenomenon, where people maintain two accounts; a Finsta (fake Instagram) account to be viewed by the general public and a “Rinsta” (real Instagram) account in which they choose to show a more vulnerable and, therefore in their view, more “real” presentation of themselves.²⁷ Is this a departure from earlier modes of self-presentation, and has social media software introduced a new awareness of the performativity of the self?

Moving the horizon a little further out, perhaps one could also entertain the thought of a self of software. Though they have existed for decades,²⁸ the development of chatbots has sped up considerably in the past few years, on the tail of developments in machine learning and increased computing power. Where chatbots were once limited to simple predetermined responses to hard-coded commands, modern “voice assistants” such as Alexa and Siri are an increasingly mundane part of everyday life. These are, quite explicitly, bots with an (attempt at) personality: a voice, a name, even a somewhat stilted attempt at a sense of humor. Is this technology at the point where next to a self being expressed through software by others, we can speak of the software expressing its own self? And if so, how is it presented? Can we attribute agency to the software itself in expressing it, or does that agency belong solely to the creators of the software?

These are but two pathways that may be considered for future exploration of software and the self. However the two are combined, it is clear from work done for Ego Media and elsewhere that things have been changing, sometimes rapidly, as software has become more powerful, interconnected, and central to our self-presentational practices. Primary to the work of Ego Media has been the perspective of software as a mediator for self-presentations, thus presenting the self through software as a research object. Within this there is room for various focal points, from a focus on the software and its affordances to a focus on online self-presentational practices and motivations. The sum of these studies is a showcase of how complex the relation between software and the self is, and how rich the palette of self-presentations afforded by software is.

Theme essay:

Time

Chapter:

Social media case studies